On July 21, ISSC bought Aydin Displays, a maker of ruggedized displays for military applications, for $24.5M. ISSC guided Aydin’s revenue to $16M for 2026.

Yes. ISSC really did buy the manufacturer itself, with everything that comes with it: engineers, plants, patents, and so on. This is a different animal from the handful of product lines or licenses ISSC has folded into Exton before.

We’re talking about a 40,000 sq ft plant in Birdsboro (about 45 minutes by car from the Exton plant), around fifty employees, production lines, equipment, and an entity with its own P&L and its own balance sheet. One catch: the margins weren't disclosed. Not for now, anyway. It’s possible ISSC never publishes them.

All of this points to one thing: to understand and judge this acquisition, you have to forget the paradigm inherited from the six previous ones.

Understanding this acquisition and its implications first requires knowing ISSC and its subtleties well. I’ll assume you do (just in case: here’s my full deep dive on the company).

I dug through the SEC archives to arm myself with everything I need.

First, what they bought, in detail.

Table of Contents

1. What ISSC Bought: Aydin Displays in Detail

2. Aydin Displays’ 13-Year Ownership History

3. Why ISSC Acquired Aydin Displays

   1. The Obvious Play

   2. The Less Obvious Play

4. Valuing Aydin Displays Without Published Margins

5. How ISSC Financed the $24.5M Acquisition

6. Aydin Displays’ Market Position

7. Verdict

1. What ISSC Bought: Aydin Displays in Detail

Aydin makes ruggedized displays that have to survive where other displays have gotten used to dying. On the deck of a destroyer, for example, deep inside a submarine, or embedded in an armored vehicle. The company supports "over 20 military platforms across more than 80 countries out of its vertically integrated manufacturing facility."

Their products range from 10 to 65 inches, are AS9100D certified (the quality standard for aerospace and defense), and have been sold to the US Navy and its prime contractors for 50+ years. Those contractors (BAE, Lockheed, etc.) account for the bulk of their defense sales. Direct purchases by the Navy are mostly limited to buying a few spare parts.

Fine, at first glance you can file this under the defense-focus strategy Askarpour put in place when he arrived in 2022, and hope it creates as much value as the F-16 lines. But here ISSC is playing a new game.

In its previous acquisitions, ISSC bought selected assets, left the rest to the seller, and obtained its own TSOs to get the regulatory right to run those lines (product by product). That took about a year, with some inefficiencies and a bit of lumpiness in revenue in the meantime.

Here ISSC bought the whole company for $24.5M. Of course that includes the engineers, the lines, the patents, and so on, but also, and above all, its CAGE code (the company’s identifier with the government and defense customers), its customer contracts, and its certifications and qualifications. All the assets, but also all the liabilities, which for now are unknown to us.

The $24.5M was funded by a draw on the JPMorgan facility, on a cash-free/debt-free basis and adjustable on working capital within 90 days (the impact should be minimal). No escrow, no holdback, everything is paid at closing, and any bad surprise ISSC discovers will be settled through a claim directly against the seller. So it’s either proof of confidence in what’s inside the box, or proof of confidence in the seller’s solvency, or, very probably, both.

The seller is Sparton Corporation, an Elbit Systems of America entity, itself a unit of Elbit Systems, an Israeli defense company. We’ll come back to it, because the seller’s origin is itself a source of potential synergy, and therefore of value creation.

To quickly sum up the rest:

• Elbit signed a Supply Agreement and therefore remains a customer of Aydin (terms not published),

• Elbit signed a five-year non-compete clause for the United States and Canada (Aydin sells its displays to Royal Canadian Navy frigates),

• a TSA (Transition Services Agreement) under which Elbit commits to providing Aydin with its usual support services for the length of the transition (ERP, accounting, and so on, which ISSC will therefore have to absorb).

All of it signed and closed on July 21, with an extension of the Birdsboro lease. So Aydin’s plant stays, at least for a while.

2. Aydin Displays’ 13-Year Ownership History

Usually I’d have gone for irony before disappointing you, but no irony this time: ISSC didn’t give us Aydin’s margins. No P&L, no EBITDA, in short, nothing that lets us properly judge the multiple paid for this acquisition.

But you're not here for descriptive work, so I dug a bit. My starting point: Aydin wasn’t born on July 21, 2026 (a very original thought, all things considered). It has already been sold, several times, by owners who were themselves listed. Their official filings therefore give us Aydin’s old numbers (among other things). Let’s go back in time.

2013, to be precise. Aydin belonged to Video Display Corporation, a fading American display maker that put it up for sale. The pro forma accounts filed in connection with the sale give us a precise but old snapshot of the company: that year, Aydin did $17.6M in revenue (yes, against $16M guided for 2026), and at an operating loss on top of it.

Sparton (a defense electronics contract manufacturer) bought it for $15.5M, i.e. 0.88x sales for a company losing money. The vehicle Sparton set up for the deal was called Sparton Aydin, LLC. The same entity ISSC just bought 13 years later. Okay, but what about the Sparton era?

One clue: the contract provided for an additional $6.6M earnout (an extra amount paid to the seller only if Aydin hit a target EBITDA within 12 months). I dug through the filings on both sides, and I found nothing anywhere. So the EBITDA threshold very probably wasn’t hit (”absence of evidence is not evidence of absence” saves them).

It probably started badly, but it definitely continued badly.

The ruggedized display backlog fell from $14M to $9M in just six months in 2016. Meanwhile, Sparton made acquisitions to grow Aydin: IED in 2014 and KEP Marine in 2015, for a total of $7.6M. To sum up: what little growth there was got bought, or rather rented. At a steep price. Let’s do a bit of arithmetic. Aydin alone in 2013: $17.6M in revenue. 13 years, two acquisitions and $7.6M later, revenue is guided to $16M. In nominal terms, it’s very ugly. And that's before adjusting for inflation. Before anyone nitpicks: this is still defense, margins can’t save the situation.

So ISSC is buying the company after four successive owners, none of whom managed to grow it, and it’s doing so at 1.5x 2026 revenue. Hence the importance of the margins we don’t have. It’s possible ISSC gives us more information in the 8-K/A that should come out in the next 90 days. Possible, not certain. So we have to judge without it.

Historically, the acquisition can look expensive, but ISSC is an unusual buyer with the ability to unlock forms of value that are out of reach or trivial for other owners.

Subscribe now

3. Why ISSC Acquired Aydin Displays

3.1 The Obvious Play

ISSC is different, but the previous buyers had their own particularities too.

Video Display Corporation was in its death throes. Aydin was a way to scrape together some life-saving cash, even at the cost of dumping the business. The simple fact that the company itself was losing money is very telling. A good business in the hands of a bad manager will very often become a bad business.

When Sparton bought it, it immediately had more important things to deal with: a takeover attempt by Britain’s Ultra Electronics (blocked by the US Justice Department in 2018, so years later). After the failed takeover, Sparton was taken private by Cerberus in 2019, while divesting its contract manufacturing division on the side. Aydin and its $15M of revenue at the time were almost a rounding error for a company fighting for its survival.

Now, Cerberus is private equity. The goal is to sell fast at a profit. So it sold Sparton to Elbit two years later for $380M. Aydin was more of a throw-in than a bonus. Elbit already made some of its own military displays, so we can safely say Aydin was anything but their priority.

In 13 years, nobody had the strategic reason or the means to grow Aydin (we’ll come back to why in detail later).

ISSC is the first owner for whom the company and its products matter. Displays are one of the six cockpit building blocks it already assembles. What counts in the Aydin purchase are the assets ISSC was missing.

Here are, in my view, the three things ISSC got out of this acquisition. Two out of three follow the same process as the previous acquisitions:

1. Access to a new market segment, in this case naval and land. Before this acquisition, ISSC sold exclusively cockpit parts for aircraft. Aydin is an open door to the decks of ships and the turrets of armored vehicles, to their respective primes, their programs and their replacement cycles (aftermarket, what a lovely word). But there’s more... let’s wait for point 3.

2. Good engineers. Management had already mentioned the difficulty of finding good engineers. They bought the plant 45 minutes away by car, around 50 employees, engineers included. Clever. And they’re proud of it: “bringing us exceptional engineering talent, proven display technologies, and a respected product portfolio that aligns closely with our expanding family of aerospace solutions.” The CEO mentions the engineers first. In a way it resembles ISSC’s strategy, buying certified production capacity rather than building it yourself, but with the plant and the people already trained.

3. And finally, the eternal cross-selling (a lovely word too). Aydin’s displays go into ISSC’s cockpit and mission solutions, and ISSC’s catalogue gains an entry point with Aydin’s naval customers and primes. In a sector this locked up, excellent salespeople would never have made that happen, even if paid $24.5M for the job. At first glance, the current cross-selling opportunities look relatively small compared to other acquisitions, but in a longer-term defense-focus strategy, future acquisitions will potentially (probably) improve and round out those cross-selling opportunities in land and naval.

Nothing revolutionary at first glance, it follows ISSC’s strategy completely, with the engineers on top. But that’s not all. That was the obvious part, but I strongly suspect ISSC has other “assets” in its sights, the kind you’ll never find on a balance sheet.

3.2 The Less Obvious Play

The rest is for paid subscribers.

You get 1 free article. If this one got under your skin, help yourself.

But future ISSC updates will be locked, along with every other company I cover or will cover.

For those interested: I publish 20+ deep dives like this a year, with regular follow-ups, up-to-date valuation models, and a lot more. I'll leave it at that, my content is my best pitch. Go have a look.

Read more

The slightly-less-short version: Cabral Gold goes into production in a few months and will use part of that cash flow to drill the 50 targets across its Cuiú Cuiú district (Brazil) and to prepare a PEA (Preliminary Economic Assessment, the first economic study of a deposit) in 2027, for a mine on a much bigger scale.

So what we have is a developer one step from becoming a producer, and at the same time an explorer with a lot of targets to drill and potentially a lot of discoveries ahead.

That parks the company in the two moments on the Lassonde curve that offer the most re-rating. For the uninitiated, the Lassonde curve is the valuation path a miner follows over its life. Think of it as what the “laws” of economics and psychology do to a miner. The two big re-rating phases come after a significant discovery and on the approach to production. Precisely where Cabral sits.

And it doesn’t stop there. Re-rating opportunities? Cabral has plenty. But one thing at a time.

Before we even analyze the company itself, let’s take a quick detour through the gold miners. Four pretty charts summed up in five sentences:

• Gold miners are historically cheap relative to gold.

• Gold miners are historically cheap relative to the S&P 500.

• Gold miner valuations haven’t kept up with their cash flow growth.

• Gold miner margins are historically high.

• All of it despite, or thanks to, one of the worst quarters in the history of gold.

That’s it, I’m done with the sector. No macro. It’s outside my circle of competence. At most, it should be free optionality. The only acceptable way to profit from it is finding a gold miner that justifies a thesis all on its own, without having to assume a higher gold price or a sector re-rating. Cabral Gold is cut from that cloth.

Like Atlas Salt, the setup is drowned in complexity and geological subtleties, the kind that takes an enormous amount of work to properly get your head around. I did that work, then reworked the whole thing so the uninitiated can follow it end to end.

The full write-up is 35k words. For those in a hurry, a 10-minute version of the most important bits lives here.

Before we start, a thank-you to , who put the company on my radar. Don’t let the name fool you: the guy is genuinely one of a kind, and generous enough to share his portfolio and his sharp analyses for free on X. Follow him. You won’t regret it. I don’t.

Follow him on X

Let’s go.

Table of Contents

0. The Cabral Gold situation, in brief

1. Cabral Gold and its two-phase strategy

   1. The mechanism in detail

   2. The four arguments for the sequencing

   3. The legal structure

   4. Why the mining asset is rare

2. The Tapajós and those 2 million unexplained alluvial ounces

   1. A province, in the strong sense

   2. The gold rush, and the exploration logic it imposes

   3. Tocantinzinho, the proof next door

   4. The ground: a single owner of 306 km²

   5. The Tocantinzinho Trend Deformation Zone (TTDZ)

3. The geology: a body with two states

   1. How the gold arrived there, and the value of the answer to the “how”

      1. The pump and the egg

      2. The verdict of the Qualified Person

      3. The ceiling and the free hypothesis

   2. Where the gold sits, and how you look for it

   3. The weathering profile layer by layer, and its economics

      1. The blanket

      2. The saprolite

      3. The saprock

      4. The fresh rock

   4. The inventory of Cuiú Cuiú

      1. The official total: 1,138 koz, 2022 vintage

      2. Deposit by deposit

   5. The two grade populations

   6. Why 75-80% of the ounces are in the hard rock

   7. The 2022 correction

4. Phase 1: the cash machine, as the PFS describes it

   1. Heap leaching

   2. The numbers behind this leach

   3. The front-loading of the cash flows

   4. Where the build stands, and the path toward first pour

5. Metallurgy, the eternal problem explained (relatively) simply

   1. Two numbers vs reality

   2. Percolation: where the project already had its bad news

   3. Cement: the only reagent that can miss the budget

   4. The clays and the missing data

   5. Water: the underestimated risk

      1. Four tools, and a water balance in the future tense

      2. Three choices, in increasing order of consequence

      3. The first wet season

   6. The samples vs reality

   7. Mercury, courtesy of the Tapajós

   8. The engineer’s verdict, and what it costs in ounces

6. Phase 2: the exploration, and the district thesis

   1. A little vocabulary just in case

   2. The official inventory: 1,197,800 ounces, of which 75% counted at a price from four years ago

   3. The pending update that can change everything

   4. The drilling machine, and how to read its results

   5. The pipeline, tier by tier

      1. Tier A, the defined resource

      2. Tier B, the deposits already modeled

      3. Tier C, the targets drilled without resource

      4. Tier D, the undrilled

   6. The track record, and the ROIC of the drill

   7. The Phase 2 PEA: the second Lassonde cycle

   8. So, what growth is probable?

7. The financing: a gold loan as a sword of Damocles

   1. The cash position

   2. The terms of the gold loan

   3. The related party

   4. The history of the dilution

   5. Royalties and taxes

8. The people

   1. Carter: the honest scorecard

   2. The bench: where the execution thesis really plays out

   3. Who owns

   4. Who buys

9. Brazil, Pará, and the permits

   1. Two agencies, plus a veto

   2. Where Cabral is, precisely

   3. The GU regime

   4. The real delays measured on the neighbors

   5. The water, again: the permit missing from the public file

   6. The four remaining layers: fiscal, social, political, currency

      1. The fiscal

      2. The social

      3. The political

      4. The currency

10. Situating the price: Cabral against its peers, its model neighbor and the district buyers

    1. How to value a developer correctly

    2. Cabral’s multiples, recalculated properly

    3. The Tocantinzinho template

    4. Serabi: the mirror that settles the awkward question

       1. Serabi in four numbers

       2. Where the margin of safety lives

       3. The multiple, bounded by Serabi and Aura

    5. The takeout

11. Gold

    1. The spot

    2. The model’s deck

    3. The golden lever

12. The valuation

    1. The frame: a NAV built up by phase

    2. The valuations in numbers, case by case

    3. The quarter-by-quarter self-financing test

    4. The drilling ROIC and the impact of the dilution

    5. The sensitivities, and the hierarchy of the variables

    6. My strategy and disclosure

13. The risks with their test + signal + mitigant

    1. Throughput capped from the gold-loan repayment (the only thesis breaker)

    2. Commissioning metallurgy

    3. Financing/dilution outside the capped sequence

    4. Study risk (the reset of the denominators)

    5. Gold

    6. Governance / related party

    7. First-time execution / key men

    8. ESG / social / Amazon

    9. All the other usual risks

14. Catalysts, a loaded calendar

15. Conclusion

0. The Cabral Gold situation, in brief

The setup comes down to four points:

1. Cabral owns 100% of Cuiú Cuiú: 306 km² in a single contiguous block over the biggest historical gold-panning camp in the Tapajós (Pará, Brazil). About 2 Moz of gold came out of the camp’s rivers, and to this day nobody knows where those 2 Moz came from. Wild guess: somewhere deep under Cuiú Cuiú? Bonus: 25 km away sits Tocantinzinho, Brazil’s third-biggest gold mine, built on budget. Second bonus: Cabral’s CEO co-discovered Tocantinzinho.

2. Cabral’s first production, expected Q4-26: a small mine sitting in the soft surface layer (the “oxide”), 128,903 ounces of reserve, US$37.7M of CapEx already funded by a gold loan, construction about 85% done. Design AISC (all-in sustaining cost): about US$1,210 an ounce versus gold at about US$4,000.

3. This mine already has a reason to exist: generate the cash that will drill the rest of the district (and find the source of those 2 Moz recovered from the rivers) without a dilutive trip back to the market. 75-80% of the ounces discovered so far sleep in hard rock below the surface, frozen in a 2022 estimate run at $1,800 gold, plus some fifty targets never drilled. Resource update late 2026, first hard-rock study Q2-27.

4. The market is pricing the small mine: at today’s gold price, the discounted value of the reserve alone covers most of the market cap. The whole district goes for tens of millions.

The bet plays out on three points.

1. Execution. First build for a CEO who’s spent his life making discoveries. Against that: a job site on time and on budget with more than 90% of costs locked in, and a construction manager who’s already built two Brazilian heap leaches, including the country’s biggest at the time.

2. Permitted throughput. The plant is sized for 1 Mt/yr at full tilt. Current permits cap production at a third of that (soon half, most likely). The catch: the gold loan repays in a fixed number of physical ounces starting March 2027. At full throughput, everything else in the story gets paid out of the gold that isn’t going to the lender. If permits throttle production too hard, the self-funding breaks down completely. Production isn’t at full rate yet, so today’s permits don’t bind yet, but a few desks inside one Pará agency can kill this thesis.

3. The denominators. Three quarters of the official resource dates from a world of $1,800 gold, from before 35,000 m of drilling and four named discoveries. The late-2026 update re-counts everything: four years of drilling already paid for and published hole by hole, and not a single ounce of it counted yet. A big variable with big re-rate potential. Or not.

Why now? The window between first pour and the resource update is where the information is richest and the re-rating barely under way. It’s a way to play the two best moments on the Lassonde curve at once: before production and before a discovery. As for timing, four announcements were running late. The first (first ore) landed July 9, while I was writing these lines. Three are still pending: two start-up permits plus one resource upgrade to Measured. It’s a matter of weeks.

What can break, in one line: the throughput permit combined with repayment in ounces (the only thesis breaker), start-up metallurgy (margin erosion, softened or offset by the gold price), and two upcoming studies that could make my scenarios obsolete. The full list, with the signals to watch and the consequences, is in Part 13.

To sum up: metallurgy, construction, drilling and discoveries. Forget market orders after a big-number drill result hits the wire. The edge here comes from understanding the setup and its (many) subtleties in detail.

Adding gold miner exposure is a priority for my portfolio, and this company has become my lead candidate. I’m not a shareholder at the moment, but I have cash parked and waiting on a few things before I buy. The criteria and the reasoning are spelled out below, of course.

(A word of advice: get yourself a hot drink and a source of glucose; your neurons are going to need both.)

Subscribe now

1. Cabral Gold and its two-phase strategy

Cabral Gold is a pre-revenue gold developer. Its latest bottom line is a C$8.2M loss for the quarter ended March 31, 2026 (expensed drilling + stock comp + revaluation entries).

If the schedule holds, the company pours its first doré (the raw gold-silver bar) in Q4-26 at Cuiú Cuiú, smack in the middle of the Amazon. The mine producing it is the most rudimentary, cheapest thing the industry builds: an oxide leach mine. Think of it as making coffee with a filter:

• Oxide -> the top of the deposit (the first few tens of meters) rotted in place for millions of years under a tropical climate, turning into a clay (saprolite, to its friends) you can literally mine with a shovel.

• Leaching -> you pile it all on an impermeable liner (the filter), irrigate it with a cyanide solution that dissolves the gold as it percolates down, then collect the gold-loaded solution at the bottom (the hot coffee, ready to drink). Finally, that coffee runs through activated carbon: a sponge for dissolved gold that you just wring out before pouring bars.

The mine plan calls for 128,903 ounces of Probable reserve at 0.65 g/t, with an AISC (all-in sustaining cost, the full cost of producing an ounce) of US$1,210/oz. With spot gold around US$4,000, every ounce coming off the heap leaves about US$2,800 of margin. On start-up production of 20-25 koz annualized, the CEO is talking cash flow on the order of US$70M a year at run rate. Even after a severe haircut (always doubt management’s estimates), the little plant quickly spits out more cash than the company has ever raised in a year.

Discount that stream over the 6.2-year mine life, note that even at a 5% rate you land too close to the market cap, and on to the next name. Anyone who reasoned that way just missed one of the rarest setups in mining. What we’ve just described is only Phase 1 of a bigger plan: build a district. The 6.2-year mine plan is just a way to fund Phase 2: intensive exploration and potential production at a much bigger scale.

The real thesis lives under that big pile of saprolite: about 75-80% of the ounces already discovered at Cuiú Cuiú sleep in the fresh, dense rock that has been resting under the clay for millions of years. Potentially far more, if the 2 Moz recovered from the rivers really did come from the deep rocks of Cuiú Cuiú.

To sum up: Phase 1 mines the skin of the deposit, and the cash it throws off pays to X-ray the skeleton (PEA targeted Q2-27).

1.1 The mechanism in detail

This “self-funding model” is nothing new. It’s even a staple of junior presentations, very (too) often undeservedly. Let’s check its viability for Cabral.

On the sources side: about C$50M of cash in early April 2026 after the C$20M bought deal, plus the margin on Phase 1 gold once production starts.

On the uses side:

• the remaining mine CapEx (about 15% left of US$37.7M, over 90% committed)

• US$5-8M a year of drilling for six rigs and 25-30,000 m every six months

• sustaining capital (US$8M of sustaining over the life of mine)

• the gold-loan repayment: 39 kg of gold per quarter (about 5,000 oz/yr), starting March 31, 2027.

At full throughput (1 Mtpa), the mechanics close the loop with room to spare: loan repayment absorbs about 27% of 2027-2028 payable ounces (31% with interest), drilling is comfortably covered by what’s left, and cash piles up fast. Great. Capped at 500 kt/yr, the machine jams: the 39 kg due each quarter, plus interest, eat two thirds of the ounces produced. Exploration stops being automatically free and would probably require dilution or more loans.

On paper it works. Now for the part that isn’t on paper.

1.2 The four arguments for the sequencing

First off, why this plan? Why not think big, drill the district for five years and build afterwards?

For four good reasons.

One: the initial CapEx is tiny. US$37.7M, funded by a gold loan and one raise, with no massive dilution. That’s about as close to ideal as developer financing gets. The flip side: it makes the whole thing look like a small six-year mine and hides the district. For a potential shareholder who does his homework, that’s a plus.

Two: mining the oxide is pre-stripping the hard rock. The district’s war chest sits at depth. Stripping the top and treating it as waste is CapEx thrown away. By mining the gold in the surface layers, you cut the cost of going after the future gold in the buried hard rock, by cutting the strip ratio (the tonnes of waste moved per tonne of ore). It’s a win-win.

Three: the operating platform. Cast the net wide: camp, road, grid, laboratory, Brazilian team, a production and compliance track record, etc., everything that makes Phase 2 cheaper to finance and faster to permit.

Four, and not the least of them: the cash self-funds exploration without dilution, provided the permits keep pace. If they don’t, it still shrinks the dilution needed, and the earlier points still hold.

1.3 The legal structure

Cabral Gold Inc. is incorporated in British Columbia and owns 100% of Magellan Minerais Prospecção Geológica Ltda (Magellan Minerals), the Brazilian subsidiary that carries the 30k ha of Cuiú Cuiú (an intermediate holding company, Cabral Gold B.C., sits between the two). But the lineage is older than Cabral Gold. Magellan Minerals acquired the ground in 2005, discovered Central and MG (two of the district’s discoveries), hibernated from 2012 to 2016, then got rehoused into Cabral through the RTO of a shell (San Angelo Oil) in 2017. Alan Carter (the CEO) has been at the wheel of Magellan since at least 2012, and at Cabral’s since day one.

One subtlety: the gold loan holds a first lien over those subsidiaries. Put differently, the creditor, who is also the biggest shareholder, sits on the entire district, not just the plant.

1.4 Why the mining asset is rare

Every mining story claims a rare asset. Here, I can describe the rarity factually:

• the biggest historical placer camp in the Tapajós (placer: gold naturally eroded out of the hard rock, then sorted and concentrated in river gravel), at about 2 Moz alluvial;

• consolidated into a single block of about 306 km² under a single operator (the actual ownership is more complex, we’ll come back to it);

• about 25 km from one of Brazil’s biggest producing gold mines (Tocantinzinho, we’ll come back to that too);

• holding the camp’s only 43-101 resource (the Canadian standard for mining estimates) and,

• since March 2026, the first mining Licença Prévia (the hardest regulatory wall in Brazilian permitting) ever granted on this camp in the deposit’s forty-year history.

On the balance sheet, the district simply doesn’t show up: the C$74.3M of assets is mostly financing cash (C$31.9M), a build being capitalized, and exploration expensed away over the years. Half a century of garimpeiro rush (Brazil’s artisanal gold mining) and what that rush implies about the gold still in the ground: zero accounting dollars. In defense of the accounting rules, how would you even book that? It’s the market’s job, and whether the market has done that job is the whole subject of the sections ahead.

First, the ground. A district thesis is judged on the district.

2. The Tapajós and those 2 million unexplained alluvial ounces

2.1 A province, in the strong sense

To say that “gold province” is an overused term is an understatement. Allow me, then, to justify it.

The story begins with a collision: two continental plates smashing into each other for tens of millions of years. The crust folds until it breaks, letting fluids charged with gold, up from the depths, spread along the fractures. This work site is called an orogeny. When it has worked hard enough for long enough, it leaves a whole province behind it: a region where the structures are measured at the scale of the earth’s crust, with fractures that follow each other over hundreds of kilometers, with, among them, dozens of deposits that share the same origin and the same plumbing. Consequence: the discoveries gather there, and the method that found the first deposit often finds the seventh.

It’s a very closed club: the Tintina in the Yukon, the Abitibi in Québec or the West African Birimian. The Brazilian Tapajós belongs to this category, and it is very probably its least-drilled member.

To understand why, we need a little geography. South America has a core: a basement that runs from Bolivia to Venezuela and that hasn’t moved for about two billion years. This is what we call a craton (Amazonian, in this case): the ancient and rigid heart of a continent, which has finished deforming, and which is stable enough to conserve what time had the decency to deposit there. The Tapajós Mineral Province occupies the middle of this core, on the block called Tapajós-Parima. 2B years ago, this crust was the neighbor of West Africa, close to the Birimian mentioned just above. The Tapajós is therefore a direct cousin of the Birimian, the basement that hosts a good part of African gold production.

Why, then, does a cousin of the Birimian remain so little drilled? Three good reasons:

• The Tapajós rock is very shy. It does not show itself in public (a geologist would say it doesn’t outcrop). Elsewhere, a prospector walks the ground, breaks an outcrop with the hammer and follows a vein with his eyes. Here, the entire basement has rotted in place into a soft clay (yes, the saprolite again), itself covered by sediments over time and erosion. Consequence, surface mapping “provides limited insight”, and even the trenches show nothing but rot. To summarize: 0 outcrop = classic prospecting useless = geophysics + meters of drilling = spending capital.

• Access. Jungle, rivers, all of it in a capricious climate. Nothing surprising in there being no road network suited to facilitate drilling and discoveries.

• The permitting regime. We’re talking years, even more than a decade, for a complete cycle: Cabral’s LP (Licença Prévia), the first of the three Brazilian environmental licenses, took about five years of processing.

Add on top of that an exclusively artisanal production (the alluvial gold) and the paradox becomes evident: part of the gold was already accessible enough, and it wasn’t worth the cost of spending all that capital to go looking for the potentially big treasure below.

Today, this potential treasure is barely beginning to be X-rayed, and it holds one of the central roles in this thesis.

2.2 The gold rush, and the exploration logic it imposes

The numbers of the garimpo rush, the gold recovered from the rivers (once again, synonymous with a deposit nearby), have a few specificities that need to be explained.

Officially recorded production: about 159 tonnes, about 5.1 Moz. Those who know how the region functions know that this is necessarily underestimated, a large part of this gold never passed through the official channels. The ANM itself (the federal mining agency) estimates the real number between 30 and 50 Moz. The serious estimates of the real number are closer to 20-30 Moz. At the peak, more than 250,000 garimpeiros were together pulling out more than a million ounces per year. Today about 30,000 of them remain, for about 200 koz/yr. The gap between 5 and 30 Moz will never close, informal gold will never enter the registers. No matter. What counts is that all this gold was gathered in river gravels. Now, the gold of a river always comes from elsewhere. Erosion tore it from hard-rock veins somewhere upstream, then the current sorted it and concentrated it for millions of years. And nobody has yet found the bulk of these veins.

Cuiú Cuiú is the center of gravity of this story: the biggest placer camp of the whole Tapajós (I remind you, about 2 Moz alluvial according to the ANM estimate). To give context, see it as a village of about 5,000 inhabitants put up in the middle of the 1970s, with daily air links well into the 1990s. But it’s another context that interests us here. Let’s calibrate a little. The Tocantinzinho placer, before 2 Moz were discovered there in the buried rock, weighed about 200 koz. So:

• 200 koz of Tocantinzinho placer led to a 2 Moz mine.

• 2 Moz of placer for Cuiú Cuiú, and the current estimates of gold in the rock cap out at 1.2 Moz.

Compute the potential ratio versus Tocantinzinho yourself.

116,000 m of drilling and two decades later, the key sentence of the whole case fits in one line from Carter: the bulk of the placer gold of Cuiú Cuiú remains “still unexplained” (March 16, 2026). The known deposits are very far from explaining the origin of those 2 Moz recovered from the rivers.

This is proof of nothing. Geology can be very capricious. But the argument is all the same very directional. And it’s not over. The gold-in-soil anomaly is about 7 km long at Cuiú Cuiú, against about 1 km at Tocantinzinho before its discovery. An argument, also, very directional. Count about ten boulder fields at surface (mineralized blocks that erosion detached from a nearby vein never found), some grading 74 to 92 g/t on average (the Alonso field at 91.7). Nothing here is proof of ounces. But they say where to look, and above all, that the hunt is far from over.

2.3 Tocantinzinho, the proof next door

Some fifteen kilometers southeast of the property, on the same structure, the third-biggest gold mine of Brazil (according to G Mining) has been running since September 2024, about 25 km deposit to deposit. See Tocantinzinho (developed by G Mining) as exhibit A of the jurisdiction case. It will be a comp all along this write-up, so it’s worth the trouble of detailing it here.

A few facts on the mine to start:

• P+P reserves (Proven plus Probable, the inventory committed to a mine plan): about 48.7 Mt grading 1.31 g/t, about 2.04 Moz.

• CapEx of the feasibility study: US$458M, delivered within budget (uncommon enough to be mentioned) with US$447M spent + US$11M committed at 93% completion.

• First pour July 2024, with first commercial production September 2024, also on time.

• First full year: 171,871 oz (2025) at an AISC of US$1,155/oz.

• On the financing side it stings a little, however. G Mining signed a stream with Franco-Nevada: US$250M received upfront, in exchange for the right for Franco-Nevada to buy 12.5% of production while paying only 20% of the market price, up to 300 koz delivered, then 7.5% after. Just reading it stings. We’ll talk about it again.

The market rewarded this whole journey: about C$196M of market cap at the end of 2021 vs about C$9.24B today. Of course, the raw number mixes re-rating + dilution. To be more precise: x3.9 in the single year of the developer-to-producer transition, and about x8.8 per share in 26 months.

Tocantinzinho cleared the forest path for Cuiú Cuiú.

It proved everything an investor can demand of a jurisdiction, step by step. A Pará deposit, on this structure, can obtain its three environmental licenses (the LP of the concept in 2012, the LI to build in 2017, the LO to operate in 2024; twelve years at the counter...). It can then finance itself, get built without exploding the budget, then produce what the management had promised. And at the end, the market pays. In short, the story can materialize.

It also improves the quality of the case itself. A jurisdiction that has already said yes has even more chance of saying yes, a build precedent to reproduce, infrastructure already financed (the road that serves Tocantinzinho passes through the region) and a capital market that has already touched the Tapajós. In short, the story has more chance of materializing.

What Tocantinzinho does not prove: that Cuiú Cuiú contains 2 mineable Moz. Or even that it holds far more than Tocantinzinho itself. Geology is not necessarily transitive.

The “irony” of the story: it’s Cabral’s CEO and his team (who in part followed him to Cabral) who participated in the discovery of Tocantinzinho some twenty years ago. If someone can transform all of this into a district, it’s him. As for building it, that’s more complicated. We’ll come back to it.

2.4 The ground: a single owner of 306 km²

The package comes to 30,563 ha, about 306 km². Among them, eighteen mining titles.

To read the title portfolio, one line of Brazilian administration suffices: an exploration permit gives the right to drill, but never to mine. Mining demands a mining concession, whose processing takes years. Between exploration and mining there exists a jump seat: the GU (Guia de Utilização), a trial-mining permit (capped, of course) attached to a concession application under processing. The objective is to give miners a permit that allows production while waiting for the final production permit to arrive (see it as a lubricant of a system a little too complex).

Cabral’s eighteen titles also line up on this scale: fifteen exploration permits in the pocket plus a sixteenth in application (enough to drill the district), and two concession applications under processing, each carrying a GU. It’s these two GUs that keep Phase 1 running: the Q4 2026 mine starts on the jump seats, while the real seats get processed, slowly.

The ownership. Everything is 100% Cabral’s, via Magellan Minerais, the Brazilian subsidiary. Of the 30,563 ha, 11,980 are in Magellan’s name, the direct title. The remaining 18,583 are said to be “in indirect interest”. It’s a rather unusual arrangement, which reflects in large part the social risk. These 18,583 ha are registered in the name of the cooperative of the local garimpeiros. The arrangement is deliberate. The PLG (the small-mine permit) is a regime reserved for artisanal miners that an industrial company cannot hold. By transferring eleven of its eighteen titles to the cooperative, Cabral gave the garimpeiros a legal framework to work on the property, and reserved for itself by contract the only two things that count: the right to explore this ground, and a buyback option on the titles for the day of the switch to the industrial mine. The framework agreement dates from 2006, renewed in 2017. To summarize: one layer of legal complexity (ownership structure) to face another layer of legal complexity (the permits).

You also have to put this back in the Brazilian cultural context. It’s not rare to see headlines pitting the small artisanal miners against the big mining explorers/producers. Here, the artisanal miners are directly in the title structure, and above all with the incentives needed to want this project to succeed. That doesn’t protect from a conflict, all the more since the final compensation of these PLGs is quantified nowhere. But 20 years of documented coexistence on the biggest garimpeiro camp of the Tapajós is an asset that very few Amazonian juniors can show (and the federal hardening against illegal garimpo makes it even more precious, but we’ll come back to that later).

2.5 The Tocantinzinho Trend Deformation Zone (TTDZ)

Now connect the dots on the map. The crust broke here along a NW-SE corridor of several hundred kilometers, the “Tocantinzinho Trend Deformation Zone” (TTDZ), which G Mining calls the “Chico Torres Megashear” (renaming doesn’t always mean making things simpler).

Remember, the ground does not show its rock. It’s therefore geophysics that must serve as the map. Fly over the jungle with a magnetometer (the aeromag) and the corridor appears as a linear band of weakened magnetic field, a “low” in the jargon (why low for gold, we’ll see later). This corridor threads Cuiú Cuiú, Tocantinzinho, Mamoal, Mato Velho like beads, then Serabi’s underground mines (Palito and Coringa) and São Jorge. It crosses Cabral’s property on the diagonal.

It remains a large-scale structure and it says nothing at all about where to drill, and, too bad if I repeat myself, but a structure does not make a viable mine. On the other hand, a structure that already carries one of the biggest gold mines of Brazil, with two underground mines and a half-dozen advanced deposits, and that crosses the biggest unexplained placer camp of the province, now that gives a lot of theoretical value to Cabral.

But it remains a large-scale analysis. A mine implies discoveries, which implies drilling, which implies knowing where to drill. But once again, the rock does not outcrop. And unfortunately geology does not come to save us on this and it remains the biggest problem... no, I’m kidding. Geology gives us tools that allow us to “see” where the gold is and how it organizes itself.

Be attentive, because it’s here that the potential Cuiú Cuiú value not perceived by the market takes shape.

3. The geology: a body with two states

This section is the longest of the report after the valuation, and rightly so. Everything that follows, the metallurgy, the exploration, the model, rests on a mental image that the market, in my opinion, does not have.

See this section as an investment. Take the time to understand it well and the rest of the report will read twice as fast afterward.

3.1 How the gold arrived there, and the value of the answer to the “how”

Two families of models compete to explain the Tapajós deposits. I promise, the nuance is less academic than it looks.

3.1.1 The pump and the egg

I hope this image is worth a thousand words, but it will not eliminate them.

The orogenic model. See it as a metal pump activated by earthquakes (Part 2). During the collision, the rocks compressed at depth release their water charged with dissolved gold. At each big earthquake, the large faults open at once. Voids open up, the pressure drops abruptly and whatever can rush in does, here gold-laden liquid (among other things). Then, the structure stabilizes, the fault closes again, the pressure climbs back, and the cycle starts again. Result: vertical mineralized bodies that can exceed a kilometer of depth. Summary: orogenic = vertical seismic gold pump. Once found, you mostly just follow it down.

The intrusion-related model (IRGS). Here it will take a little more imagination. Imagine an enormous bubble of magma that rises in the crust. After a while, it stops, and cools in place for a million years. This big bubble of cooling magma is called a pluton. The cooling is obviously not uniform, it goes from the edge toward the heart. Now, the crystals that form first during the cooling are very capricious, they accept neither water nor metals in their structure. Everything that does not crystallize therefore concentrates in the remaining liquid, like a sauce reducing. The more the external layers cool, the more this metal-filled liquid concentrates in the internal layers, and the more the pressure rises. After a while, the pressure is such that the metallic juice is expelled into the fissures of the pluton’s roof (think of an egg in its shell that has cooked too long and whose white has come out slightly through the top of the shell). And there the story stops. The gold and the metals that accompany it stay put, hence the three signatures of the model: a bismuth-tungsten-molybdenum mix (the metals that travel with gold when the source is magmatic and that also concentrate in the liquid). The grades are low but consistent, which makes it a signature. The vertical extension is however bounded compared to orogeny (the gold never descends lower than the pluton that expelled it). Summary: IRGS = an egg that cooked enough to break the eggshell, with cooked egg white that intruded because of the pressure but stayed close. The difference to remember versus the orogenic model: the seismic pump reactivates as long as the collision lasts, while a pluton exudes only once.

The verdict of the Qualified Person

Right. And Cabral in all this?

Let’s ask the QP (Qualified Person, the accredited geologist who puts his license on the line for the estimate)1. The verdict: orogenic (it hasn’t changed from 2022 to 2025). For him, it’s the seismic-pump scenario that switched on during the Trans-Amazonian orogeny (about 2 billion years ago, when West Africa was still its neighbor). The deposits would have formed between 6 and 12 km under the surface, and two billion years of erosion had the courtesy of bringing them back up within reach of the drill.

I could say “trust him”, but this global mental picture is precisely what the market doesn’t have. So I owe you a full justification.

Here, we have five observations in support:

• The rock that hosts the gold cooked at depth, and its degree of cooking can be measured (that’s the regional metamorphism): consistent with a formation several kilometers under the surface (therefore, orogenic).

• The gold cuts across several generations of granites. If a pluton were its father, the gold would stay glued to this pluton, like the white around its egg. That does not seem to be the case.

• Almost all the zones carry high-grade shear veins. Translation: veins born in moving faults (the territory of the pump).

• The veins are banded and laminated layer by layer. The vein quartz was laid down in stacked layers like tree rings, and the rock around it was broken, re-welded, then re-broken and re-welded again and again (”polyphase breccias” in the jargon). If you have followed: each layer is a stroke of the pump and the rock kept the count.

• Last observation, the most disputed: almost no trace of the magmatic model. There is indeed molybdenite and scheelite at Machichie (the indicator minerals of molybdenum and tungsten that you find in the metallic juice with the gold in IRGS), but neither correlates with grade. Roughly, where these traces exist, the gold does not follow them.

So much for the QP’s reading. But reading rock is not quite an exact science for a human. The opposite camp reads these same minerals completely differently, and to be fair, they have a point.

First, they do not agree on the interpretation of the presence of this molybdenite/scheelite. The mineralization is entirely housed in granites, exactly where the egg white that came out of the shell would go. What’s more, scheelite is one of the standard IRGS indicators in Hart & Goldfarb (the authors of the model). Nothing decisive either.

What do the neighbors think of it? G Mining classes Tocantinzinho as IRGS while Biondi and his co-authors argue there’s nothing to rule out reading it as orogenic. To summarize, the debate still runs across the whole province, and it’s far from settled.

The ceiling and the free hypothesis

What does it change for Cuiú Cuiú? The ceiling.

A pump promises a relatively vertical continuity: similar camps get mined at 1, 2, even 3 km under the surface. Currently, the deepest hole of Cuiú Cuiú stops at 430 m (it was still mineralized when the drilling stopped, “open” in the jargon). If it is indeed a pump, the current 1.2 Moz would only be the upper floor. An egg makes a very different promise: mass around the pluton, and on the periphery potentially a copper-molybdenum bonus (porphyry: the class of giant low-grade copper-gold deposits).

Now, Cuiú Cuiú conveniently ticks boxes on both sides. The pump has its five observations but the egg has its own clues too. It’s just that none has ever been drilled:

• At Central and PDM (two deposits of the district), the drill intersected diorite porphyry dykes (blades of magma injected into the host rock), the kind of fingers that escape from a bigger mass that stayed below (our famous pluton).

• North of Machichie, coarse molybdenite and scheelite concentrate around a protrusion of granite that pierces the basement.

• To sow even more doubt, this protrusion sits on a magnetic low (which could be the signature of a buried pluton).

If I translate: something magmatic probably sleeps under the north corner of the district, but nobody has ever gone to verify.

To hammer the point home: 2.5 km from Tocantinzinho, a target named Santa Patricia shows an 8 km copper-in-soil anomaly. Serabi is already drilling copper-gold-molybdenum porphyries on its Matilda target. The whole province is starting to show signs of buried eggs, and if these egg pieces exist, they’d add to the thesis (extra metal on top of the gold already counted). Obviously I pay nothing for this porphyry hypothesis, and neither does the model.

But know that the possibility exists, that it is free, on ground already paid for. For the mine under construction, it changes almost nothing. The geologists’ debate fixes the ceiling of the district but the mine runs the same either way, whoever wins.

3.2 Where the gold sits, and how you look for it

Very interesting, the geological debates, but where to look for the gold in these 306 km² then? The answer comes down to two simple ideas, plus one tool.

Idea one: the gold follows the pipes, and the pipes have a predictable geometry. The magnetic map of the district says everything at a glance.

The trunk jumps right out at you. That dotted line crossing the package on the diagonal corresponds to the TTDZ of Part 2. An attentive eye will notice it follows a deep-blue band (an extinguished magnetic signal, more on that in a second). Then come the beads, those yellow points (the deposits) that cluster along this thread.

You will notice that these deposits are slightly on the periphery of the trunk. When two blocks of crust slide along such a trunk, secondary breaks open at an angle, like the oblique tears you get when you slide two plates with dough between them. These are the branches (the “splays”, in the jargon), and here they are oriented E-W. It’s these branches that carry deposits like MG, Machichie and Jerimum Cima. The magnetic zoom on MG and Machichie clearly shows these extinguished (blue) bands that follow the E-W orientation.

If I translate all that into investor language: the deposits line up along two directions, each predicting the other. Find the line, and it tells you where to drill next. Shorter still: the gold follows two known lines, and you can see them on a map.

Idea two: the more the rock has suffered, the richer it is. It’s granite, and granite is too rigid to fold, so it breaks. The more it broke, then re-broke (the breccias of 3.1), the more often the fluid passed, and the more that transformed the rock at depth (i.e., gently depositing gold). We need extremely complex chemistry, so hold on: gold stays dissolved as long as it does not meet iron. When the fluid meets iron, it turns it into pyrite, and the gold drops out with it. There, I hope you were holding on well. Let’s now apply this very complex chemistry.

Central (a deposit of Cuiú Cuiú) is wrapped in large red halos, tinted by a barren iron oxide. See it as the pantry. The iron there was consumed, and the gold dropped out next to it as grey pyrite. Same logic for the blades of dark rock that cross the zones (the dykes), barren themselves, but rich in iron, where their edges trap the gold. Summary: look for the broken rock, and look for the iron. The gold is right next to it.

All that remains is to find the broken rock and the iron.

It’s time to make all this geology pay:

1. The basement granite (everything that is under the saprolite, therefore what interests us for Phase 2) contains magnetite, a mineral that is... magnetic. Fresh rock therefore emits a measurable magnetic signal.

2. The fluid that deposited the gold also destroyed the magnetite on its way (during the transformation into pyrite).

3. Therefore everywhere the gold passed, the rock lost its magnetism.

4. Fly over the jungle with a magnetometer and the gold corridors show up on their own in “extinguished” bands on the map (those deep-blue bands).

5. Icing on the cake: the more extinguished the signal, the more the rock was transformed, and therefore the higher the grade (it’s more a correlation than a direct causation).

So far we’ve reasoned at the scale of the whole structure. But this tool offers much finer granularity. Since 2025, Cabral flies this magnetometer under a drone (therefore finer than the old airborne survey). The drone has already revealed two unknown faults that frame the Central-Mutum-PDM corridor.

The rock does not outcrop, so nothing is visible to the naked eye from the surface. But thanks to the geological structure itself, the magnetometer replaces the hammer and the eye of the prospector. That’s why the targets number in the dozens. Cabral literally owns a machine for targeting the 306 km² of the package. Thank you James Maxwell.

3.3 The weathering profile layer by layer, and its economics

Now for the pure vertical. Here is the vertical cross-section that governs the project (it’s an example from the PDM deposit, but fairly representative).

Remember, the tropical climate rotted the top of the deposit in place and this rot changes character as you go down. Let’s follow the drill down through four layers.

3.3.1 The blanket

See it as a rug thrown at random over a parquet floor. The rug is made of whatever’s around it, literally. Slope debris, sediments, even the old tailings of not-very-conscientious artisanal miners. A far-from-uniform rug, 0 to 40 m thick (much less on average).

The pattern of the rug says nothing at all about the floorboards below, there is almost no continuity between the two (”unconformable”, say the geologists). It’s even a bit of a liar, frankly: the rug widens to more than five times the width of the floorboard zone below. So the rug in no way implies floorboards below. Let me repeat: the rug does not imply the presence of floorboards below.

For the material itself, it is not very dense: about 1.6 t/m³ in the model for the blanket (the bulk density measured in columns even drops toward 1.0), against more than 2.7 for the granite below. About 0.35 to 0.5 g/t of gold you can mine with a shovel. It’s the surface so zero waste to move, and it leaches very well (cyanide that flows along the heaps to recover the gold at the bottom, filtered coffee) with 88-91% design recovery at MG and Central.

Only one nasty by-product turned up: pockets of lake clay. Nasty because too fine to stack without plugging the heap (the cardinal sin of a heap leach, we’ll obviously come back to it). The mine plan sends these localized pockets directly to waste, gold included (the cost is trivial, it’s the right choice).

3.3.2 The saprolite

The heart of Phase 1, down to 60 to 80 m depending on the deposit.

These are the floorboards. They were there before the rug, but they too have rotted. The hard minerals became clayey, the sulfides became iron oxides (that again), but the rock still kept its overall geometry (same direction, same dip, same thicknesses). See it as a photocopy of the hard body below, but softer. The CEO insists on it: the mud “has the same gold content as the underlying rock”. It also keeps the grade, or even improves it. I use the active voice on purpose: weathering itself removes part of the barren rock but leaves the gold in place (sometimes with a bonus of gold redeposited by surface waters). Here too a shovel will do, still no crushing-grinding.

But it remains clay. Make big heaps of it and it plugs. If it plugs, the cyanide does not flow and the gold stays stuck. The solution is therefore to knead the clay with cement before stacking, in order to glue the fines into permeable granules (agglomeration) which, in turn, are kind enough to let the cyanide (and the gold) through. Recovery varies a lot by deposit, from 90.5% at MG to 72.5% at Central. Careful, these are design recoveries; reality could be harsher, or kinder. We’ll come back to it at length in Part 5.

3.3.3 The saprock

It’s a transition layer between the saprolite and the hard rock below (the rock literally hardens as you go down). It’s not always present, and its thickness varies (a few meters up to 20 m at Central). Nothing economic in it under the current plan, it’s pure waste to move.

3.3.4 The fresh rock

At last.

Below 60-100 m. The density approaches or even exceeds 3. You therefore have to drill, blast, crush and grind. Once ground, the gold gets recovered in two ways. Half by gravity (the coarse gold separates out by density) and the rest by CIL (carbon-in-leach).

CIL is the heavy and industrial version of Phase 1’s cyanidation. We go from filtered coffee to filtered beer. You grind the rock into fine powder that you then stir into a cyanide solution in big tanks (beer). The laws of physics don’t change from one tank to the next, so the gold dissolves here too, and you throw activated carbon (the sponge) straight into the tank. The carbon captures the gold as it goes; all that’s left is to wring the sponge to recover the gold. Here it’s much faster and more efficient than Phase 1, with lab recovery rates of 90-97% (just a guide, not the design number). But we’re talking about a real plant; it’s much more expensive than a heap and a liner. That is Phase 2, the 2027 PEA and the 75-80% of oz already counted (which therefore excludes all the potential oz of the district thesis).

Two quick notes before we head back up to the surface.

Nature literally spent millions of years crushing and pre-grinding the top of the deposit. Phase 1’s US$37.7M of CapEx is just that free work, cashed in.

The second is a bonus hunting tool. The speed at which a rock rots depends on what it contains, in our case, pyrite. When pyrite meets rainwater and oxygen, it oxidizes and makes sulfuric acid. This acid attacks the rock much harder than the rain alone, diminishing the density. So above a mineralized zone, the rot reaches deeper than normal. And the reverse is also true: a barren dyke without pyrite holds up better, and so the rotted layer thins there. You can therefore literally map the depth of the mud to guess where the mineralized zones sit below. To summarize, an abnormally deep saprolite is a clue to ore below. Even the mud does geophysics.

Subscribe now

3.4 The inventory of Cuiú Cuiú

3.4.1 The official total: 1,138 koz, 2022 vintage

We’ve built up enough theory to move to practice. Over to the deposits.

According to the 2022 MRE (Mineral Resource Estimate, the official resource estimate):

• 604.0 koz Indicated (21.6 Mt @ 0.87 g/t)

• + 534.5 koz Inferred (19.8 Mt @ 0.84 g/t)

• = about 1,138.5 koz over five modeled deposits, of which 232.9 koz in oxide and the remaining 905 koz in fresh rock.

Inferred: you extrapolate between the holes (up to 100 m beyond the last hole), these are resources you’re legally barred from booking as reserves.

Indicated: the extrapolation is more limited (50 m or less, here) and you’re allowed to hang engineering on it.

Above that again sits Measured, but Cuiú Cuiú doesn’t have a single ounce of it yet.

We were talking about 2 Moz out of the rivers, whose source in rock remains to be found. And here, we are not even at 1.2 Moz counting the oxide, without a single Measured ounce. Let’s square all this.

First, 2022 is old. Only the oxide of MG, Central and Machichie was re-estimated for the PFS. The primary, for its part, stayed frozen at 2022. So on top of being potentially underestimated relative to what the district can offer, the official estimates aren’t even up to date. Hang on to that detail.

Phase 1’s reserves are all oxide and all probable: MG 82,912 oz, Central 29,959 oz, Machichie 16,032 oz. 6,178 kt @ 0.65 g/t = 128,903 oz. And a global update, targeted for the end of 2026, that has to redo everything across six or seven deposits. So even the oxide resources aren’t up to date, and soon will be. It’s even one of the big potential short-term catalysts.

All of this obviously ignores every target not yet drilled. The district-scale thesis is a matter of years.

3.4.2 Deposit by deposit

Time for the guided tour, from ripest to greenest.

• MG (the flagship): the district’s template architecture (a lattice at roughly 1 g/t with rich veins inside). The 2026 infill (drilling between the existing holes) confirms the model in aggregate (”in line with expectations”, per the company, on 124 holes published, 41 still pending), with standouts far above. It’s MG that roughly pays the next five years, and its mining started July 9 (guided for end June, so one week late, immaterial).

• Central (the veteran): it’s the deposit where the garimpeiros scratched the surface the most, the one where the drill went deepest in the district (430 m, still mineralized), and the one that already carries the biggest hard-rock resource. It is in the mine plan and we already know that its saprolite is the worst-leaching ore of the project.

• Machichie Main (the MG clone): “remarkably similar”, according to management, with the egg clues of 3.1 and the most visible gold on the property. Its basement is drilled only to 150 m: the clone still has to be confirmed at depth.

• PDM (the 2026 surprise): its rug is growing, and above all the drill snagged Mutum, a never-tested basement zone in an almost untouched 2 km corridor (right on a fault the drone had just revealed, thank you James Maxwell).

• Jerimum de Baixo (the sleeper): counted in 2022, never drilled by Cabral since. The update will say what it’s worth, if it’s worth anything.

• Jerimum Cima (the headline magnet): the best hole ever drilled in the district, an open system, and zero official ounces. Its whole case hinges on the end-2026 maiden resource. Very probably a hit or miss.

• Machichie NE (the blind discovery): nothing at surface, with spectacular but scattered holes. But only six holes, at this stage it’s more an argument than a deposit.

• CN and Mutum (the targets): interpretation, clues, and drilling to come.

To summarize:

• 3 deposits pay Phase 1 for now,

• 4 others await their first official number end 2026,

• and discoveries that keep coming at a steady clip (Machichie NE 2019, Jerimum Cima 2024, Mutum 2026),

• and that could even accelerate if Phase 1’s cash flow arrives (more than 50 potential targets).

As for the speed, the cost, and the drilling, that’s for Part 6.

3.5 The two grade populations

On one side, rich veins at 87 g/t (no, it’s not a typo), on the other, an average grade of 0.65 g/t for the reserves.

First, the district contains two types of grade populations (like Tocantinzinho):

• a homogeneous background running at 1-2 g/t in fairly wide breccias (the MG envelope “averaging about 1 g/t”, Jerimum Cima’s 82.6 m @ 1.0);

• and thin but ultra-rich columns (for example, the 87 g/t).

Even averaging the two populations, you don’t land on 0.65. Two ingredients are missing. First, dilution: a shovel doesn’t mine with a scalpel; recovering the rich veins means taking the lean halo around them and a little waste at the edges (we explained it in Part 2). Then, and above all, the economics: at the pit plan’s gold price, even mud at 0.35 g/t pays its way and more. So the pit takes tonnes of it, deliberately, and each lean-but-profitable tonne that improves the economics drags the average down. Hence the 0.65.

This average is also (and above all) a schedule. The richest zones first (even diluted), coming out at about 1 g/t. The leanest wait for end of life. Result: about 0.97 g/t fed in year 1, 0.77 in year 2, then a slope that ends toward 0.45-0.65. For the why, it’s almost always the same reason: time value. A higher-margin dollar in year 1 is worth more than in year 6.

A small subtlety here: the nugget effect. Picture the gold in the rich veins as a raisin cake (as opposed to a plain sponge cake, nicely uniform if made by a conscientious pastry chef). Now, a lab assay only tastes a minuscule slice of the whole cake. Put another way, most slices probably miss the raisin, and the average of the assays understates the true grade of these veins. The QP spells it out, and the standouts of the 2026 pre-production infill at MG (far above the model) point the same way, even if the aggregate of the 125 holes comes out demurely “in line with expectations”.

3.6 Why 75-80% of the ounces are in the hard rock

The arithmetic first. In the 2022 MRE, the oxide carries 232.9 koz out of 1,138.5, or 20.5%. The rest, 79.5%, is fresh rock buried under the floorboards.

The 2024-25 oxide additions bring the split back toward the “about 75%” that management cites (logical: only the oxide has been re-counted since 2022, Phase 1 needed its reserves; the primary still awaits its turn). The asymmetry in proportion is purely anatomical: the oxide is the skin of the deposit, the hard bodies are its skeleton. A skin has a capped thickness (about 60 to 80 m). The skeleton, for its part, is vertical: it goes down as far as the drill has gone, 375 to 430 m depending on the deposit, and further still. For the same footprint, the skeleton’s volume dwarfs the skin’s, inherently. The sketch in 3.3 says all this at a glance.

This distribution is a kind of ceiling. The oxide is relatively easy to find and measure compared with the primary. Whereas for a volume of oxide already found, there’s almost certainly more primary still to discover. Each deep drilling campaign adds ounces to the primary. Phase 1 therefore funds the X-ray of a Phase 2 that’s still of unknown size.

3.7 The 2022 correction

I hope you remember that the presence of a rug does not necessarily imply the presence of floorboards below, because it’s time to tell the story of those who forgot it.

Until 2021, Cabral “lived” on inherited models (a 2011 estimate, reworked in 2018). Here’s the horror show, in four acts (squeamish geologists, look away):

• drill collars positioned by compass (with errors going up to 35 m!);

• basement grades projected up toward the surface through a saprolite that had never been drilled;

• the reverse, rug grades magically projected down into the floorboards;

• “default” densities, straight out of the manuals.

Management knew it had to rebuild the models and carried out a complete rebuild in 2022 (drill collars re-surveyed by GPS, saprolite drilled systematically, densities finally measured). The results were not kind: several basement zones in the model “did not exist”. The rug’s spread had created phantom ounces.

But it was for the best. The purge that killed these phantom ounces gave birth to new ones. The re-modeling made the Indicated jump 253% (to 604 koz) and isolated for the first time a leachable oxide resource of 15.4 Mt @ 0.47 (233 koz), which Carter said at the time “could be a game changer for early low-cost development”. He clearly had a nose for it. In four years these resources went through two PFS studies and the mine is under construction. Phase 1, even the whole plan, was literally born from an error correction.

I draw three things from this story:

• The origin of the two-phase strategy is geological first, financial second. Without the discovery of the rug-floorboard unconformity, nobody would have isolated the oxide, and there would be no Phase 1 (and therefore not this deep dive).

• Management discovered that its inherited resources were in part fictional. It wrote it in black and white and rebuilt on top. The timing surely played a part; notably, the update was demanded by the BCSC (the British Columbia market cop). But the purge itself, and what it turned up, is to their credit. As a potential shareholder doing his homework, I want to see exactly this: a model born from a methodological purge.

• The permanent caveat is that each new rug (PDM, Jerimum Cima) will carry the same ambiguity until its floorboards are drilled in their own right.

At Cuiú Cuiú, the root proves out hole by hole. And that’s exactly what Phase 1’s cash is meant to pay for.

4. Phase 1: the cash machine, as the PFS describes it

4.1 Heap leaching

We need to talk again about the filtered coffee and the beer. Quick recap, just in case.

Phase 1 (the filtered coffee of Part 1): you stack the agglomerated saprolite on a liner, you bombard it with cyanide that percolates and comes out charged with gold at the bottom. You run the whole thing through columns filled with carbon downstream of the heap (the sponge waits for the gold at the drain). The industry calls that carbon-in-column (CIC).

Phase 2 (the beer of 3.3.4): here you have tanks filled with ground slurry (the rock was too hard). You can’t filter it cleanly enough to run it through columns, so you throw carbon directly into the tank. You stir it so it captures the gold in place, then you recover it with a screen (the carbon grains are bigger than the ground rock). The industry calls that carbon-in-leach (CIL).

As a potential shareholder, what interests me is also what isn’t there. No mill, no agitation tanks, and above all no tailings dam (the toxic mud lake of conventional plants, an operational and permitting nightmare all by itself). For Cabral, we have a minimalist heap-leach plant that transforms 0.65 g/t ore into margin on one single condition: that the cyanide solution reaches all the way down.

Cabral has a few quirks that affect this flow:

• This clayey, wet saprolite. A simple mineral sizer that breaks up the clods under a 50 mm grid and that’s all. Nature already paid for the crusher (3.3).

• Despite that, the clay would plug the heap if you stacked it as-is (and we really do not want that). So everything goes through an agglomeration drum with 10-15 kg of cement per tonne to glue the fines into permeable granules.

• The norm in the industry is to have a permanent pad on which you stack layers of ore successively. You start with a first lift of about 10 m, you leach it then you leave it there. You add the next one on top, you leach and you let it flow. And you start again for years, sometimes ending up with heaps 150 m tall. The advantage is enormous: zero unloading, one single liner amortized over years, and the old lifts even keep yielding residual gold along the way. Cabral cannot do that. The heap would plug too fast. The solution was to have four cells on which you add ore (about 100k tonnes per cell), you let it leach 60 days (120 days with rinsing, unloading, etc.) and the spent ore goes to the stockpile and you start again with a new heap on the same cells.

The on/off has an obvious cost: permanent double handling (load, unload, start again). But it’s in my opinion an enormous advantage for a management that has never operated a mine before: each cell is a 60-day chance to experiment, at a far lower cost of error than a permanent pad. Management has plenty of chances to optimize its cement recipe, the height, etc., everything that repeats from one cycle to the next. Summary: less efficient, but more forgiving. For a first build, which must serve to finance a bigger one afterward, it’s the best trade-off.

4.2 The numbers behind this leach

We have enough background to interpret the latest official numbers (Ausenco’s updated PFS, effective July 2025):

The reserves:

• 6,178 kt @ 0.65 g/t = 128,903 probable oz, 100% oxide, spread over three pits (MG 82.9 koz, Central 30.0, Machichie 16.0);

• Pits designed with gold at $2,250/oz ($2,220 net of refiner charges) and a Brazilian real at 5.75.

• Plant at 1.0 Mt/yr nominal, with 6.2 years of life;

• Strip ratio 0.78: less than one tonne of waste per tonne of ore, the gift of 3.3 (elsewhere, this ratio runs to multiples);

• Mining contracted out (no fleet to buy, more CapEx avoided);

• Recovery: 85.4% in weighted design, 87.8% in the cash-flow model. The gap with the design is a matter of weighting (we will come back to it in depth in Part 5);

• Life-of-mine production: 113.2 koz recovered, 112.6 koz payable (99.5% after refining deductions).

The costs:

• Initial CapEx of US$37.7M, one of the smallest I have ever seen (thank you, 100 million years of free grinding). US$8M of sustaining over the life and US$1.1M for closure.

• OpEx: mining $3.50/t, Central→MG haulage $1.41/t, processing $10.00/t, G&A $1.5M/yr. Total: about $18/t processed, or about $994/oz of site cost.

• AISC: $1,209.9/oz on the base deck (the deck: the study’s set of price assumptions), $1,243.5 on the spot case;

• Royalties: 4% of gross revenue (Brazilian CFEM 1.5% + Osisko 1.0% + Versamet 1.5%), or about $99/oz at $2,500 gold and $132/oz at $3,340. The first is mandatory; the third, rather good news. We will talk about it in detail in Part 7.

The total economics:

• At $2,500/oz, we have an NPV of $73.9M, an IRR of 77.8% and a payback in 0.85 years.

• At $3,340/oz we have an NPV of $137.8M, an IRR of 139.3% and a payback in 0.6 years.

• At current spot (my quick estimate extrapolated from the PFS sensitivities): at $4,000/oz we have an NPV of about $200M and a payback probably near 0.4 years.

The study is one year old and even its bull case is below the current spot after a strong retracement. At current spot, give or take a few tens of millions, the market cap covers the mine and throws in the district for free. That’s what I call good optionality.

One important caveat: the 77.8% IRR assumes a CapEx that’s already financed and committed. The gold loan has been drawn since November 2025, construction is well advanced (85% at July 9), and 90%-plus of the costs are locked at firm prices.

Financing risk has all but disappeared for Phase 1. Construction execution risk is significantly reduced. Production is coming. One of the best zones, if not the best zone, of the Lassonde curve.

4.3 The front-loading of the cash flows

The mining sequence attacks MG first, Machichie in the second year, and saves Central for last. On the grade fed in, that gives:

• year 1: 0.97 g/t (management even guides “well above 1 g/t” over the first twelve months);

• year 2: 0.77;

• afterward: it diminishes down to 0.45-0.65.

And the pre-production infill adds to it: 25 m @ 7.47 g/t from surface, including 10 m @ 17.09 (the famous RC737 hole), material so rich that it will pass through a gravity circuit (the density sorting of 3.3.4) before even seeing the heap.

All that for about 36 koz recovered over the first two years, or about 32% of life-of-mine production. Year 1’s AISC comes out at $731.8/oz, still nearly 40% under the life-of-mine average. The margin on an oz at $4,000 is more than healthy.

This front-loading is the schedule’s one luxury: most of the cash arrives before the first repayment of the gold loan (March 2027) and above all before the first wet season. But this luxury depends on throughput. A throughput well below the theoretical figure can literally break the plan. It’s not the most probable case, but it clearly has its place in a bear case.

The plant is precisely sized for a 0.85 g/t grade while the reserves grade 0.65 on average. That has the advantage of being able to manage the front-loading (notably for the carbon and elution circuits). Sizing on the life average would have throttled precisely the quarters that count. And the quarters that count are above all the first ones.

4.4 Where the build stands, and the path toward first pour

State of the build at July 9, 2026: about 85% built with more than 90% of costs contracted: “on budget, on schedule” according to the CEO. The dry circuit is finished and mostly commissioned (the sizer is installed and the agglomeration drum is turning). The mining of MG started this week and the first agglomerated ore is starting to go onto pad 1.

The ADR plant (the unit that extracts the gold from the cyanide solution and pours it as doré) was tested and partially commissioned in Perth before shipment (an unusual de-risking step for a junior). It has arrived in Brazil and should be on site end of July. The water-management ponds are finished and pad 2 of 4 should finish this month (with the regulatory authorization for cyanide use finally in hand).

Big positive point: the build came through the 2025-26 rainy season (December-April) on schedule. For a first build in the middle of the Amazon, it’s a real execution data point to add to the team’s track record. On the human side: 308 people on site, 100% Brazilian, 61% of them from Pará.

Here is the remaining path toward first pour, as guided by management:

• Mining and stacking: started. Great.

• ADR installation on site end of July. In principle, not a source of trouble.

• Commissioning of the wet circuit in Q3, its timing depending on how the dry circuit progresses.

• First gold “early in Q4”, then commercial production and ramp-up in Q4 2026.

The process is fast once launched: Machichie’s 2025 columns recover 88 to 96% of the gold in 22 to 34 day cycles. The first ounces will therefore follow the stacking very closely.

But the variables that interest us, potential shareholders, will come after this schedule. The data that confirms or dents the thesis (the real recovery by pit, the real cement dose, the behavior of the heap under the rain) only arrives between Q4-26 and the first wet season of 2027.

A functioning build does not imply a mine that works, let alone a stock that performs. That’s exactly what the next section catalogs: what can break at that point, and what the PFS itself says about it when you read it down to the appendices (more than 400 pages...). Small spoiler: it’s long, but it’s the price of an edge.

Share

5. Metallurgy, the eternal problem explained (relatively) simply

How much gold will the heap actually give up? It’s the tens-of-millions-of-dollars question of this Part 5.

Management announces 88% average recovery, and the few write-ups out there repeat the number and call it “conservative”. The number is “honest”: it comes straight out of the PFS. But “conservative” clearly doesn’t survive a read of Section 13 and its test appendices (how many shareholders actually read them?). What you find there is less flattering and above all more interesting:

• a global design at 85.4%, but calibrated to a hair on laboratory columns, without a scale-up cushion;

• a few weak spots conveniently parked at the end of mine life;

• and one decisive piece of data that the company generated but never published.

We are going to take all that apart, piece by piece. It’s the most technical section of the report, but it’s there that you earn the right to have a real opinion on Phase 1, and on its capacity to finance Phase 2, so roughly the whole thesis. Let’s go.

5.1 Two numbers vs reality

Take a tube of one or two meters and fill it with agglomerated ore. Lovingly irrigate it with cyanide for weeks and measure the gold that comes out at the bottom. That’s the process behind every recovery number we’re going to talk about here.

The PFS gives two numbers:

• “85.4%”: the design average weighted by tonnes and built domain by domain: MG blanket 88%, MG saprolite 90.5%, Central blanket 91%, Central saprolite 72.5%, Machichie 88%.

• “87.8%”: the recovery the cash-flow model achieves over the mine’s life, or 113 koz recovered (out of 129 contained). The gap with the design is a simple matter of weighting: 85.4% is tonnes-weighted, 87.8% ounces-weighted. Now, the ounces live mostly in the domains that leach well. Round, and you have the “88%” that management often trots out.

What interests us is what these numbers assume.

The key parameter of the model is the solution-to-ore ratio: how many liters of cyanide you run per kilo of ore. Naturally, the more you rinse, the more you recover. Now, a real heap will only receive 2.4 L/kg over its 60-day cycle. The design takes it into account, and honestly, fairly cleanly:

• it keeps only the 2024-25 campaigns (the ones where the agglomeration recipe was dialed in, unlike the 2023 tests, which were fairly... questionable2);

• it reads each column at the exact point of the real plan (60 days, 2.4:1) rather than at its final number, which would surely have flattered the result;

• then it knocks off 1% for process losses (because you never know).

As a shareholder, my problem is what the design does not do. A laboratory column is a “perfect” miniature heap. A heap agglomerated with tender care by a technician, watered uniformly to the mL. The column is never compacted under five meters of ore, let alone by machinery driving over it. In short, a commercial heap has dead zones that its lab miniature doesn’t reproduce (all models are wrong, but some are useful).

The industry knows it and has sought to measure this lab-to-field gap for a long time, and Kappes (the founder of the very laboratory that did these tests) had the good sense to quantify it for this class of ore: 92-95% in columns would produce “85% or greater” on a well-conducted production heap. So 7 to 10 points of discount, for the best-run heaps.

On one side we have Cabral’s design at 85.4%, on the other we have Kappes’s “well-executed” floor at 85%. To summarize: 85.4% is a good student’s score, with very little room to maneuver (hence the appeal of the on/off cells, which hugely limit the cost of error compared to a permanent pad).

And then there is Central, the last little subtlety. Its saprolite (about three quarters of the tonnage of its pit, by reconstruction from the resource mix, the reserve split not being published) leaches at 72.5% in design. No excess of conservatism here, it leached at 74-77% in columns. It’s the only domain where the design itself admits a structural weakness (the Central domain accounts for 23% of the reserve ounces). The plan directly sequenced it at the end of the mine life so that discounting does less harm to the NAV. Logical.

5.2 Percolation: where the project already had its bad news

The real risk of a clayey heap leach comes down to one condition: the heap must drink, and it is thirsty. When the fines plug somewhere, the solution doesn’t stop flowing altogether. Gravity always helps it find shortcuts, but it ends up taking only those shortcuts (it is not very original3). The gold of the zones it avoids still exists. You duly paid for it, mined it, stacked it, watered it. But it stays stuck.

One small point in the company’s favor: the permeability data is bad, and they published it anyway. Better still, the design respects it. That’s rarer than it should be, and you can probably thank the CEO’s skin in the game for that.

Let’s start with the test protocol. Again, it’s a scaled-down reproduction: you crush an agglomerated sample under the equivalent of X meters of ore, then you look at whether it still drinks. The results:

• 10 kg/t of cement: two of the three MG composites fail at all heights (the high-grade saprolite, for its part, holds at 5 m);

• 15-20 kg/t: most pass at 5 m, fail beyond;

• 20-35 kg/t: the best cases cap at 10-15 m;

• More specifically, the Machichie blanket fails even at 5 m with 15 kg/t.

I remind you that the industry runs at 9-10 m per lift. This ore only takes 5 m. And even then, only with cement added. The single lift and the on/off of 4.1 suddenly make more sense. I have little doubt that a more... promotional management would have buried these tests in an appendix and kept the 9 m so as not to scare the market.

Good intentions are all well and good, but we’re here to produce gold. And there, the tests aren’t very kind. To summarize, the 5 m is a feasibility floor. There is no floor above and several samples needed 15-20 kg/t of cement just to pass, while the budget provides 10-15. There is probably a “liar”, and we will quantify the impact of the lie just after.

Another worry: the test is a photo. The heap is a film, a film series even, to be precise. The photo answers only one question: this material, crushed under X meters, does it drink at time T? As we saw, a heap lives and ages 60 days:

• the watering detaches fines,

• the current drags them toward the bottom where they accumulate in layers,

• the granules soak up and soften,

• and each passage of equipment packs the whole a little more.

After several stages, a heap that drank perfectly on day 1 may have plugged itself by day 30. Obviously, no static test will have seen it coming.

Let’s take a concrete case. Bald Mountain, Nevada (a Barrick Gold heap leach, one of the biggest gold giants). A successful mine with a competent operator. Yet probe holes in a pad in full operation found a thin layer of fines, next to nothing, that was making the solution stagnate. In 17 holes out of 32! The gold was waiting patiently below, and it probably waits still. If that happens to Barrick in Nevada, it can happen to a first build in the Amazon.

What to conclude from it? First, remember that the permeability of a heap is re-decided every day, at the agglomeration drum. Batch after batch. No PFS can guarantee anything in advance and as we saw, Cabral’s has no height margin to absorb a bad batch. It’s not an all-or-nothing: if it goes badly it’s “just” an erosion of the margins; and at “worst” it will need to fund itself some other way to keep financing the potential district. But I am an investor. And it’s the second most important point to watch in the coming months. Management knows it, besides, and even has the decency to write it: agglomeration “must be closely followed”.

The CEO said in an interview that he was so excited by this project that sometimes he could not sleep. I’d bet sleep sometimes keeps him waiting over one too-capricious agglomeration scenario.

5.3 Cement: the only reagent that can miss the budget

Cement is the glue of agglomeration, but above all, it’s a consumable.

The budget provides 11,819 tonnes per year, or 11.8 kg/t at 1 Mtpa. The design, for its part, is quoted at 10-15.

What the lab tests say of it:

• 2022: 24 kg/t;

• 2023: 10 kg/t (the questionable program);

• 2024: 14.8-20.1 kg/t;

• 2025: 14.6-15.1 kg/t.

So outside 2023, the lab never came in under 14.6. I remind you that the budget pays for 11.8. Even at 11.8 cement is among the main OpEx lines.

The sensitivity calculation is simple and I obviously built it into my model: ground that demands 18-20 kg/t instead of about 11.8 makes the cement line half again as expensive, or about $70/oz of AISC (I included the incremental diesel cost). Almost trivial at $4,000 the ounce, already more painful at $3,000.

For the connoisseurs: cyanide is here a quiet line. No sulfides that would consume the cyanide in the place of the gold. No preg-robbing (natural carbon in the ore that would act as a rival sponge, re-capturing the gold as soon as it dissolves). Arsenic is low too. In practice, the lab-to-field rule of KCA (the laboratory of the same Kappes) says that a real heap consumes 25 to 33% of what the columns consume. Applied to Cabral’s columns, the rule predicts a real need of 0.06 to 0.16 kg of cyanide per tonne. The budget allows 0.25. Plenty of headroom.

Summary: one of the main questions, but not decisive either, is how many kg of cement are consumed per tonne of ore processed.

5.4 The clays and the missing data

Everything we said in 5.3 depends on one single variable: which clay. For what interests us, there are two big possibilities:

• Kaolinite, which is born from a granite rotted under a humid climate (exactly the recipe here). It doesn’t swell when mixed and docilely lets itself be glued into granules.

• Smectite (think cat litter). It drinks water into its own structure while swelling, migrates, and has a tendency to re-plug the channels (the “voids” where the solution is supposed to circulate) that the cement had created. It’s a heap killer, and the prime suspect in the Bald Mountain plugging (5.2).

So, kaolinite or smectite? In what proportion? Domain by domain?

The company has the answer.

Three programs (2022, 2024, 2025) sent samples to two specialized laboratories for XRD analyses (X-ray diffraction, which identifies the mineral species) and CEC (cation exchange capacity, the proxy for swelling clay). The PFS mentions these orders. Then nothing more. Over more than 400 pages of PFS and no clay species, no CEC value, no table. The test isn’t ambiguous: kaolinite sits at 3-15 meq/100 g, smectite at 80-150. The words kaolinite and smectite literally appear nowhere in the report. And no result published in the others either, while the analysis orders themselves do appear.

Let’s reason Bayesian (it should be a pleonasm): a profile rotted on granite under Amazonian rain gives dominant kaolinite. Smectite needs the opposite: base cations that stay in place, so restricted drainage or a rock richer in calcium and magnesium than a granite. The one place I would still look is Central, where the drill logs diorite porphyry dykes (see 3.1) and where the saprolite happens to leach worst in the whole project. A diorite carries more calcium and magnesium than the granite around it, which is exactly what smectite needs. Next, if the results were really bad, building the plant would be a strange choice, above all for a CEO who put several millions of his cash into the company (but not impossible, it’s been seen before). But all the same. A company that pays three times for a decisive analysis and doesn’t publish the result, you can’t give it the benefit of the doubt.

I asked the company for these results and got nothing back. Until that changes, I owe it to you to price in this uncertainty by being hard on my estimates of the cement needed to reach the intended production.

5.5 Water: the underestimated risk

Here I stray the furthest from the consensus, because it quite simply doesn’t come up.

Let’s do a little arithmetic: 2,158 mm of rain per year on the site vs 840 mm of evaporation. Or a delta of about 1,300 mm that does not evaporate, of which 70% concentrated from December to April.

It counts because the leach circuit is supposed to be a closed circuit. The same solution turns continuously: watered at the top, the pregnant solution recovered at the bottom, the solution poured over the carbon to recover the gold, re-dosed with cyanide, sent back up, etc. Nothing is supposed to leave it. But closed systems only exist on paper. The rain that falls on the heaps and the ponds joins the party. Four to five months per year, structurally more water will therefore come in than goes out. This surplus must leave, and as this water is cyanide-bearing, it can only leave treated.

5.5.1 Four tools, and a water balance in the future tense

The design answers with four tools:

• Rain covers (raincoats for the heap), deployed in the wet season on the cells that are not in active leaching. That keeps the rainwater from getting contaminated and lets it be drained off as “clean water” before it enters the circuit. But the active cells stay open to the sky, so the rain will inevitably fall on what is being watered and will join the pregnant solution;

• Five ponds (about 76,500 m³ in all4) sized for the 24-hour once-in-a-century storm, i.e. 211 mm (1/10 of the annual rain fallen in one single day). The storage calculation assumes the covers are in place.

• Diversion channels around the installations (nothing surprising).

• And a detox-discharge train (the chemical destruction of the cyanide with peroxide and copper sulfate, before releasing it into the environment). Surprisingly, it is presented as a backup: the installation is “zero-discharge… under typical annual climate conditions”, and the train will run “only… as required”. Typical, okay, but it’s almost certain that atypical conditions will show up for a meaningful stretch, above all in the Amazon.

One important element is missing from the list (and from the PFS), the one that would have completely changed this part: the site-wide water balance. “The site-wide water balance will determine whether treatment and discharge… are necessary”, in the future tense, and the list of recommended work adds “more detailed process and site-wide water balance is required”.

5.5.2 Three choices, in increasing order of consequence

Three choices make me raise both eyebrows.

One: the covers. The literature of the same Kappes documents their mechanical difficulty in the field (Rio Chiquito, Costa Rica: the cover “has not worked very well because of the mechanical difficulties of moving the cover”). Making them the first line of defense against a five-month monsoon is a fairly significant operational bet. If Rio Chiquito repeats here, it’s the assumption behind the storage calculation that collapses.

Two: the drip. Cabral will water its heaps by lines of pierced pipes that seep the solution without projecting it. The alternative would be to spray via a sprinkler, which disperses instead of letting flow. On the tropical projects followed by KCA, about 7% of all the pumped flow leaves again as vapor (all sources combined). That’s a little over 6 m of water/year shed for free over the watered zones. Kappes’s conclusion is that a well-designed sprinkler heap manages to hold the water balance up to 2,500 mm of rain/year. So more than falls at Cuiú Cuiú. That’s exactly why the five tropical heap leaches in his survey all irrigate that way. The drip, which Cabral chose, does the opposite. It minimizes evaporation. A virtue in an arid site, but clearly not in the Amazon.

As for the reason, I have two candidates. One: a sprinkler puts cyanide solution into the air as mist, and an operating permit still sitting in a drawer in Belém is not really where you want that discussion. Two: spraying onto a 5 m lift of agglomerates that already fail their percolation tests above that height is an excellent way to crust the surface and drag the fines down (the failure mode in 5.2). Both are good reasons, but not explicitly theirs: the word "sprinkler" appears nowhere in the PFS. So I'll grant them the trade-off, but I mark them down again for the silence.

Three: the status of the discharge. “Zero discharge in typical conditions” means discharge in wet conditions, which come back every year. Let’s be honest, discharging the treated surplus in the rainy season under the tropics is accepted practice, if not exactly routine. Fine. But the PFS gives neither the limits nor the seasonal capacity of the discharge. For the cyanide-bearing water, we get just one adjective, “acceptable”, zero numbers. The word outorga (the Brazilian water permit that covers withdrawal + discharge) literally appears nowhere in the report, and the section that promises “Water Rights” in its title says not one word of it. We will come back to it.

5.5.3 The first wet season

I can seem to be nitpicking, and I am probably even nitpicking right now. But I have seen, and myself lost, too much money in a mining thesis for not having nitpicked enough. The devil isn’t in the details. The devil is the details.

Here is an example where nitpicking would have been necessary. In July 2012, Ghana’s EPA stopped the Tarkwa heap leach (about 1/4 of the site’s production), belonging to Gold Fields, on a problem of conductivity of the discharged water linked to salinity (a routine parameter). The stop lasted three and a half weeks, or about 15,000 ounces lost, and the lasting fix, reverse-osmosis stations, cost tens of millions of $. We are talking about a world-class mine managed by a major operator with metallurgy that worked, and production stopped on the decision of an agency. If we transpose into Cabral’s context, the scenario “the heap works but the water management jams” is one I don't see discussed much.

And it’s not over, the rain also hurts the upstream side. The ore arrives from the pit with up to 20% stated moisture (the tests measure in reality 22.1%). If it is wetter, the screening can slip, the agglomeration too (the QP flags it himself). Now, the drum has almost no water margin before making mud instead of the expected granules. We have already seen numerous precedents in the region, and Cabral doesn’t have the option, like others, of switching to dry grinding.

To be entirely fair, the “it just works out” option exists. Sansu, in Ghana, held 2.5 m of annual rain with 60,000 m³ of ponds for 3,000 t/d. Cabral, at 2,740 t/d, has comparable storage, bigger even. But, and I insist on this but, Sansu watered with sprinklers, with the evaporation valve running. Cabral stacks the same storage, certainly, but without this valve.

Hence my timing conclusion: the real commissioning test of this plant will be the first wet season in operation, or December 2026-April 2027. First pour will only be a dress rehearsal.

5.6 The samples vs reality

We’re nearly at the end of the audit, I promise. But one ultimate qualitative judgment remains for us to make: the sample itself. Ten samples, six of them 2024 composites (averaged over MG and Central) + four 2025 composites (Machichie). For 6.2 years of plan. It’s clearly limited and even the PFS documents these limits:

• Part of the 2024 drilling fell outside the optimized pit. So that’s tested ore that will never be mined. Great.

• The Machichie samples were taken at surface (for lack of available drill core), crushed much finer than the plant will, and tested in columns of 2.3 m of ore, vs 5 m in reality. So literally three gaps that improve the test conditions relative to reality.

The QP duly noted these gaps and recommended filling them by... simulating reality more faithfully (enjoy the pleonasm): a large column representative of year 1’s ore with variability tests to fill the uncertainty. As he writes: “a small number of tests… can result in a lower accuracy of production estimates”. These tests have, to my knowledge, not yet been done by management. And while we inventory what the PFS puts at its own charge, two caveats I hadn’t passed on yet: the high-grade veins of Machichie excluded from the test columns where the “design recoveries may not be achieved”, and the cement dosage calibrated “based on a limited number of tests”.

Speaking of the QPs, a few details worth the detour. The metallurgy QP (Raponi, Ausenco, 40 years in the business) has never visited the site: his interpretation rests on the laboratory reports and the geological model. The geotechnical QP (Elfen), for his part, spent two days there, with Pierina and Lagunas Norte on the pedigree (roughly, heaps in difficult climate). KCA is the world reference of the discipline, and a vendor with the decency to grade its own predictions. Points to weigh for attributing a degree of confidence to all these percentages.

5.7 Mercury, courtesy of the Tapajós

The ore carries mercury, up to 1.8 g/t on the tested samples. In another province, it would almost be an appendix line, but in this region, garimpo mercury is the environmental issue of the region.

The plan treats it seriously. The mercury follows the gold all along the circuit, up to the ADR plant, and the PFS imposes capture and abatement equipment there. It’s already budgeted and without major metallurgical impact.

5.8 The engineer’s verdict, and what it costs in ounces

Let’s summarize all that.

The ore is of the right class: saprolite is the most predictable heap-leach ore there is, relatively sparing in cyanide. The process is well known: Ity (Côte d’Ivoire) leached cement-agglomerated saprolite for decades for more than 1.2 Moz and Yatela (Mali) ran the same drum-conveyor process for a dozen years.

The execution in this context is relatively new: this ore in this type of climate with this team.

The design is honest and adapted to the main variables: the 5 m lift in on/off is the right answer to tests that ruled out anything better, and the 85.4% already integrates Central’s true number.

The flaws are localized and known (at least in this deep dive):

• an agglomeration without margin for error;

• a cement probably under-provisioned in the budget;

• a clay mineralogy withheld, at least from investors;

• water management treated in an appendix, with choices I judge shaky without more information;

• a year 1 tested on almost nothing.

My technical bear case. A recovery that settles at 78-82% (full lab-to-field haircut on MG + Central under its 72.5% via slow veins), a cement at 18-20 kg/t. Cost: about 6% fewer ounces, about US$70/oz more AISC. If I enter that into the inputs of the mine model: US$10-25M less NPV at US$2,500 gold, only about 12% at US$4,000.

To summarize: the metallurgical bear is a margin risk at the current gold price. For it to break something, it would have to be combined with collapsed gold AND capped throughput. I see it as the uncomfortable variable in the case, very probably badly understood by the market, but which has the advantage of having a big cushion underneath: the per-ounce margin.

What would make me change my opinion, quarter by quarter, in one direction or the other:

• the clay reports (XRD/CEC) published (or not);

• the big year 1 column, at design ratio and cycle (not likely);

• the reconciled recovery by source pit: MG first but above all Central;

• the pregnant solution flow at the foot of the heap (roughly the pulse of the percolation). The design promises 242 m³/h, and any plugging will read there first;

• the cement tonnage against budget (5.3);

• and the behavior of the site between December 2026 and April 2027.

A clean wet season, even with a somewhat higher cement cost, and this section becomes a paragraph in the next report. A failed wet season, and it’s the thesis that changes shape. For that, see you in April 2027.

Share

6. Phase 2: the exploration, and the district thesis

I described the cash machine in Part 4, we are finally going to get to what it is supposed to feed.

My work here: transform “50 targets and spectacular holes” (the pitch) into a tiered inventory with probabilities and expiration dates (an analysis).

Let’s have some fun.

6.1 A little vocabulary just in case

Small reminder:

• Inferred = mapped hypothesis < Indicated = measured to within 50 m < Measured.

• Reserve = Measured + Indicated that survives the modifying factors (dilution, recovery, costs, price) in a committed mine plan.

• Proven derives from Measured, Probable from Indicated.

Cabral = 128,903 oz at the most demanding rung (100% probable) + about 1.2 Moz at the resource rung. The market pays about $1,970 an ounce for the probables and about $212 an ounce for resources (reconstructed in 10.2). I am going to spend this section debating the denominator.

Small subtlety, (too) often decisive: a resource is a function of price before it’s physical inventory.

Raise the gold price in the pit shell (the pit envelope that defines what is mineable) and three things happen without a meter of drilling:

• the cut-off drops (tonnes already drilled become ore)

• the optimized pit widens/deepens,

• marginal zones can be connected (it becomes economically viable to do it).

That translates into more resources mined at economic margins, resources that were going to be mined but that get mined for cheaper, or operational simplifications that let you mine faster. If it’s not clear, examples will clarify.

6.2 The official inventory: 1,197,800 ounces, of which 75% counted at a price from four years ago

The resource of record, per the latest AIF (the Canadian equivalent of the 10-K): 670.5 koz Indicated (26.05 Mt @ 0.80) plus 527.3 koz Inferred (20.16 Mt @ 0.81), about 1,197.8 koz.

Broken out by type:

• the oxide: 287-292 koz depending on the document you open (the gap, about 5 koz, comes from two cuts of the Machichie shell), supposedly fresh from 2024-2025;

• the hard rock: about 905.5 koz (450.3 Ind + 455.2 Inf, of which 157.6 koz in an underground scenario, too deep for a pit, all of it going to Inferred), estimated... in 2022, with a pit shell at $1,800 an ounce. The Inferred accounts for about 44% of the total resource (and about 50% of the Phase 2 hard rock alone).

These 905 koz of hard rock were therefore counted at less than half of the current spot, before 35,000-40,000 m of drilling, and before four post-2022 discoveries (Machichie Main, Machichie NE, PDM primary, Jerimum Cima). That leaves a lot of room to raise this number. Nothing like Cabral’s own past to show it.

Between 2018 and 2022, a rise in the gold price lowered the cut-off from 0.35 to 0.26 g/t. Add the re-modeling from the purge (3.7), and the Indicated jumped 253%. Without a new deposit.

Re-cutting the 905 koz with a shell at $3,500-4,000 is mechanically the least complicated gain in the whole case: all it takes is the existing data, a good computer and a competent geologist (and/or a very good AI). The real question is the magnitude of these gains. We will come back to it later. For now, we just shouldn’t kid ourselves: these ounces are often leaner and deeper than the ones from new drilling. Okay, it inflates the denominator. But it also dilutes their quality in part. And a potential acquirer knows that.

The elephant in the room is the timing: this inventory is stitched together from three reports at different dates (primary July 2022, MG/PDM oxide October 2024, Central/Machichie oxide July 2025). The EV/oz calculated on it mixes ounces counted at three different gold prices. It’s a fairly messy valuation, but it’s the only one we have for now.

6.3 The pending update that can change everything

The company promised a global resource update “later this year” (May-June 2026 releases, so H2 2026).

On the menu:

• 6-7 modeled deposits versus five in 2022, a “healthy jump” according to Carter,

• maiden resources targeted for Jerimum Cima, Machichie Main primary, Machichie NE and the PDM primary (the Mutum zone of 3.4.2),

• a redo of MG and Central,

• and a first possible move of MG into Measured after the infill (165 holes and 5,767 m, 41 results still pending to this day).

Two observations of discipline.

One: the “MG to Measured” upgrade was promised “by the end of this month” in the May 26 release, then in the June 11 one. Still “in preparation” in July. That’s two deadline slips in two months. Nothing serious for a developer, but the metronome of scheduling promises has to be read with that in mind.

Edit July 21, 2026: the last 41 holes are out, infill closed at 165, tail “in line with expectations” (nothing that beats RC737). Measured upgrade now promised “by month-end.”

Two: any growth number written before this update, including mine below, will expire in Q4 2026. It’s the big short-term catalyst. It can completely invalidate my scenarios, upward or downward. Given the time I spent on this deep dive, they had better be... no, nothing.

6.4 The drilling machine, and how to read its results

Capacity, first. Six rigs in June 2026 (an interview number; the releases only confirm four rigs), five already on site at last check. Four are diamond rigs: they bring up an intact core of rock and offer reference samples that allow better estimates. The two others are RC (reverse circulation): they grind the rock and bring up chips. It’s faster, it’s cheaper, but no core. The construction and exploration teams are kept separate, so building the mine did not slow the drilling. Two details only come out of the interviews: the two RC rigs would belong to Cabral (no contractor to pay by the meter), and the second-half target would be 25,000-30,000 m.

The results now, and above all how to read them. It’s an art to handle with care. Here are four rules for doing it.

Rule 1: the value of an intercept (the mineralized slice a hole cuts) is the gram-meter = thickness x grade. Cabral publishes its own internal ranking, which simplifies the calibration. A rough rule of thumb: under 30 g·m, it’s noise. Above 100, serious. Above 250, it’s exceptional:

• The best hole of the district, DDH372, makes 9.5 m @ 87.4 g/t (of which 2.9 m @ 285.5), about 830 g·m.

• The fifth, DDH384, makes 107.6 m @ 2.5, or 269 g·m.

• The fourteenth is an old friend: RC737, the star hole of the start-up (4.3), 25 m @ 7.47 from surface, about 187 g·m.

Rule 2: the true width. Each Cabral drilling release carries the warning “True widths may be up to 50% of actual drill intercepts”. Translation: the drill cuts the zones at an angle and therefore measures longer than the real thickness (it’s simply trigonometry). The 9.5 m @ 87.4 therefore makes, at best, 4 to 5 m of true width. And the 107.6 m of DDH384 is a length pierced at 60° in subvertical zones: the 107-meter layer does not really exist.

Rule 3: a hole is not a resource. To count ounces, you must prove that the mineralization continues between the holes. Remember, holes spaced 50 m at most for Indicated, and an extrapolation at most 100 m for Inferred. An isolated intercept, even at 830 g·m, is therefore worth zero ounces until it’s surrounded by dozens of holes. And it’s better that way. A lot of value has been destroyed in exceptional holes surrounded by waste. Take Jerimum Cima, management says so itself: “not currently part of the existing resource base”. To try to transform that into resource, management bet on the analogy holding with MG, where the rich veins follow each other over 700-800 m. It’s an argument for deciding where to drill. But it’s the drilling results that will have the last word. They always have it.

Rule 4: a grade without context means nothing. 25 m @ 0.7 g/t from surface in saprolite, that’s 17.5 g·m and 2027 cash. A shovel will do and the plant is already paid for. 10 m @ 8.7 in hard rock at 100 m depth, that’s 87 g·m (5x more) and yet it’s only an argument for the future Phase 2 plant with the $300-500M of CapEx Carter floated in an interview (nothing concrete) for production in 2029 at the earliest. That’s two more years of discounting and all the uncertainty that comes with it.

6.5 The pipeline, tier by tier

On to the complete inventory. I classified it from the closest to the furthest from the “bank” according to my own judgment, as always when we talk probability. I base myself above all on analogies with the deposits already modeled. My end-of-section scenarios rest on these assumptions.

6.5.1 Tier A, the defined resource

1,197.8 koz. 100% probability, it’s already in the bank. Remember that the freshness isn’t uniform (oxide 2024-25 vs primary 2022).

6.5.2 Tier B, the deposits already modeled

This is low-risk growth. Two mechanisms drive this growth.

The infill first: drilling between the existing holes tightens the grid, and ounces already counted in a low category move up a rung (from Inferred toward Indicated, even a first possible Measured at MG). That creates few new ounces, sure, but no matter. The goal is to make the existing resources more bankable.

Second method, the step-outs: drilling where the 2022 model stopped for lack of holes while the bodies remained open. Here the goal is clearly to add new ounces, but following a model that already exists and that raises the probabilities of success. MG finished its infill (the 166 RC holes of 6.3) and Central, 21.5 Mt @ 0.88 (Indicated + Inferred, for a total of about 608 koz), is well known. Hence my high probability of 70-85%: you drill where you already know the gold probably is. I put the gross incremental content at 80-150 koz (8 to 14% of MG+Central). I could be a little more optimistic, but optimism rarely rewards mining shareholders.

One variable remains, maybe some of you are waiting for it: the gold price. We explained that a rising gold price allows transforming waste into viable ore. Is that the case here? First, it mostly applies to the 905 koz of primary. But enormous uncertainty still remains above all, and the distribution probably has a fat tail toward 0. I do not know if it’s an all-or-nothing, but the nothing is not negligible. So I count 0 additional ounces for this variable. I see it as a free lottery ticket.

6.5.3 Tier C, the targets drilled without resource

The main source of the potential 2026 re-rating.

Here all the published holes prove there is gold, but none proves a continuity. The 2026 update will be the judge target by target: either the framing holes connected the hits and resources get added, or they didn’t and on to the next. Hence the coin flip: 40-60% depending on the target.

Here the concerned deposits are (in decreasing order of potential): Jerimum Cima, Machichie Main, Machichie NE and Mutum (the PDM primary). I spare you a new description of the deposits, but here is the summary visual just in case (plus it bumps up my ROI on the time spent making them, win-win).

Gross content: 150-300 koz for Jerimum Cima. My reasoning is that MG (the flagship) already carries about 460 koz of total resource for a 700-800 m system (6.4); Jerimum Cima already outlines a 900 m system. Adding 150-300 koz is betting it will be worth between 1/3 and 2/3 of an MG at equal size. Here too no heroic hypothesis, just one more deposit in the district average with a conservative haircut.

Following this logic, I take 80-180 koz for the rest of the tier (0 for Mutum which has only one hole, let’s not get carried away).

6.5.4 Tier D, the undrilled

Here we enter pure optionality, the one that makes this bet attractive in my eyes, and where the future district scenario can live.

Forget the holes. Here we do pure prospective geology. We’re two rungs below anything concrete, but the clues are serious. More than 50 targets on a 10x15 km district, while the deposits currently modeled fit within a core of a few kilometers.

Remember also the gold-in-soil anomaly of about 7 km, against about 1 km at Tocantinzinho before its discovery. And about ten boulder fields (2.2, blocks that erosion tore from a vein never found), gathered at surface and assayed, with averages that stop you cold: Alonso, 91.7 g/t over 23 blocks along 950 m, more than the best hole of the district; Tracajá 74.9; Medusa 24.7; isolated samples up to 3,727 g/t, 5,700x the grade of the reserves.

I insist, nothing is less sure. These blocks only give an indication of where to look, no more, no less. The drone-mag, the targeting machine of 3.2, sorts the pile, AI-assisted, over about 20 km².

And so, what is the probability that a prioritized target becomes a deposit one day, and what is the quality of this potential deposit? I have no idea. What is the probability that several deposits get found and make the district thesis true with millions of oz? I have strictly no idea. The situation is much too complex.

I know neither the probability of finding a deposit, nor the potential value of this deposit. I can’t even ballpark the distribution of probabilities. To resolve the district thesis, I would have to calculate these probabilities and the values associated with these potential deposits for each of the potential targets, and calculate the expected value. Impossible. Even ballparking it, impossible. So I had to find a workaround, but more on that later.

6.6 The track record, and the ROIC of the drill

Two questions remain for me to answer on the drill: has it already proven that it transforms meters into ounces, and is the dollar spent well used?

On the first, the verifiable history is fairly limited. 171 koz Indicated in 2018. 604 koz Indicated in 2022 (1,138.5 with the Inferred), but most of it comes above all from the model purge (3.7). It’s more an accounting correction than an effective drill.

Four years later we’re up 59 koz in total, the Indicated having gained 66 koz and the Inferred given back 7, while four named discoveries piled up off balance sheet. Thin, but the context explains it. The pedigree of the man himself is excellent: five discoveries from scratch, including Tocantinzinho. The end-2026 update will be Cabral’s first real test. The drill drilled, the dollars were spent, the results published hole by hole. The question that remains: how many ounces will be counted.

On the second, we can at least quantify. C$1.93M (drilling 1.74M + assays 0.19M) for 10,087 m in the last half, or C$192/m. The 35-40,000 m drilled since 2022 therefore cost about US$5M direct (at the 0.71 exchange rate), say US$9-13M once loaded with the rest (geologists, camp, geophysics, which I round generously). What all these drilled meters buy is a range of 310-630 gross koz (sum of tiers B and C). On this base, the prospective creation cost falls in the model’s range, about US$16-30/oz.

An ounce the market pays about US$212 for at Cabral. In recent Brazilian M&A, once normalized to the current gold price, we are in a similar range: $200-260 (I detail in 10.5). One dollar in the drill = $7 to $13. And I remind you that I was conservative in my estimates and that I stripped the district angle out of the base. Cut it in half and we’re at a 3x at least. Weight by my 6.5 probabilities and we’re still at 2x to 10x. Suppose even that I have it all wrong and that Cabral drills at the industry’s average cost (about US$62 an ounce according to MinEx, 2009-2018), it still comes out to nearly 3.4x. Any way you cut it, the drill pays for itself.

Obviously, this is only prospective and the update will be the moment that sets the denominator. What’s more, the numerator varies (Cabral’s market cap). Nothing is fixed. But the trend is clear. Drilling is by far the best use of Phase 1’s cash, and it’s precisely Phase 2 of the plan. And about time.

6.7 The Phase 2 PEA: the second Lassonde cycle

How much is Phase 2 worth? Zero NPV in this write-up.

Management’s calendar:

• resource update end 2026,

• then PEA in Q2 2027 (the first economic study of the hard rock).

A PEA (Preliminary Economic Assessment) is the rung under the PFS in the ladder of studies: a relatively conceptual mine plan, a ballpark CapEx, an NPV just as ballpark, and above all the right to count the Inferred (a right the PFS doesn’t have).

To summarize, the update recounts the warehouse, and the PEA is the first quote (in pencil) for the plant that would empty it. Then a PFS end 2027, a construction decision “probably in 2028” and a build over 2028-29.

The orders of magnitude (careful, all these numbers come only from interviews): 150-250 koz per year, of the same caliber as Tocantinzinho, 7-10x Phase 1, a CapEx “3, 4, 500 million… may be more”, also x10 Phase 1.

One hard piece of data already exists: hard-rock recoveries tested at 90-97% in the lab (the gravity and the CIL tanks of 3.3.4). Two facts put this in a little context. One fact: the LP of March 10, 2026 covers the complete project (”1 million tonnes/year, and beyond”, says the release). Concretely, the five-year permit wall is behind for both phases. Only the LI then the LO remain, and they’re shorter. Two: the financing, estimated at US$300-500M, will be a 2028 problem (still according to management’s forecasts), and backed by Phase 1’s cash flow (and by the re-rating that may have happened by then).

Hence the refusal. To give an NPV to Phase 2 today, you would have to build the plant in a spreadsheet: take the interview numbers above, draw cash flows from them, then multiply the whole thing by a probability of success. The industry calls that an rNPV. Right now, I call that science fiction: invented assumptions, multiplied by an imaginary probability. Your time and mine are too precious for that.

I chose something more modest. I value the hard-rock ounces like inventory, at a per-ounce price that reflects the degree of confidence:

• $80-150 the Indicated,

• $30-70 the Inferred.

It’s 2-4x under the $200 that the market pays for ounces of this type on average, but it’s necessary. The oxide starter carries its value in DCF because it arrives in a few months. The 2027 PEA will transform this denominator into an NPV question, for potential revenue in 2028 at the earliest. With the discounting that comes with it. Hence my significant haircut.

I’ll note all the same that the PEA will change the nature of these objects. They will have earned the right to belong to the NAV (if there are ounces, obviously), with all the Lassonde-curve arguments that come with it (discovery euphoria, development trough, re-rating on the approach to production).

6.8 So, what growth is probable?

It’s time to assemble all these hypotheses.

The bear: 1.20 Moz.

Nothing new on the board. 0 additional conversion, all this work led to nothing more. The advantage of this scenario is that it depends (almost) not at all on the future. The dream.

The base: 1.50 Moz.

1,198 koz in the bank + 90-115 koz of infill and step-outs (weighted range of tier B: 56-128) + 120-150 koz at Jerimum Cima (weighted: 60-180) + 55-80 koz on Machichie Main, Machichie NE (weighted: 32-108).

A quick note: at the base, the model splits the new oz into 40% oxide and 60% hard rock (the oxide share drops when the increment rises). As a reminder, the 40% have a plant nearly paid for and ready in a few months. The 60% hard rock will go on to fatten the PEA.

The bull: 2.00 Moz.

Tiers B and C converted almost in full (about 470 of the 630 gross koz), plus roughly the rest coming from the first successes of tier D. Why 2M? Simply because it’s management’s target and the number they keep repeating.

And here I need to add an extremely important caveat that will deeply shape all the rest of this deep dive, including, even above all, the valuation and my own strategy.

That’s it for free subscribers.

Marketing 101 would have me tease what’s below, play on your frustration, all that. I’m too tired for that. I should have published hours ago. A corrupted file decided otherwise.

You already know what’s down there: the remaining 40% of the deep dive, same caliber as what you’ve read so far.

For those interested: I publish 20+ deep dives like this a year, with regular follow-ups, up-to-date valuation models, and a lot more. Go have a look. A lot of you have already joined, and I know from experience most people are convinced long before they get here.

If you’re on the fence, or just have a burning urge to read to the end without paying: one free article. Order from the house.

Read more

Here’s a free post, on me. Spend it one of two ways: on this short-form thesis, which says a lot but not everything. Or on the big one, a long read, but the complete picture. Your choice.

Read more

Let's go. (Full thesis here)

Read more

Build-A-Bear (BBW) took the idea of building your own stuffed animal, and took it very seriously.

Thirty years later, it’s one of the most recognized brands in America, a company with an exceptional run in the market (in the statistical sense of the word) and a valuation that looks nothing like its operating performance.

I’m not a shareholder yet, but I intend to be, under certain conditions. All of them spelled out, of course.

We’re in this together for the next 25,000 words, so let’s not waste time.

Table of Contents

0. Table of Contents

1. The Short Thesis: Build-A-Bear’s Bull Case in Brief

2. Anatomy of the Bear: The Business Model

   1. The Original Business Model, in One Gesture

   2. The Products and the Average Basket

   3. Beyond Make-Your-Own: Wholesale and Mini Beans

   4. The Brands and Licensing

   5. The Customers: From Kids to Kidults

   6. The Doors: Stores, Formats and Franchises

   7. The Unit Economics

   8. Behind the Scenes: Supply Chain and Operations

   9. Competition

3. The Macro Backdrop

   1. The Context: Consumer Sentiment

   2. The Behavior: How Consumers Actually Spend

4. Build-A-Bear’s History, Brick by Brick

   1. A Study in Pink: The Founding and First Boom (1997-2007)

   2. The Valley of Fear: Crisis, New CEO and Turnaround (2008-2019)

   3. The Adventure of the Second Stain: Covid and the Record Run (2020-Today)

5. The Financials

   1. The Income Statement

   2. The Segments: Two Companies, One Multiple

   3. The Balance Sheet and the Flows: The Capacity to Wait

   4. Return on Capital (ROIC)

   5. The Latest Quarter: Q1 2026 Results

      1. The Quarter Itself

      2. Analyzing the Walmart Partnership

6. The Crux: The Bear Case and the Four Risks

   1. The Four Problems

   2. Falling In-Store Traffic

   3. Clickocalypse: AI Search and E-Commerce Traffic

   4. Tariffs, Ô Tariffs: Supply Chain and Pricing Power

   5. SG&A and Margins

   6. The Overall Read: Weighing the Risks

   7. Other Risks

7. Management and Capital Allocation

   1. A New Dawn: The CEO Transition

   2. The New Management

   3. Capital Allocation

   4. Incentives and Compensation

   5. Insider Buying/Selling

8. Valuation: What Build-A-Bear Is Worth

   1. Napkin: A Reverse-DCF First Pass

   2. How Much, and Under What Conditions

      1. The Variables

      2. Growth by Segment

      3. The Contribution Margins

      4. The Rest of the Assumptions

   3. So, How Much? IRR by Scenario

   4. Sensitivities: What Moves the Return

   5. The Free Upside: Optionality You Aren’t Paying For

   6. Limiting the Cost of Waiting and Uncertainty

   7. Limiting the Losses

9. Conclusion

1. The Short Thesis: Build-A-Bear’s Bull Case in Brief

Starting point: the market prices Build-A-Bear (BBW.NYSE) at 7.3x earnings (price: $31, market cap: $392M).

At that level, you’re at least looking at a retailer at the bottom of its cycle, if not a company in real trouble, possibly both. So there are two questions to settle.

• Is this really a cyclical bottom?

• Is BBW really in trouble?

The first would take more arguments than a short thesis can carry, so let’s assume the worst-case answer just in case: yes, it’s a cyclical bottom, with all the trouble that comes with it.

Here’s how the company “copes” with that trough: BBW just booked its fifth record year in a row. $529.9M in revenue, $67.2M in pre-tax income, and EPS of $4.26, also a record. And it’s guiding for another record year on the top line.

A battered valuation that keeps printing record after record is always hiding something. In this case, almost certainly an opportunity.

The bet comes down to three things.

One: the weakness is cyclical. It may already be past its low, but it doesn’t even have to be. Michigan sentiment hit 44.8 in May and is already back to 49.5 in June. Build-A-Bear’s traffic has tanked far harder than most retailers’, with a real soft patch starting in 2026 (not for the obvious reasons, we’ll get to that, trust me). Weak, sure, but still a record year. The company guided for another record year, even after cutting guidance. I know, guidance isn’t a strong counterargument, but it was worth mentioning. Don’t worry, we’ll dig into the margins, the cash flows, and a few 10-K footnotes that say a lot more than any conf call.

Two: the engine of value creation has already changed in nature, and the stock probably hasn’t caught on. I know, it sounds like a cliché, but I’ve got the numbers to back it up. Commercial (wholesale plus licensing) is growing more than 20% a year (at high margin), some BBW products just landed in Walmart, and the company runs an international franchise that eats almost no capital. At 7% of revenue, commercial is still small, which is probably why it’s not in the price. Small, but it matters. Covering all of it will take some rummaging through the accounts and a few thousand words.

Three: the balance sheet pays you to wait. Zero debt, about $65M in operating cash flow, $40M in free cash flow, enough to buy back its own shares at 7.3x earnings (19% of its shares bought back over the last four years) and ride out a long trough while the fragile ones cut into their spending. Time is the enemy of bad businesses and the ally of great ones. BBW can wait a long time. A very long time.

Laid out like this, the asymmetry is almost embarrassingly simple. The eternal question: am I right, am I wrong, by how much, and at what odds. The eternal answer: here are my arguments and my numbers, judge for yourself.

Subscribe now

2. Anatomy of the Bear: The Business Model

Let’s start with the bones.

2.1 The Original Business Model, in One Gesture

A kid walks in, scans the shelves, picks out a deflated plush shell, and hands it to an employee who fills it right there at the stuffing machine. Before it’s stitched shut, the kid slips a little fabric heart inside, one they’ve rubbed, kissed, and wished on. And there you have it, the “Heart Ceremony.”

They dress it, give it a voice and/or a scent, pick out shoes, name it, and walk out with a birth certificate in its name. That’s the founding gesture of Build-A-Bear. Hold onto it as the atom of this whole thesis: a brand that industrialized a moment of attachment, then spent thirty years spinning out every variation of it.

2.2 The Products and the Average Basket

That Heart Ceremony is psychologically devious. Through the ritual, the toy itself gets personified: the child has “birthed” the plush and given it a name. You instantly reach a level of attachment that goes well past anything bought off a shelf at Walmart. And a different kind of attachment means different behavior, financial behavior included.

A family walks in expecting to pay the sticker price on the plush: twenty-something dollars for a standard model, around ten for the entry-level one, more for a large format. Except the bare plush is only a starting point. Dressing it runs $5 to $15 more (more if the outfit comes from a license), giving it a voice or a sound, $6 to $15, a scent, about $5, and it can keep going for a while: shoes, glasses, a bag, and so on, let your imagination run wild.

All of it is easier to justify once the plush has a name. By the time you reach the register, the $25 bear is worth $60. The kid walks out in awe of their personified, personalized plush, the parents walk out in awe of the company’s commercial strategy. The “build-then-dress” model is a machine for turning a token of affection into extra lines on the receipt. The added cost of the ceremony itself is fairly minimal for the company: that work would have been done at the factory anyway.

On top of the core products sits a whole range of variations. At the high end, licensed or limited-edition plush. They go for $40 to $75, with some collector pieces north of $100.

At the very bottom, the “Mini Beans,” small plush at around ten dollars, work as a loss leader and an affordable collectible (these are exactly the ones that recently showed up at Walmart). Between those two ends, the full range. The result shows up in a stat management likes to quote: average basket per transaction went from about $35 in 2012 to more than $53 in 2021, driven as much by the number of items as by unit price. The same visit now brings in half again what it did ten years ago.

Take all those plush types, multiply them out, add licenses, add dozens of different accessories, and you’ve described almost everything BBW sells.

Worth a special mention: the Bear Cave, a corner of the site for the over-18s, launched in 2019. Nothing NSFW, just licenses better suited to adults. The hardest thing I found is still pretty soft: a drunk bear (or at least one about to be), another crying blood, and a “you turn me on” t-shirt.

The ritual has also moved from the counter to the screen. You can build your bear online, on the “Bear-Builder.” It has to be a minor feature, because I tested it (purely for research, I promise), and “clunky” is being generous. It felt like a twenty-year-old Flash game running on Windows XP.

E-commerce has become, in the company’s own words, “its biggest store.” In practice it’s mostly a storefront that never closes, over-indexing on teens, adults, and collectors while the physical store stays focused on kids and families. To illustrate: when BBW added the option to record a voice message on the site, “Record Your Voice,” to drop straight into the plush, online sales grew double digits. Behind that double digit sits the chance to wish someone a happy birthday “directly,” do a gender reveal, send congratulations of any kind, and so on. To be fair, the digital bet was a bold one for a company that sells experiential, hands-on stuff, but it paid off (not without friction, as we’ll see later).

Once all of that is in place, the company “just” has to track sales in real time, push what works, prune what drags, while timing releases around the licensing and seasonal calendar (e.g., plush dressed in national colors for the soccer World Cup). That release cadence, synced to movie and event drops and orchestrated season after season, gives collectors and kids a reason to come back. Like a lot of retailers, the quality of the business model comes mostly from tightly managing its release strategy and its inventory.

This is the perfect moment to tell you a story that colors everything that follows in this deep dive.

2.3 Beyond Make-Your-Own: Wholesale and Mini Beans

Late 2023. The company is stacking records across all its segments and realizes the power of its brand has outgrown the make-your-own concept. Build-A-Bear is now Build-A-Bear. Might as well use the brand. So they create a line of plush that wouldn’t be self-assembled, stamped BBW, that the company could sell straight to retailers. The idea is to distribute plush, pre-made, in bulk, under the BBW brand (and the margins that come with it), on a regular basis.

That idea is what gives birth to the Mini Beans line in February 2024. Small pre-stuffed plush at $10, made to collect. The CEO lays out her strategy on a conf call:

“One of the reasons we launched Mini Beans was not just to put them in our stores, but as a proof point of the power of the brand to stretch beyond the make-your-own concept.” (Conf Call, Q2-24)

At first she sells them mostly in BBW’s own stores, but as the wins pile up, other retailers start taking interest within a few months:

“We are in the process of working with other retailers, not only here in the US but across the globe.” (Conf Call, Q2-24)

May 2025, the Mini Beans are available in other stores (Hudson, Applegreen, etc.). In the end, 2M units sold and a segment posting +30% YoY. The next quarter (Q2-25), Mini Beans climb to 80% YoY.

March 2026, management announces a partnership with Walmart to put Mini Beans in 1,500 of its stores. Naturally, it’s a collectible, so they roll out Walmart exclusives. The upshot: “a multimillion dollar wholesale order, now hitting shelves in approximately 1,500 Walmart locations across the U.S.” (Press Release, March 16, 2026)

What comes next, we don’t know yet. On the last conf call, the CEO pointed out there’d been no real feedback from Walmart yet. This Walmart partnership (and others with potential wholesalers) is one of the most important variables for the valuation. We’ll come back to it many times.

2.4 The Brands and Licensing

This is probably its least visible asset and, to me, its most powerful: a portfolio of more than 75 licenses (Pokémon, Star Wars, Toy Story, Harry Potter, NFL, Wicked, you name it). BBW has become, like LEGO, a meta-brand: a backdrop brand you can set almost any character on, and have the customer build it themselves if they want.

Each license brings its own audience, its own drops, and so its own reasons to come back. The downside is that it carries the main flaw of meta-brands: the loss of a key license. If Nintendo, Disney, or Sanrio suddenly stopped licensing their rights, the hit to BBW would be significant. In reality it remains win-win, and no major license has ever been pulled from BBW.

To limit that risk, BBW again did what Lego did: build its own brand. The Mini Beans are one example, all the more so since they also stack the meta-brand angle (a Hello Kitty Mini Bean, say). It owns other in-house IP like Promise Pets, Honey Girls, or Merry Mission, though none of it is as developed as the Mini Beans (and yes, we’ve seen crossovers like Promise Pets Mini Beans).

It has also tried its hand at other media. Since 2019, BBW has had its own studio, Build-A-Bear Entertainment, set up with Sony. So far the track record is thin:

• Honey Girls, a musical feature released in 2021 on streaming/DVD, based on IP that’s been around since 2015

• Glisten and the Merry Mission, an animated film released in November 2023, based on IP around since 2014.

Naturally, all of it got spun out across different media: radio, apps, video games, and so on (nothing worth mentioning). The box office stays modest, but profitability is beside the point here. The goal is to anchor the Build-A-Bear brand in the minds of kids and their parents. The company won’t put it this way, but the goal is to build a “mini-Disney” (to be fair, that’s the dream for nearly everyone in the sector). A universe of characters and stories you can re-adapt at will.

Its most recent release is an animated series on their YouTube channel, KABU. At last count, views topped 2 million and reportedly drove more than $1M in plush sales (yes, through Kabu Mini Beans).

All of it currently serves mostly to sell plush. Net capitalized entertainment assets are worth only $4.5M on the balance sheet. But what matters is that customers buy the plush, and they do, just not all of them the same way.

2.5 The Customers: From Kids to Kidults

At its opening in 1997, Build-A-Bear is a workshop for kids aged three to eight, planted in a shopping mall. The shift happens in the 2010s: the brand stops aiming at the child alone and goes after the teen, the adult who gives gifts, and the adult who collects. Today close to 40% of sales go to teens and adults, elegantly labeled the “kidult” segment. The company grew up with its audience.

This isn’t a BBW quirk: kidults now make up nearly 28% of the global toy market.1 BBW’s place in that segment, what weapons it plays with, and against whom, that’s the job of Part 6. For now, the point to keep is that BBW went from the 3-to-8 segment to the 3-to-45 segment, and the more time passes, the wider that segment gets.

Ages vary widely, and so does behavior. Let’s break it down.

The store is mostly for kids, their families, and young teens. The first driver is birthdays and parties: more than a third of revenue revolves around them. A birthday doubles as a showroom for every kid there. Among the twenty-odd children and their families who came for the party, some will come back, maybe even for their own birthdays, with other kids. The second driver is the store itself. The store is built to be its own display window, whether it sits in the middle of a mall walkway, inside a store that breathes childhood, or in a tourist destination.

Online, you shift to the near-adult teen and to adults. Both can be either collectors or there to give a gift. Hence the Bear Cave. The occasions are more varied: a new baby, the holidays, a graduation, but above all, Valentine’s Day. Bears appeal to lovers enough that Valentine’s Day became the single biggest sales day in the brand’s history in 2026.

The inevitable gift card is in the mix too, and it’s not trivial. As always, it adds a few accounting subtleties not worth getting into here. Either way, the revenue gets booked at some point, whether the card is spent or not. $6M in unredeemed gift card revenue last year.

Tying it all together is the loyalty program, the “Bonus Club” ($1 spent = 1 point). To join, you hand over some data, which translates into a customer base the brand owns outright to fine-tune its marketing and its offers. 8 in 10 Americans know the brand (I’ve been holding back the word moat this whole time, but come on... oops), so marketing serves more to maintain existing affection than to create it.

That marketing happens everywhere: owned content, licenses, email, and above all, social media. Close to a billion views when the company dropped its Baby Yoda plush. Strings of shorts racking up tens of millions of views each (some sponsored, some not), made by moms or young “influencer” couples. But this social-media marketing mostly serves the owned-store sales; for e-commerce, the advertising runs mostly online. Speaking of points of sale.

2.6 The Doors: Stores, Formats and Franchises

The doors are many, and the count is instructive. As of January 31, 2026, Build-A-Bear has 662 locations worldwide:

• 375 corporately owned, of which 333 in the US and Canada, 42 in the UK and Ireland.

• 178 are third-party operated. Think of it as grafting the BBW experience inside someone else’s destination: cruise ships, zoos, theme parks, and so on. BBW sells the merchandise and the turnkey experience, and the third-party operators run the place, pay the rent, and carry the risk. The vast majority of these locations are in North America/UK.

• 109 are international franchises. Here, a foreign operator pays for the right to open a store with the Build-A-Bear brand and products. These third parties run the place, pick the location, negotiate the lease, pay it, and so on. In short, once BBW grants them the right to use its licenses, they’re the ones making the decisions, taking the risk, and weighing down their own balance sheet. BBW sells them the merchandise and the turnkey experience (BBW’s only upfront capital outlay). It’s an extension of the BBW brand, so the choice of operator matters to BBW, all the more since one operator often runs several BBW stores in the same country.

As for size: “So we have formats that we can operate from 200 square feet to 10,000 square foot locations profitably.” (Conf Call, Q3 2024)

Within that range, a whole variety of formats:

• Traditional Mall Store. Full-format big footprint in the mall. It was dominant in 2015 (90%+ of the mix), now a minority (this detail will matter a lot in the next part);

• Discovery Format. A modernized version of the Traditional. Smaller, with a redesigned layout and better productivity per square foot. It’s the majority of the current mix (about 54%)

• Concourse shop. Around 200 sq ft in a mall corridor (no walls to rent). The idea is very flexible leases and, above all, almost no capital. The aim is mostly to create brand exposure cheaply and with high versatility

• Shop-in-Shop. A store inside another retailer. Typically a corner of a Walmart. BBW sells its own BBW products in its corner of the store. This is different from the partnership where Walmart itself sells BBW products on its shelves.

• Tourist destination. The name says it. BBW is an experience, and it overperforms where people are already buying experiences. It’s been one of the DTC drivers for a few years now.

• The multi-level flagship. Just one of these: Orlando’s Icon Park, designed as an attraction in its own right in a tourist zone. It opens H2-26.

• Seasonal. This one’s for the short term in specific situations. Leases, usually small, from 2 months at the shortest to 18 months at the longest.

BBW doesn’t own the walls of any of its stores: it leases them all. So it has to negotiate its leases. But a brand that pulls in customers is a brand that pulls in landlords too. One of the big variables that reshaped BBW is its leverage in negotiating leases: “particularly in the post-covid environment, where we renegotiated practically every single lease. And I think that utilizing the power of the brand has leverage, and I believe we’ve done a very good job at that” (Conf Call, Q3 2024)

Let’s unpack this a bit, because it’s one of the biggest variables on the balance sheet. Most leases index rent to activity: a base rent plus a variable piece that kicks in beyond an agreed annual sales threshold. Some drop the fixed part entirely, and the landlord then takes only a percentage of the store’s revenue. When a location slows, the rent slows with it. Many leases also carry a sales-based exit clause, over two windows: years three to four, then years six to seven, exercisable depending on the case by the landlord or by BBW.

In North America, the term was deliberately shortened to realign the fleet faster: that flexibility is one of BBW’s strengths. But it has a price: if commercial rents are higher at renewal, BBW will pay more in the vast majority of cases. In the UK and Ireland, it’s both simpler and more constraining: leases reset every three to five years, aligned to the market but “upwards only.” So base rent can rise but never come back down, even when the market turns. Remember, this is a minority of the company's directly operated stores.

Alongside, the digital side: buildabear.com, the Bear Cave, the marketplaces; and since a web order is picked up in store as often as it’s shipped, click-and-collect in the UK and Ireland, in-store pickup in the US, every store also doubles as a mini-warehouse.

Since this spring, the shelf space of 1,500 Walmarts all at once. And, at the far end, outbound licensing, which puts the brand’s name on shelves it will never run. E-commerce depends on the stores logistically, but the physical stores also depend on e-commerce: “Most of our guests, from what we can tell [...] have a tendency to go online first [...] find a store, plan a visit, plan a party - [these] web pages are our largest visited web pages. So we know that the guests are using our buildabear.com as a source of information and planning for their store visits.“ (Conf Call, Q2 2025)

That’s the present; for the future, the trend is clear: 64 net openings last year, only 7 of them owned, 40 with partners, 17 franchised. They’re guiding for around 50 new locations in 2026. Worth noting the international footprint went from 19 to 37 countries in two years. Enough to soften the blow of the gloomy US consumer. For owned-store growth (so North America + UK + Ireland), the market is probably close to saturation (management will obviously never put it that way). Growth in that segment very likely won’t come from store count.

2.7 The Unit Economics

$530M in revenue last year, for net income of $52M. Why only $52M is the job of Part 5. The company files that figure under three segments:

• retail, stores and web combined, 91.7%

• commercial, wholesale and outbound licensing, 7.3%;

• international franchise, almost pure royalties, 1%.

This revenue split hides a lot of important things, but we’ll keep that, too, for Part 5. For now, let's keep it descriptive. To set the scale of one unit: a North American store runs around $1.2M in sales a year, about $518 per square foot, at roughly 56% gross margin.

The company’s economic strategy comes down to four pillars:

1. Organic sales growth by widening the audience (collector kidults in particular) and by raising the basket (price and units per transaction)

2. Location expansion, mostly asset-light internationally through franchises. Last quarter there were a few “closures,” but mostly format upgrades or relocations.

3. Wholesale and brand licensing. Wholesale = selling the pre-stuffed products in 1,500+ Walmarts (note also a Los Angeles showroom). Brand licensing = letting outside manufacturers and distributors sell products tied to a brand BBW owns. This is different from a franchised store. It’s the same as when BBW pays a royalty to Nintendo to make its own Pokémon plush, except here BBW is the one getting paid for others to make and sell its brands as plush. The main brand involved here is Honey Girls.

4. Gifting and personalization. The idea is to push the brand beyond the childish in-store experience toward the gifting occasion, to stretch the moments of “consumption.” If I drop the management-speak filter: get adults to buy more often.

I’d file 1 and 4 under intensive growth (more value per customer in mature domestic markets), and 2 and 3 under extensive growth (generating other revenue streams). But we’ll get into the numbers later.

There’s only one thing left to see to understand the business: behind the scenes.

2.8 Behind the Scenes: Supply Chain and Operations

You won’t be surprised to learn that Build-A-Bear makes nothing itself: it buys three-quarters of its plush from five suppliers, based in China (a little over half) and Vietnam (the rest), after patiently cutting a China dependence that ran above 90% before 2020. The diversification only lasted so long: China and Vietnam were among the countries hit hardest by tariffs (we’ll come back to it in detail, count on it).

It doesn’t manufacture, but it answers for everything: toy-safety compliance, independent lab testing, factory social audits, and so on. After the various children’s-toy scandals of recent decades (look them up, it’s worth it: little beads, say, whose coating turns into GHB once swallowed), these constraints weigh on gross margins (but it’s a fixed cost you can’t optimize away, so be it). Its track record is reassuring: only 6 minor recalls, one of them in early 2026 (nothing significant).

As for logistics, it owns a 350,000-square-foot warehouse in Groveport, Ohio, plus three other centers (California, England, and Shanghai) run under third-party management. Around 1,200 full-timers and 4,300 part-timers depending on the season to keep all this machinery running (excluding franchises/partner-operated). A single store is unionized (end of 2025).

The company tends to its reputation as an employer, and it shows: on Fortune’s 100 Best Companies to Work For list for a decade, through 2018, and singled out again by Newsweek in 2026.2 That’s anything but trivial. Value creation is a human business: happy employees are more productive employees, and that carries through to customer satisfaction.

For the doubters, the Fortune 100 Best Companies outperformed the S&P 500 by 2,000% over the last 27 years (note that BBW was literally part of that light-blue line for 10 years).

2.9 Competition

Who are all these happy employees fighting? The 10-K distinguishes two types of competitor (I’ll add a third):

• The typical toy seller: Ty, Jellycat, Hasbro, Mattel, Lego, Amazon, and so on. The publicly listed ones are Hasbro, Mattel, Pop Mart (the Labubu people), and Funko (thank Private Equity and its healthy appetite for this kind of company for the thin list of comps). Call it “toy competition.”

• Next, the company also competes against everything “that fights for leisure time and budget.” Think movie theaters, restaurants, theme parks, zoos, and so on. Call it “experience competition.”

• Finally, the company competes against every business fighting for the best spots in malls and tourist locations. Call it “location competition.”

Three circles of competitors, and not one of them really doing the same thing. Here’s a broad map of the sector’s competition. It’s neither exhaustive nor the final word.

As for toy and location competition, the close competitors number in the hundreds. On the experiential side, direct competition is more limited but also more direct: a family can hit a restaurant and then the BBW next door. The odds they go to the Lego store and then to BBW are much lower. For location competition, it can be against just about anything.

If you overlay the circles (experience, toy sales, location) and factor in size/geography, the most relevant comp in my view would be Lego. But Lego isn’t publicly listed. The Hasbro and Funko comparison is too limited on too many variables. The least-bad comp is still Mattel: similar margins, cash flow generation, and valuations (around 6x EBITDA on both sides). But operationally, they’re still pretty different animals.

The real comp for BBW is BBW itself. That’s the one I’ll use for almost all of the rest of this deep dive. Beyond the too-distant resemblances and the thin public information on the candidates, the decisive argument is that a competitor can also be a partner.

The best example of all this is the BBW store partnered with Hello Kitty. Sanrio, a competitor in the sector but also a longtime BBW partner, opened a store built on the Hello Kitty IP but recentered on BBW’s make-your-own experience. BBW doesn’t own the concept, they only invented it. Anyone can sell make-your-own plush and slap their own IP on it. Sanrio could have done it alone. Instead they chose to do it through a partnership with BBW. Why? Because BBW has outgrown its own concept; it has become a brand in its own right, one it can be operationally and financially worth teaming up with. It shows BBW’s unique place in the ecosystem, and how a global competitor (in the toy segment) can be a partner in the specific. Since 2026, two more stores partnered with Hello Kitty and Friends, directly operated by BBW, have opened.

To sum up, BBW is:

• A brand that industrialized a moment of attachment and resells it at every price, from $10 to $100+;

• Available at every age, from the crib to the office.

• Through every door, from the store it owns to the Walmart shelf, from the cruise ship to the screen;

• Under almost every license, from Pokémon to the NFL, all the way into its own films.

• Spreading across the world through franchises that cost very little capital.

• Whose growth comes mostly from franchises/partner-operated and from the commercial segment (wholesale & licensing).

To sum up the summary, BBW is a very good idea, building your own teddy bear, pushed to the end of its combinations. Add a dose of collectibles and licenses, mix it all to taste, and you’ve got today’s BBW.

Now that we have a good picture of the company, one thing remains to describe: the macro backdrop BBW operates in.

Hedge funds pay tens of thousands for a thesis like this, then keep it to themselves. Want a good chunk of that, regularly and for free? Subscribe.

Subscribe now

3. The Macro Backdrop

3.1 The Context: Consumer Sentiment

You probably know this already, the macro backdrop is... complicated. Gloomy, the US consumer would say. In May 2026, the University of Michigan consumer sentiment index drops to 44.8, the lowest in the whole series3, which is nearly 50 years old. It bounces to 49.5 in June but stays 19% below where it was a year ago.4 The Conference Board tells the same story its own way: 93.1 in May, down, with two in three households saying they’re cutting back.5

Digging through the data, you can date the trigger of this record low: in spring 2026, the flare-up between the United States, Israel, and Iran sends crude soaring. Gasoline tops $4 a gallon for the first time since 2022, a dollar more in a single month (+35%), the most brutal monthly jump in decades. Worse than Katrina and worse than Ukraine.6 On an already tired consumer, it’s one tax too many. My take is that the uncertainty tied to AI, and to a world changing faster than probably any time in human history, helps make the consumer a little more uncertain and uneasy.

What’s objective is that the safety cushion is thin. The household savings rate fell to 4% in Q1-26 (vs 6.2% two years earlier).7 Enough to absorb one surprise at most. Meanwhile, credit-card debt is brushing its record, at $1.25T.8 Delinquencies are rising, concentrated in the most fragile borrowers. Nearly 1 in 5 student loans are now delinquent.9 In short, the American consumer has few reserves and a lot of debt.

Okay, that’s the macro backdrop painted Francis Bacon style, but mood and behavior aren’t the same thing. The gap between the two has never been wider.

3.2 The Behavior: How Consumers Actually Spend

Retail sales rose 0.5% in April and 5% year over year.10 Unemployment holds at 4.3%, with positive job creation and wages up (3.4%).11 Clearly a different picture. And that fear doesn't stop them from spending. For now, they're at the stage of picking and choosing.

That trade-off is clearly directional. In January, services spending rose $105.7B while goods spending fell $24.6B.12 The signal is in the detail. Apparel and footwear lose ground, autos (and gasoline, obviously) lose even more. To be fair, this trend has been around since Covid, but 2026 saw it accelerate. 75% of Americans say they get more value from an experience than from an object, and more than half feel they own too much stuff.13 So when confidence drops, or budgets tighten, the object gets sacrificed for the experience.

This shift claims victims, to the point where the retail picture is starting to look like a graveyard. More than 8,000 chain stores closed in 2025, and the same number is expected in 2026.14 Retailer bankruptcies keep coming: QVC (nearly $7B in debt)15, Eddie Bauer (175 stores)16, Francesca’s (more than 400 stores liquidated)17, and I could go on for a while. What they have in common: sellers of undifferentiated goods and apparel. Meanwhile, US mall traffic is climbing (3.5% for open-air centers vs 2.2% for enclosed malls)18, the real winners being the ones filling up with experience and dining (I know, dining is an experience, but I mention it because its share of the experiential is very significant).

One big uncertainty remains, this time on the corporate side: tariffs. On February 20, the Supreme Court struck down a good chunk of the tariffs imposed by the Trump administration, opening the door to some $170B in refunds.19 No surprise, the administration hits back at once with a 10% tariff on everything under Section 122.20 For our purposes here, China stays taxed around 17.5% and Vietnam around 10%.21 For anyone who imports what they sell, which is to say nearly all of American retail, that cost line is a high-variance variable.

To sum up, a hesitant consumer, but one still spending, dropping the object for the experience. On top of that, import costs you can’t model with enough certainty at 12 months.

For BBW, which sells both experience and physical goods, the picture isn’t obvious. The string of records is clear, but the weaker quarter, the lowered annual guidance, and the tariffs add uncertainty.

The “truth” is in the digging and in weighing the variables against each other. All the better, otherwise the opportunity, if it is one, probably wouldn’t exist. The digging starts with understanding BBW’s history and how its business model was built and evolved.

4. Build-A-Bear’s History, Brick by Brick

4.1 A Study in Pink: The Founding and First Boom (1997-2007)

The first brick is the most important: a whole new category. Maxine Clark, a former Payless executive, opens the first workshop in the St. Louis mall in 1997.22 The idea reportedly came from her neighbor, a ten-year-old who was hunting in vain for a trendy stuffed animal and who, with childlike originality, supposedly said: “why don’t we just make it ourselves.” Clark tests the hunch on a panel of kids, pays a design firm to refine it, and sniffs out the essential: a plush costs $5 to make and can sell for $50.

The first store is a smash: the line spills into the parking lot, $400k of plush in a few months, up to $600 in sales per square foot.23 Clark was one of the pioneers of a new genre, “retailtainment”: not quite a store, not quite an attraction. The word itself wasn’t even coined until two years later.24 The potential is obvious to Clark, and soon obvious to everyone. The next step has to be to lock in the concept fast by scaling.

The second brick, then, is scale, and she executes it almost perfectly. $5M raised for 11 stores in 1999, then 100 in 2002, an IPO in 2004, and soon a Build-A-Bear in nearly every American suburban mall. In 2007, revenue peaks at $470M. The concept is good, but the timing was perfect: baby-boomer parents who liked spending on their millennial kids, who were crazy about personalizing everything. Hard to imagine a better start.

Competition shows up anyway. On paper, after all, the concept isn’t hard to recreate. The Bear Factory, launched by Hamleys in the UK25, clones the model fairly “closely.” Close enough to try to register the name “Build-A-Bear” (already taken, of course) before ending up bought by BBW itself in 2006 for $41M.26 That’s the origin of its sizable UK footprint, and in hindsight it was objectively the best call. The brand didn’t yet have enough weight to take on competition abroad, and the acquisition let it export the brand into a country culturally very close to the US, so probably just as receptive. Another notable direct competitor, Teddy Mountain, launched in 2003, climbs to 120 stores27, then abandons retail for wholesale kits for zoos and parks after 2008 (for lack of profitability).

The irony is that even its inventor stumbles on it: in 2005, BBW, probably drunk on its own success, thinks it holds a universal recipe and extends it to dolls, the “Friends 2B Made,” which it shuts down by 2009. Those 4 years may seem harmless, but not if you read the filings from the time. F2BM very quickly performed well below management’s expectations, and an efficient capital allocator would very likely have killed the project much sooner. The seeds of the future disaster were already there, even visible: Clark stayed locked into the strategy that had grown BBW, and was very resistant to change despite the feedback, sometimes obvious, from reality. The subprime crisis would put all of it on display. The idea to keep from BBW’s first decade is that the concept can be copied, but not everything around it. And it’s the “around it” that creates the value.

4.2 The Valley of Fear: Crisis, New CEO and Turnaround (2008-2019)

The BBW of 2007 was built exclusively on malls, and those were about to die en masse. The subprime crisis, sure, but also e-commerce taking off, the iPad in 2010 concentrating kids’ (and some parents’) leisure time onto a single device. Revenue falls from $470M to $380M, and fiscal 2012 ends in a $49M net loss, a good half of it from writing down the British goodwill. The valuation drops to $50M. It was priced as dying, and to be fair, it was.

Faced with losses like that, the board had to act. In 2013 (only in 2013, read the conf calls, it’s almost painful), Maxine Clark, whose official title was “Chief Executive Bear” (not gonna lie, I find that pretty cringe), steps down. The board goes looking for a turnaround specialist, Sharon Price John. Before the shoe company Stride Rite, she spent years at Hasbro and Mattel, keeping licensed toys alive (Star Wars, Disney, and so on). To fix Build-A-Bear, the board bet on a brand-and-licensing specialist.

John arrives and spends weeks analyzing the stores, line by line. Her read is clear: the business is broken, but the brand isn’t. So she puts on her turnaround hat and closes 50-60 money-losing stores in two years. She relocates some, shrinks others. She’ll say what stunned her most at the time was that 5% of sales were online, at the height of e-commerce’s golden age. She relaunches e-commerce right away and puts her omnichannel strategy in place. A few weeks to map every hemorrhage, a few months to stop them, and a few years to repair the foundations. Then she could prepare for the future. Clark, the former CEO, was locked into her strategy centered on the in-store experience and growth through store count. It was her baby; shrinking it or significantly reorienting the strategy meant contradicting years of decisions, with escalating commitment. A new CEO was the only cure.

John makes another observation: the kids of 1997 have become adults. According to the stats, they’ve even become nostalgic. And nostalgia can be bought. She steers the sales and marketing strategy beyond young children. 8 years later, the teen-and-adult share has gone from a trivial slice to 40% of sales in 2022, above the sector average. And that nostalgia keeps growing among the new adults. It changes the business model deeply: the customer is won in childhood and stays won while growing up, switching channels (now a collector, he gives one to his boyfriend or girlfriend, then brings his future kids back, and so on). So over the years the company becomes a multigenerational annuity.

To feed that broadened audience, John triggers her licensing reflex. Build-A-Bear becomes the canvas the biggest names land on: Marvel, Disney, Pokémon, Star Wars, the NFL. The announcement of a Grogu bear, the Mandalorian’s “Baby Yoda,” pulls in nearly a billion impressions in a few days, and the bear sells out in two hours. Every major license is a door that closes on the copycats: a kiosk can stuff a bear bought in China, but it can never dress Darth Vader (legally). The company pushed the logic further and makes its own characters, through Build-A-Bear Entertainment: films, in-house heroes you don’t have to rent and whose margin you keep entirely. That segment stays relatively marginal for now.

Meanwhile, BBW leaves the dying malls for the places the customer actually is, kiosks, tourist zones, cruise ships, up to 35% of sales outside the mall by 2022. BBW is a brand businesses want, and it finds its place just about wherever it likes, often with strong negotiating power. Alongside, it opens a back door for adults, the Bear Cave.

Each of these initiatives was one more brick, pushing the would-be imitator further away. The wall becomes a moat in its own right, and each additional year fortifies that moat.

There was still one big misfire. In July 2018, a “Pay Your Age” promotion (pay for your bear at the price of the child’s age, sometimes called “Count Your Candles”) draws crowds so dense the stores have to close for safety, 1,500 people in the Manhattan store alone (I did the math, that’s 6.3 people per square meter, a critical-level crowd density). The year ends in a loss. It was a very particular moment that gave away two very important things:

• The brand was powerful enough that its demand spilled past its own walls: the moat was even wider than assumed.

• But the business was still fragile enough to lose money anyway: the rebuild was far from finished.

They’ll keep running this promotion in a more controlled way (it still exists), and each time it remains one of the best ways for the company to get itself talked about (directly, or through its customers on social media). Above all, the promotion requires people to sign up for the Bonus Club loyalty program, which not only builds a bit more loyalty but, more importantly, captures their information (email, phone, age, child, and so on) to run extremely targeted ad campaigns. That’ll matter later.

4.3 The Adventure of the Second Stain: Covid and the Record Run (2020-Today)

Then comes the trial by fire, Covid, the lockdowns and store closures. On paper it’s the worst-case scenario for a business that makes the product on site. And this is where the successive bricks John laid over 7 years, without ever thinking of a pandemic, literally save the company:

• She renegotiates a large part of these leases to lighten the balance sheet substantially.

• She goes all-in on the e-commerce she’d built up over the previous years.

• She reorganizes the logistics of her click-and-collect, maximizing efficiency while using the entire network of empty stores.

Customers order their bears online, pick them up, and BBW survives. The year takes a $23M loss. The next year, the company posts $47M in profit, its best year. BBW reaped the fruit of those hard years of rebuilding John put in place, at what looked like the worst moment in the company’s history. It not only survived, it came out of Covid stronger than ever. The nostalgia economy that followed the end of lockdown finished transforming the company.

BBW survived the subprime crisis and an ill-fitting operating strategy; it took nearly 10 years to rebuild, which is what let it survive its second-biggest crisis. The company went from a broken business (but a working brand) to a very good business and a brand at its peak. I read all these changes quarter by quarter, conf call by conf call, and I’ve rarely seen a turnaround executed this well. Shareholders at the time expected a 2-to-3-year turnaround, but John knew getting the company back afloat would take much longer, and made all the hard decisions to create long-term value.

Also worth noting in this period the influence of activists: Cannell (Cannell Capital LLC), and above all Kanen (Kanen Wealth Management LLC), who helped turn a very good business into a very good business that rewards its shareholders. It was in fact in Kanen Wealth Management LLC’s 13-F that I discovered Build-A-Bear, and took part in what would become one of the great shareholder runs of the era. I actively encourage you to keep an eye on their largest positions regularly.

From there, the machine is running. The company strings together years of record revenue. Slight stagnation in net income in FY24 and FY25 (we’ll see why later), but EPS keeps setting records thanks to buybacks. The company has paid dividends for 2 years. The moat now expresses itself in two “currencies.” Price, because an inimitable brand makes you pay for its name (fine, it’s still a toy retailer, we’re not talking ASML pricing power). And capital, because third parties compete for the right to operate the banner to the point of funding its growth: 57 of the last 64 net openings were on a third party’s balance sheet, with franchise payments. You don’t rent out merchandise.

That’s how the moat was built, brick by brick:

• a new category,

• a scale that absorbs its clones,

• a near-shipwreck that proved the brand outlived its business,

• a rebuild that filled the hole and then, on top, widened the gap,

• gone multigenerational,

• licensed from the biggest names,

• fed by its own licenses,

• operated in part by others who pay for the privilege.

The consequences of this moat for the present and the future are the more or less direct subject of the parts that follow. But I want to quote the former CEO, which gets at the essence of those consequences: “we believe that we have the opportunity to appeal to more people, in more places for more products and more occasions.” (Conf call, Q3-25)

I’d call that a halo effect.

We’ve finally covered the qualitative. Time for the numbers.

Refer a friend

5. The Financials

Time to keep several promises I’d put on hold.

Before we start, two important clarifications:

• I use the company’s calendar: the fiscal year ended January 31, 2026, is FY2025. The quarter ended May 2, 2026, is therefore Q1-26. Aggregators like Tikr and the rest shift everything by a year.

• A word on the calendar, because it’s a trap: the fiscal year ends in late January and runs one extra week roughly every six years (last time in 2023), which inflates the year-over-year comparison base a little. As things stand, for the company, we’re in Q2-26.

A little spoiler up front: no important accounting subtleties and no shady behavior from the company.

5.1 The Income Statement

FY2023 had 53 weeks. Pro-rating it down, 2023 net income is $51.8M on a 52-week basis. Don’t forget it.

What jumps out is the stagnation in net income around $52M since FY23, despite 10% revenue growth over the period.

To understand it, you have to pull the accounting thread.

Gross margin. One quirk for the DTC (most of revenue): it includes merchandise costs, but also store occupancy cost (including depreciation and impairment), royalties, tariffs, warehousing and distribution expenses. All wages sit in SG&A (which can flatter gross-margin comparisons with competitors). It rose 1.4 points over the period, and in fact it has risen every year since Covid. Note too that tariffs already had an impact in FY25. First point, the company has an ability to raise prices, small but steady, in a period when inflation was running above its historical levels. That’s a real edge for a retailer with thousands of competitors (from the restaurant to Hasbro). All well and good, but above all it doesn’t explain why net margins stagnated.

The tax rate. It’s always above 22% from 2022 to 2025 (2021 was a Covid-related exception), except in 2023 when it’s 20.4%. That gap comes from a one-off tax benefit tied to a return to profitability in the UK, which allowed past losses from UK asset writedowns to be deducted. Had the tax rate been 22% in 2023, net income would have been $51.7M and would have followed the same ordering as revenue (a record every year through 2025). Nice, but adjusting for tax, net income stagnates even more around $52M.

Pre-tax income. A clear improvement from 2022 to 2023 (+7.2%), a modest one from 2023 to 2024 (+1.2%), and a trivial one from 2024 to 2025 (+0.1%). The stagnation shows up here too, so the explanation has to be in SG&A.

SG&A. Simple: it has offset gross margin by a bit more every year since 2023. It went from 39.3% in FY22 to 43.3% in FY2025. +1.3 points a year on average. So that’s where the problem comes from, but SG&A is too big a bag. We need more granularity, and for that you have to look at Note 15 of the 10-K:

• Selling expenses directly tied to the segments: the majority, roughly the costs of the stores (minimum-wage increases, promotions, and so on) and of DTC e-commerce. These went from 28.0% in FY23 to 30.2% in FY25 (as a share of DTC revenue). About 60% of SG&A.

• Unallocated corporate overhead: basically, SG&A expenses not attributable to a segment. These went from 13.8% in FY23 to 15.4% in FY25 (as a share of total revenue). About 40% of SG&A.

Let’s focus on the total contribution of the 3 segments, that is, how much each segment brought in subtracting only the expenses specific to each (leaving aside the costs common to all 3): you see an increase of $10.7M (from 137.3 to 147.9). This is the metric management uses to make its capital-allocation decisions. I want to be sure I’ve been clear enough, so let me restate it. From FY24 to FY25, the contribution of the 3 segments rose $10.7M (each segment’s revenue minus its own costs). Over the same period, corporate overhead rose $10.5M.

Segment growth was eaten, within $200k, by the growth in head-office costs, that is, everything not specific to a segment (pre-tax lands at +$0.1M, so the rest is trivial). I haven’t seen this mentioned in any of the analyses I’ve read, even though management has spelled out that it decides based on these numbers.

To sum up:

• From 2023 to 2024, net income didn’t move because of a lower tax rate in 2023.

• From 2024 to 2025, net income didn’t move because management reinvested all of its operating gains into its own structure.

The question left is whether these costs are a new cost floor or a transitory investment.

The market seems to price the first, but I think the reality is more nuanced. First point, they’re opening their new flagship store in Orlando in 2026. That store has been in the works for years, and part of these costs very likely corresponds to one-off costs tied to preparing the opening. Second, management has spelled out that it hired talent that falls into this corporate overhead. That talent is a form of investment. It may be a new floor, but we can expect a return on the investment in that talent. For the rest, it's speculation, but it’s quite likely that part of these costs simply becomes a new floor.

That leaves the very bottom line, earnings per share, which reconciles everything. It compounded at 8% since 2021 despite the flat net, because the company bought back about 19% of its shares in 4 years. We’ll come back to it in the management part.

5.2 The Segments: Two Companies, One Multiple

Here too we’ll put ourselves in the same decision frame as management. It focuses on contribution margin by segment (the segment’s gross margin, minus the selling expenses directly attributable to it).

DTC (stores plus web) puts up a 25.9% contribution margin and grows about 6% a year. A good retailer, capital-intensive, paying rent, salespeople, and inventory. Commercial (wholesale, outbound licensing, entertainment) puts up 53.1% and grows 23.5%. Twice the margin, four times the growth.

Where does the gap come from, when the two show roughly the same gross margin (DTC 56.1%, commercial 55.6%)? From everything commercial doesn’t carry, which is a lot. No lease, no sales team, minimal marketing, and above all zero incremental capital: it’s the partner (Walmart, the zoo, the cruise ship) who runs the point of sale and carries the inventory, or the licensee who makes and sells it themselves.

The royalty and the wholesale drop almost naked to the bottom of the segment. DTC, on the other hand, pays the whole list above for every dollar it takes in. Result: commercial is 7.3% of revenue but 14% of total segment contribution, and above all, it alone provided about a third of the contribution growth between FY2024 and FY2025. On 7.3% of revenue. That’s enormous. There’s the engine I mentioned in the short thesis, in numbers.

Let’s temper that a bit.

One. This is still a fairly recent cruising speed, and it varies with management’s choices. In FY2023, commercial contribution topped out at 36% ($9.2M out of 25.4), weighed down by $4M in direct costs (the cost of Build-A-Bear Entertainment content and a film released in 2023). The 53% of FY2024-FY2025 is the real pace, but a studio that reinvests can pull it back down. One to watch. I think management won’t take risks on that side until they have a bit more visibility on the current risks (the focus of the next part).

Two. A pure license would put up a 90% gross margin. At 55%, it’s mostly the wholesale taking over. The royalty layer on top remains a minority.

Three. Above all, contribution margin excludes corporate overhead (whatever isn’t attributable to a specific segment). In reality, both engines (DTC and commercial) “free-ride” on the same structure and the same brand, which come in part from that $81.5M of corporate overhead.

Four. Timing. This segment adds $3 to $4M of contribution a year. The headwind (tariffs + overhead) takes out 10 to 16. The engine is still too small to cancel the headwind in the short term, but it improves the quality of the mix, well hidden behind the flat net income. It’s a non-problem for a shareholder who digs and who sees this 7.3 P/E as an opportunity. If the trend holds at 23% (the current pace), commercial reaches a third of group contribution in 7 years and parity with DTC in a decade.

The March 2026 Walmart wholesale (1,500 stores) is a wildcard, maybe the wildcard (the old Walmart partnerships like a little BBW stand inside the store aren’t counted here). On assumptions of $15 to $30M in revenue, with a margin heavily shaved by Walmart, say 35% rather than 53%, that’s $5 to $10M of contribution, half the current commercial segment from a single partner, with a template that can extend to other mass distributors. Keep this in mind, because it’s one of the most important variables for the bull case.

5.3 The Balance Sheet and the Flows: The Capacity to Wait

Time to prove that the balance sheet pays you to wait.

The simplest fact, and the most important: zero borrowing. No financial debt on the balance sheet, at any period close. The only thing that resembles it is $127M of lease liabilities (the fixed minimums capitalized under ASC 842), and on top of that, part of the real rent stays variable, indexed to store sales and expensed as it goes: when a location slows, part of its rent slows with it. A cost structure that breathes with activity, set on a debt-free balance sheet. There’s the starting point of this capacity to “hold out a long time.”

Next, of course, the cash. BBW isn’t sitting on a war chest. $26.8M of cash at the end of FY2025 is modest. It’s a leftover from the Kanen period, which convinced management to spend the $75M of cash that was lying around at the time on buybacks. The fortress is in the cash-flow generation and the absence of debt, coupled with an undrawn $40M revolver, signed late 2025 and running to 2030. Just in case. In case of what, I don’t know. It was raised from $25M to $40M in 2025. I won’t say it’s a sign of potential M&A, but the increase is curious all the same. The stash itself is thin, and will stay that way as long as the company doesn’t see financial trouble ahead. BBW’s strength is in its flow rather than its reservoir.

Here’s the flow: operating cash that swings between $47 and $65M depending on the year. Note that the $47M low is due to a tariff-related inventory pre-buy (working capital swallowed the rest). CapEx climbs from $8 to $25.5M with the pace of openings, IT, and above all the Orlando flagship. FCF runs around $40M, excluding exceptional CapEx.

What does this flow go to. Returns to shareholders, all of it. From 2021 to 2025, the company generated $167M of FCF and distributed $173M to shareholders ($108M of buybacks and $65M of dividends). The small $6M gap was drawn from cash. $173M returned to shareholders in 5 years, and the company is currently worth $392M.

Of course, if the company needs this flow to deal with a tough situation, the reallocation will happen instantly. As for investments, that seems unlikely to me. Management has shown little appetite for them, and since the Kanen episode, the company is very focused on returning value to shareholders (given how the stock has behaved since, you can see why).

As long as the machine holds, and it holds comfortably for now, the arithmetic of buying back at 7.3x is almost absurd, in a good way. There’s currently $47M left authorized on the program. At $31, that’s about 12% of the float. A company retiring 5 to 6% of its shares a year, at 7.3 times earnings, near its low, mechanically turns every year of waiting into value per share. Time is the ally of good businesses.

A few special mentions on the balance sheet and cash flows, just in case:

• The roughly 18% inventory increase in FY25 is tied to tariff-related pre-stocking. Beyond that, it’s not operationally informative.

• Receivables, up 33%, come from the $13.2M tariff refund booked as a receivable from the government, and partly from the rise in wholesale/commercial, which naturally carries B2B receivables where DTC collects cash/card directly. Nothing operationally significant beyond that.

• Only $0.4M of cash is restricted out of the $26.8M.

• When I say no debt, I mean financial debt. The lease is a form of debt, but it’s built directly into the business model. The whole analysis of the capacity to wait rests mainly on financial debt and FCF, so leases are barely relevant to the reasoning.

• The significant rise in net PP&E in FY25 is primarily driven by leasehold improvements on new store openings and the ongoing Orlando flagship buildout.

• The $70.9M of property, plant and equipment includes the wholly owned distribution center (valued at $31M a few years ago).

• No goodwill, no defined-benefit pension plan, and no material litigation or significant provision.

• The move in cash from $44.3 to $27.8M (from FY23 to FY24) corresponds to some buybacks and a special dividend. Everything for the shareholder. Thanks, Kanen.

5.4 Return on Capital (ROIC)

For a retailer like BBW, the most faithful metrics would be everything that touches operations directly: comparable store sales, ATV, and so on. Management has unfortunately become stingy with that kind of information over the last few years. So we’ll have to rely on accounting measures, biased as they are.

In two years, ROIC went from 33.1% to 23.6%. Yet the company improved operationally. So either the P&L is lying or ROIC is. Here it’s ROIC, more precisely the denominator. About half of invested capital comes from ROU (right-of-use) assets (31% from NWC and 19% from PPE), basically the accounting capitalization of operating leases under ASC 842. When BBW renews a lease at a higher rent or signs a longer contract, the standard mechanically books it as an asset. To sum up, the denominator jumps while not a dollar of capital is deployed.

When you control for that variable, you still land on a ROIC dropping from 62.5% to 45.5%. So it’s even worse? This time, it’s the numerator lying. FY25 EBIT absorbed about $7M of IEEPA duties (out of $11M total tariff-related costs for the year), struck down by the Supreme Court in February 2026 and being refunded to BBW. A cost the year carried without it being permanent. Adding it back to EBIT, you normalize ex-lease ROIC around 50-51%. As for inventory (+18% in FY25), that’s growth working capital tied to Commercial (+53% over two years), so it’s real capital that belongs in the denominator.

We’re still at the lowest of the last 5 years ex-leases. Here the “culprit” is management, or rather these investments via CapEx (+$7.2M in 2 years). First, the Orlando flagship, weighing on CapEx for 2 years with a sharp acceleration over the last few quarters and an opening planned for H2-26. Then the corporate overhead (mostly FY24 to FY25), weighing on the numerator and corresponding partly to investments and partly to inflation. I can’t strip them out, because they’re expenses directly tied to operations, but for now their impact weighs on both parts of the fraction simply because of timing; if these investments pay off, they should raise the numerator and lower the denominator (mechanically for the Orlando flagship) by H2-26, but probably mostly from FY27 on.

To sum up, ROIC is deteriorating, but a large part is tied to accounting artifacts, and the rest to operational timing. Someone who takes the accounts at face value without really diving in might just see a business whose revenue grows without growing its net income, and that’s getting less and less efficient. BBW’s accounts have screened very badly for 2 years. But a close look at these metrics shows a company whose engine of growth and value creation is gradually reorienting, that holds up relatively easily against a tough macro backdrop, and that should soon reap the fruit of its recent investments.

5.5 The Latest Quarter: Q1 2026 Results

5.5.1 The Quarter Itself

A little detour through the thirteen weeks that ended May 2, 2026, the ones where the stock stumbled. Here we’ll only put numbers to the situation; the full analysis is the next part.

Revenue falls 2.4%, but the information is in the mix: retail loses 5.1% while commercial gains 43.6%. It’s the mechanism shown in 5.2, but sped up. The asset-light model offsets the retail weakness. Note that the Walmart wholesale isn’t yet at its full potential. The coming quarters will say more about that.

Consolidated gross margin gains 700 bps, but that’s essentially an accounting artifact tied to the tariff strike-down and the $7.0M refund booked straight into cost of goods on the quarter. Cleaning all that out, you land on a 58.2% gross margin, 140 bps more. Despite tariffs continuing into 2026, the company raised its margins (after stripping out the refund impact; note all the same the favorable comp effect tied to “anniversarying lower prices last year”).

SG&A doesn’t cheat: +310 bps, but I’m saving it for the next part.

Operating cash fell $6.6M while pre-tax income rose $4.3M, a $10.9M gap. Someone’s lying. In moments like these, always believe what the cash says, as a precaution. Rightly so, because here it’s the earnings that lie.

They include about $7.0M of tariff refund (a one-off). Adjusting pre-tax income for the refunds, you land on $16.9M, a $2.7M drop from Q1-25. Compared with the $6.6M OCF decline, you get a $3.9M gap. $1.7M of that gap is due to the tax on the refund (useless complexity, just know it’s a one-off), the rest is mostly working-capital swings. In short, after a few adjustments, nothing relevant here.

Last point to flag, the company bought back almost 3 times more shares in the period than in the prior quarter. I’m saving the capital-allocation analysis for the management part.

Got something out of this? Restack it, or forward it to the one investor who'd actually read it. That's how the undiscovered get discovered.

Share

5.5.2 Analyzing the Walmart Partnership

This partnership is the most important optionality and one of the essential variables of the bull case. I don’t want to settle for management’s silence to judge it, so I dug a bit while waiting for the next quarter.

Walmart placed “a multimillion dollar wholesale order” with BBW to sell its products directly in its stores (run by Walmart, not a corner of the Walmart where BBW sets up its stand). BBW ships its products, Walmart puts them in its stores, handles everything, and hands BBW its cut. That’s how, from March 16 to May 15, 1,500 Walmart stores saw trend pods filled with BBW-stamped Mini Beans appear, might as well use the brand.

Let’s make a rough estimate of the order’s size. Take the middle of the range: $5M. Assume it covers only the plush (not the derivatives like plush clothing). BBW should make about $2.5 on the plush (sold for about $5 on the shelf). 5M/2.5 = 2M plush. Divide by 1,500 stores and that’s about 1,330 plush per store. Divide by 60 days and that’s about 22-23 plush to sell per day per store, about $110/shop/day. I could be off by a factor of 2 to 3, but it doesn’t change the conclusion: it’s clearly a small test order for Walmart, especially for a collectible.

At the same time, you could buy them directly on the Walmart site: mystery bags, plush packs, collectible sets, and so on. There are a few reviews (the biggest has 44), the vast majority positive. They’re still available on the Walmart site. Residual stock? A top-up of the partnership? Hard to say, all the more since it may not be tied to this partnership directly. Some are on promotion, others out of stock. Here too, it doesn’t really help decide.

I combed social media for more information. First, on the BBW-specific subreddit (mostly collectors and slightly too seasoned parents), reception is positive overall. The vast majority of cases where I saw the brand-dilution-through-Walmart argument were in analyst pieces on the partnership rather than from customers themselves. Positive reception, but nothing showing it reaches beyond the usual BBW sphere. I didn’t see an abundance of content either. The fans liked it. The less-fan-but-still-fond-of-the-brand crowd were pleasantly surprised to see it again in an unexpected place. Influencers made content about it. But no sign of real buzz like there’s been for other BBW successes. Not cold, but not hot either.

In the same stretch, BBW opened a wholesale showroom in Los Angeles and built out the dedicated team. The aim is to land other partnerships with other wholesalers. You don’t build a showroom on a failure, but the silence caps the odds of a total success:

“In the first quarter, we expanded the team for this important strategy and opened a new Build-A-Bear showroom in Los Angeles to help serve our wholesale account base. This is partially based on the confidence from the recent launch into 1,500 Walmart locations featuring our popular Mini Beans collection.” (Conf Call, Q1-26)

While I was at it, I analyzed secondary-market price trends, but nothing significant there either. Prices remain close to in-store prices (not always easy to judge, since in-store prices themselves vary). Here too, it’s not negative, but you can’t claim it’s positive. Which is the perfect conclusion for the analysis of this partnership. Keep it in mind for the valuation part.

We’ve finally laid all the groundwork we need; we can finally get into the important part. The one that justified these 10,000 words. Don’t worry. There are still 15,000 left before we can decide whether the bet is worth it, why, from what price, and under what conditions.

6. The Crux: The Bear Case and the Four Risks

6.1 The Four Problems

Let’s set the scene again:

• US consumer confidence was at its lowest in decades, with a slight uptick in the latest data, thousands of stores closing, an import cost no one can pin down at twelve months, and a consumption shift toward experience.

• Build-A-Bear posts its fifth straight record year, its biggest sales day in history on Valentine’s, and guides to do it again. It opens franchised stores abroad. It buys back its shares and pays dividends on a balance sheet built to hold. Big but: in the first quarter, it flinched.

This section focuses on that flinch and its implications for the future.

Let’s size the problem. In Q1-26 (ended May 2), revenue falls 2.4%, to $125.3M; retail sales give up 5.1%, and domestic traffic drops 7%, after outperforming the national trend for 2 years while the national numbers were falling. It’s a surprise to management too, which had guided “roughly flat.” Management pins the shortfall on a “broader macro shift,” cuts its revenue forecast for the year, and warns the second quarter will be weaker still. Pre-tax income rises, but that’s misleading: excluding the $7M one-off tariff refund tied to the prior year, it falls to $16.9M adjusted, down from $19.6M last year. The business itself is shrinking. To sum up: a real revenue drop, a real operating miss, and a cut guidance.

We’re done with the branches, on to the roots. I’ve flagged the 4 problems that, to me, could justify a problem in the business, at least on the surface. On their importance, I ranked them by their impact on the last quarter (Q1-26). From most to least important:

1. Falling in-store traffic. In Q1-26, domestic traffic fell 7%. That figure is worse than the national trend, after 2 years of outperforming that trend all through the down cycle.

2. A total shift in marketing. In late 2025, Google pushes its AI-generated answers more aggressively to the top of search. As a result, organic ranking collapses. As a result, free traffic to buildabear.com melts away. It’s a broad trend: on queries topped with an AI answer, the organic click rate drops from 1.41% to 0.64%.28 Meanwhile, US organic search traffic falls 2.5% year over year29, and zero-click heads toward 70% of queries.30

3. Customs tariffs. BBW imports everything from abroad: 51% from China and 44% from Vietnam. Management guides for about $10M in net cost over FY26. Beyond the amount, it’s the uncertainty that’s the problem.

4. Wage inflation and SG&A overall. In Q1-26, SG&A goes to 44.8% of revenue versus 41.7%, or +310 bps.

We’ll spend the next half hour of reading sizing, working through, and quantifying the impact of all these problems.

Before we start: in my reading of analyses of the company, I came across, more than once, two problems cited as major ones that, to me, don’t deserve the title at all. I mention them here to head off future questions:

• Competition from Pop Mart’s Labubu and from Blind Boxes in the Kidult sector (collectors especially). This is objectively wrong. Yes, there was clearly a mania around the products in 2024-2025 (Pop Mart did a 20x in 1 year), but not only has interest in Labubu dropped sharply since (Bloomberg is already talking bubble bursting31), the secondhand resale prices have collapsed.32 In any case, the competition was largely in the e-commerce segment, which is still the minority part of DTC. False problem and false consequences.

• Offsetting the drop in organic traffic (2) would supposedly have a decisive impact on SG&A (4), and BBW would have to “fix marketing to fix its margins.” This is largely incorrect, with the numbers to show it. But the explanation needs more detail. I’ll come back to it in point 2.

Here we go.

6.2 Falling In-Store Traffic

Over the 2 years the confidence trough was forming, BBW’s traffic had beaten the national. As recently as March, the company described its 2025 performance as “outpaced the national average” and ended the year barely down (less than 1% while the national benchmark lost about 5%, Conf Call, March 12, 2026). In Q1-26, while BBW loses 7%, US retail traffic is positive, around +1.7% in March, the sixth straight month of gains.33 In one quarter, the gap flips: from about four points above the national to eight below.

So the problem seems to belong to BBW alone. Really? As is often the case, the reality reads between the lines.

Most of the hole comes from existing stores (-$7.5M), while recent openings recover a good chunk of it (+$3.5M). Everything else (closures, FX, gift cards, and so on) counts for almost nothing. Meanwhile, the average basket rises, driven by units per transaction and a few well-placed price increases. The diagnosis is clear: fewer visits to existing stores.

Okay. The comp first. Q1-25 was the biggest quarter in BBW’s history: double-digit growth, rising margin, and traffic carried by launches that crushed it with teens and adults. The last two years benefited from a string of collector drops that fed those records. Q1-26 doesn’t get a product release like that, but is still compared against it. The same volatility that put BBW above the national now puts it below, once you compare.

If it’s a question of release timing, is it temporary? According to management, yes. The bulk of the releases is loaded into the second half. I’ll spare you the plush collections hiding behind that “loaded calendar,” but note also, even mainly, the opening of the Orlando flagship in the second half, the company’s biggest store. The business is inherently lumpy; the first-half dip looks a lot like a hole in the schedule, and that’s management’s view: "Retail segment performance will remain back-half weighted in 2026." (Conf Call, Q4-25)

Okay, the comp was tough, but there’s a real story behind it. Teens and adults are the cohorts that dropped off most. If you’ve followed along this far, you probably guess why: e-commerce. As a reminder, the site is a storefront for collectors that makes them want to go to the store (sometimes out of pure desire, sometimes for an in-store exclusive), and that traffic got cut by Google’s zero-click. We’ll cover it next, but the idea is simple: everything is connected.

Another point, minor but it had an impact, the weather: “severe weather across large portions of the country.” Management had sized such a shortfall at $2M, but for the prior quarter (January’s weather).

Finally, of course, the macro backdrop. We covered it in depth in part 3. Let’s quickly recap the conclusions:

• Sentiment is far more pessimistic than behavior

• The impact is felt much more on replaceable physical objects than on the experiential

• A few signs of improvement are taking shape.

• Even if the cycle drags on, BBW is equipped to survive, both on its balance sheet and on its P&L, with growth abroad.

Surviving isn’t enough for us. What’s the probability the damage is significant for BBW if the cycle drags on or even worsens? Based on all my assumptions and calculations, I estimate the probability of a net-income decline at 15.0% or more to 24.4%, with a 2-sigma confidence interval. Ah, if only... I do have a hunch, but it doesn’t count. The only things I can use are the present and tests for the future.

For one, Valentine’s 2026 was the best revenue day in the brand’s entire history. When marketing matters less, customers still show up at BBW. Easter held the comp against last year. The Frosted Animal Cookies collection (the things you end up writing as an investor...) sold through in under 2 weeks. Above all, right after the new collection launched (March 2026), management was already noting a “trend change” in traffic. The schedule-hole thesis looks more plausible than a demand problem.

To find out is simple. The second half sees BBW’s 30th anniversary, the Harry Potter series release, the other collections, the Orlando store opening. The comps aren’t as tough for Q3 and Q4. If, despite all that, second-half traffic stays below the national average, the schedule-hole hypothesis is eliminated. Note that Q2 suffers largely the same problems as Q1, hence the focus on H2.

6.3 Clickocalypse: AI Search and E-Commerce Traffic

(The word isn’t mine: management dropped it on a conf call, borrowing an industry term. When management names its own problem, you might as well take it seriously.)

Let’s lay it out properly. Since Google started answering directly at the top of the page, the free click to merchant sites has collapsed. Sharon sizes it at “double-digit impact to the direct click-through… from organic search” (Conf Call Q4-25). The desire to buy hasn’t moved: when the customer arrives, they convert and spend more, basket and units up. It’s a discovery problem: the flow of strangers who used to land on buildabear.com via a search.

E-commerce demand falls 26.1% in Q1-26, after -13.6% in Q4-25 and -5.5% across 2025 as a whole. Not only is the decline significant, it’s accelerating. It's bounded, though: the direct revenue impact is only $2.9M less, versus the $6.1M drop in net retail. E-commerce is still the minority part of DTC, but as we said just above, e-commerce works to send flows back to the stores, so this hole in e-commerce hits harder than the $2.9M.

That said, AI isn’t entirely to blame; management admits it delayed e-commerce upgrades while migrating its legacy inventory system (Conf Call Q4-25), against a backdrop of tough license-launch comparisons. So the -26% inflates the real structural hole attributable to Google.

The question left is who takes the hit. BBW is structurally better sheltered than its competitors. Luckily, we already have a fair amount of data on the topic. AI devours informational traffic and long-tail sites, but it barely touches the navigational/transactional. Short queries, like “build a bear,” trigger an AI answer only 8% of the time, versus 60% for questions.34 AI coverage of e-commerce actually fell year over year, from 27% to 18.5%: Google protects purchase intent, where it monetizes (it also has to compete on the effectiveness of its search engine and AI’s impact on it). Over the period, the biggest sites gained organic while the tail lost it.35 And buildabear.com’s traffic runs precisely on direct and on the typed brand name, so these are queries AI barely touches. Add the retailtainment base, where most sales happen in store, brand recognition for eight in ten Americans, an owned customer base you re-engage without paying anyone, and 1,500 Walmarts where discovery is physical, filtered by no algorithm. Management says it better than I do:

“We’re actually very well suited for what is expected to be the positive attributes of the brands… One is that you’re a strong branded company, so people are really looking for you and your brand. One is that you can create unique and engaging content. Another is that you have the ability to speak directly to your consumers… we have, like, an 80% capture rate at our stores… so we can send them direct emails.“ (Conf Call Q4-25)

Digging through the old conf calls, you read that this isn’t the first time the problem shows up. Back in 2024, web traffic had clearly dropped off, attributed to a fall in organic search amid competitive conquesting. A few quarters later, traffic back and “improved organic search results” (Conf Call Q2-24), via an SEO overhaul and a better product calendar. The volatility of online distribution is recurring here; it’s a near-constant game of adaptation, and the team has already proven it knows how to adapt. I’m still surprised, though, that e-commerce investments in 2024, notably the arrival of a Chief Customer & Digital Officer and a Chief Revenue Officer, led to such a clickocalypse impact over two quarters, accelerating. It’s clearly a problem that mustn’t last, or it calls management’s operating quality into question.

Let’s get to the false problem I mentioned in this part’s intro. A recent analysis claimed SG&A margins were BBW’s main problem and that to fix them, it had to fix its marketing and its reach problems via Google. The idea was that to maintain its retail growth, BBW has to offset the loss of the free internet-search lever through SG&A. To me that’s an analytical error. On paper, the analysis can look right: SG&A rose to 44.8% in Q1-26, versus 41.7% a year earlier. But the analysis falls apart the moment you break down the increase:

“SG&A expenses were $56.1 million or 44.8% of total revenues, compared to 41.7% last year. Higher wage rates and investment in talent, followed by general inflationary pressures and timing of longer-range investments, all contributed to the 310 basis point increase.” (Conf call, Q1-26)

Not a word about marketing. This margin problem is more a reflection of the fourth problem mentioned above than of customer-acquisition cost, all the more so given what we just covered in this subsection. One point can be conceded: part of the “talent” mentioned refers to talent recruited precisely to deal with, among other things, the clickocalypse. But these salaries are on the order of a hundredth of a margin point, at most a tenth. You objectively can’t call that BBW’s main problem.

Management is in fact guiding toward a reallocation of marketing rather than an increase, with less organic and more direct contact. “There will be a decrease in our SEO spend and an increase in direct marketing” (Conf Call Q4-25), paired with an upgrade of the product schema for AI engines, email, and social. A big part of this shift costs very little when the brand does the work in place of the budget: the Frosted Animal Cookies collection generated nearly 0.25 billion media impressions in a week, largely in earned content (Conf Call Q4-25), and its PR approach topped PR Newswire’s March releases “due to its ability to catch both eyeballs and AI agents” (Conf Call Q1-26). Content is now optimized as much for the AI as for the customer.

Here too the problem isn’t as bad as it looks, but it's real. I see two big unknowns.

The first is a question of speed: BBW will adapt to AI search, the only question is how fast versus its competitors. The second is a question of amount: how much extra SG&A to hold DTC growth. It’s possible the analyst from earlier is right, but only in the future. For now that doesn’t seem to be the case, and the power of the BBW brand makes the probability of a truly significant impact (several basis points) fairly low.

Let’s still assume it costs 3 points of SG&A margin. The TTM cost would be $16.2M. Let’s not assume DTC revenue rises thanks to that SG&A spend (even though that’s its reason for existing). Alongside that, BBW is still a company growing double digits in its Commercial and International Franchising segments at higher margins. It’s also a company with a strong brand that has the ability to raise prices (and exercises it). That probably won’t offset $16.2M, but at least half is entirely plausible.

All of this rests on the present problem. But woe to anyone who underestimates the impact, and above all the evolving impacts, of AI on just about everything. Everything can change very fast, for better or worse. And this Google click problem could be nothing next to a risk caused by some new AI feature in a few months or years. That risk is quantifiable neither in probability nor in the force of its consequences.

The only thing I can say with certainty is that today, the Google click problem has a real, bounded impact, one management has a strategy for and is in the middle of deploying, with weapons many of its competitors don’t have, partly offsettable, with a brand power that makes a bleed of several margin points very unlikely. At 7.3 times earnings, this risk alone doesn’t justify the multiple.

6.4 Tariffs, Ô Tariffs: Supply Chain and Pricing Power

BBW imports nearly all of what it sells: 51% from China, 44% from Vietnam. And until 2025, it cost nothing extra (plush came in duty-free): “toys prior to this situation were tariff-free. This was just not something that we really had to deal with” (call Q4-25). The “tariff” line fell as if by magic from the regulatory sky. Exactly the kind of cost a slightly too nervous Mr. Market loves to extend in a straight line.

On February 20, 2026, the Supreme Court struck down the entirety of the IEEPA tariffs. The refund was already covered in the accounts section. Let’s just recall it’s an expected $13.2M refund, nothing that really concerns us for the future except that it flattered the comps and, above all, that it’ll muddy future comps. In the same move, BBW brought its 2026 estimate down from $16M to about $10M. Speculating on the tariffs themselves would be a waste of time. The only thing worth our time is how management can respond, and how it has.

In 2025, BBW absorbed $11M in tariffs ($0.65 of EPS) and still posted a record pre-tax of $67.2M with gross margins up on the year. The shock was absorbed and the business improved. As mentioned above, BBW used its moat: selective price increases on the most in-demand products, pushed above the gross line. In Q1-26, ex-refund, retail gross margin gains another 190 bps, “primarily driven by selective price increases” (10-Q).

The main problem with tariffs is the uncertainty they create: “we’re not talking about just tariff rates going up and down. We’re talking about tariff rates changing from country to country on some monthly basis” (Sharon, Q4-25). And it’s not over.

Today’s 10% comes from a temporary surcharge put in place hastily after the Supreme Court decision, expiring around July 24, 2026. No one knows what will replace it. The Court of International Trade (CIT) even struck it down on May 7, 2026, a decision immediately appealed. The administration is already turning to a more durable lever aimed at China (Section 301), and two such investigations name Vietnam specifically. The Vietnamese rate, down to 10%, could therefore climb back toward the 20% it showed before the ruling. On the refund side, these are just receivables on the balance sheet whose timing depends on the US Treasury. In short, the uncertainty cuts both ways: the tariff can keep falling or snap back up, and BBW can’t build its plan on either. The only certainty, the show must go on.

As we saw earlier, these tariffs are offset by growth in commercial and franchises. And these segments are not subject to tariffs: “Inventory held for our international corporate and partner-operated stores is not subject to tariffs” (Q4-25). The fast, high-margin segments, commercial and franchise, ship from Asia to other countries, outside US customs.

Does the $10M/year cost, offset by the brand’s pricing, partly dodged by the asset-light mix, and whose first nuisance is unpredictability rather than amount, help justify a multiple of 7.3 times earnings? No. Certainly not. The last problem, on the other hand...

6.5 SG&A and Margins

I said I ranked these problems by order of impact on the last quarter. So it comes last. But to me it’s the most significant long term, and the one that most easily supports a bear case. Yet it’s rarely mentioned. By management or by analysts.

Behind the wall is the part you came for: every remaining risk and the weight I give it, the valuation model, the bull/base/bear IRRs, and the exact price and conditions that would make me a buyer. Plus every deep dive I’ve written, and every one still to come.

I’m not selling a portfolio and a track record. A track record is mostly survivorship dressed up as skill, and it reverts to the mean, if not below it, at the next real bear market. I sell something duller: full deep dives on small and micro caps nobody knows or bothers to understand, and the follow-up after. You pay me to give you everything you need to make your own call.

I’m not done with this name: the next update, with Q2 and the first real Walmart numbers, lands for subscribers. If that’s your kind of research, the button’s right below. On the annual plan, it’s €25 a month. Cheap, for work that pros read.

Read more

“If you want to be happy, be.” - Aleksei Tolstoï

A soft note to open, because from here on, it’s blood, sweat and tears.

I’m starting from one observation: the geopolitics of diplomacy is giving way to the geopolitics of conflict. Decades-long tensions are crystallizing into wars no one can win.

From that observation, grim but pragmatic, three sectors stand out. They're the ones I expect to shape my portfolio over the decade ahead.

My job here: the thesis for each sector, the names to play it, and why each one earns its place.

Subscribe now

Sector 1

The first is the most obvious. Too obvious, on its face: defense.

From 2016 to 2025, global military spending rose 41% in real terms. The military burden reached 2.5% of GDP, the highest since 2009.1 Last year’s NATO summit in The Hague raised the target to 5% of GDP by 2035 (at least 3.5% of it on core defense).2 Europe’s recent Readiness 2030 rearmament plan is designed to mobilize up to €800B.3 The U.S. national-defense 2027 budget request tops $1.5T, the largest request in U.S. history.4

Plenty of billions, plenty of GDP. But the qualitative side matters just as much.

Japan is in the middle of its biggest rearmament since 1945. Turkey is emerging as a military-industrial power. Russia has tipped into a war economy. And the 2027 clock is ticking: the deadline by which Xi has ordered the PLA to be ready for an operation against Taiwan (not an official war date, but close enough to matter).

So far this is first-order analysis. Anyone who regularly reads the right papers gets here. No alpha in that.

(That’s the theory. In practice, I’ve shown plenty of times that the main source of alpha, created or destroyed, has never been the information. It’s who uses it. But I don’t need that argument here.)

Let’s jump straight to the second order: then what?

First point: a budget line is not a cash flow. Any outfit that depends on the government will tell you as much, bitterly.

NATO’s 5% of GDP, Readiness 2030’s €800B, the U.S. national-defense’s $1.5T: these are promises. By definition, a promise is supposed to be unbreakable. In politics, it’s... different.

But a signed contract is a whole different animal. When Finland signed for 64 F-35s in 2022, Lockheed locked in cash flows for decades: deliveries through 2030, so MRO into the 2060s (not all of it; partner contractors take their cut).5

The order book, then, is the buffer against the political cycle. The moment a promise makes it into the order book, it becomes a commitment that will turn into cash flow (though the contract can still be modified, stretched out, even canceled under certain conditions). Watching the order book is therefore the best way to read the sector’s economic reality. And what a reality. The defense backlog has grown for the third straight year (book-to-bill above 1). And combined revenue for the world’s top 100 aerospace and defense groups blew past $1T for the first time in 2025.6

“Fine, okay, this takes a bit more analytical nuance, but nothing earth-shattering either. Still not convinced there’s alpha in it.”

And you’d be right. We’ve still got a few more “then what”s before we get to alpha.

Then what?

The order is only the tip. When Kongsberg sells a NASAMS air-defense system to Denmark for €500M, that €500M is the appetizer.7 The main course is the recurring revenue over the decades that follow: the aftermarket. Decades of recurring revenue, higher-margin than the OEM sale and, above all, non-cyclical. You can defer buying a platform. You don’t defer maintaining a fleet already in service. So a weak order book in any given year has limited impact as long as there’s an installed base to maintain. To sum up: the sector has gone from an industry that sells machines to one that sells “subscriptions” to keep them running.

Then what?

Here we get into a wrinkle you rarely see flagged in sector theses: the shift in the economic unit of recurring revenue. Take a typical MRO invoice from a few years ago: a hydraulic actuator here, a wiring harness there, a set of turbine blades, and so on. Physical parts, plain and simple. They take time, hands and skill to replace. Today, that same invoice can tell a very different story: a mission software update, a threat library subscription, a licensed capability unlock, and so on. Here, once the code is signed off and certified, a few clicks and it's deployed.

We’ve gone from recurring revenue built on hardware and cost-plus contracts to recurring revenue built on electronics and embedded software (physical MRO obviously hasn’t gone away). Investor hats back on: the economic unit of recurring revenue has shifted from tangible assets to intangible ones, with the margin expansion that follows. Code hits margins that metal never will.

Any valuation of the sector has to account for this shift toward non-cyclical, higher-margin recurring revenue. But at best that justifies paying higher multiples. It doesn’t get you to a buy. Where’s the alpha, damn it?!

To find it, let’s recap:

• Allocation to the military sector has gone up.

• It’s supposed to keep rising, and the only real gauge of that is the order book.

• Order books are rising and hit records in 2025.

• Those order books are a source of non-cyclical recurring revenue, part of it higher-margin.

Then what?

How do you spot the real winners among all these overflowing order books? Truth is, I’ve already given the answer. Take a few seconds to think about it. Could be fun.

If demand is no longer the constraint, the only thing left is supply. And supply comes down to two questions:

• How fast does the company convert its order book into cash flows? (In defense, delay is almost the rule.)

• What margin does it make on each unit produced?

The big winners will be the ones that can burn through the order book fast, lean heavily on the aftermarket, and earn the fattest margins there. So a software-led mix (OEM margins are too uniform, and OEM revenue is rarely big enough to matter).

How do you find them, and, above all, at the right price?

Let’s start at the end. Forget the upbeat headlines. As we’ve seen, promises aren’t cash flows, and what little alpha those stories carry has already been scooped up by better-informed players. What we’re after is the bad news. A budget decision postponed. An investment plan scrapped after a change in government, and so on. This is noise. The order books are full, and will stay that way for years. Almost anything that knocks the price down without affecting the order books directly or the company’s ability to convert them efficiently can be read as a buying opportunity.

Okay, that’s the when. Now the what?

Worth your time? Worth someone else's.

Share

The Picks 1

First off, forget the defense majors, even the pure-play defense SaaS names trading at indecent multiples. Alpha lives in small caps almost by definition, and in this case especially in Europe (where the software shift hits hardest).

Unfortunately the sector has already run hard since the Russia-Ukraine war, and today’s micro caps need far more than a sector tailwind to work: an unproven technology, shaky contracts, and so on. So I've set them aside for this sector.

One note that holds for the whole post: I don’t own a single one of these companies. They’re candidates I’m actively tracking to play these sector theses. And for anyone still in doubt: I’m not a financial advisor and this isn’t financial advice. Your money, your judgment, your risks, your returns. The fine print here.

1. Theon International Plc (EURONEXT:THEON): a bet on the resilience of an exceptional operating margin, with expansion upside from the mix.

In 2025, Theon posted a 26.2% adjusted EBIT margin, high for a company that's still very hardware-heavy. It held that margin even while integrating Harder Digital, a company it acquired to lock in supply of image-intensifier tubes (a critical, supply-constrained component). The legacy business probably runs higher margins than the consolidated figure, though that obviously depends on mix and on the integration. What matters isn't so much the size of the order book as its quality: €2.3B of soft backlog plus options (€1.4B of it pure soft backlog), which gives visibility through 2029.

The real upside is in the mix: the new digital products should more than double in 2026, to roughly 20-25% of revenue, while management targets 20% of revenue from platform products over the medium term. Management has to keep executing, hold the margin on each product, and shift the mix to lift margins overall. The secular trend handles the rest.

At 20x EV/EBITDA and 20x P/E, this isn't a compelling entry (not for me, anyway). The stock is objectively very volatile, so patience stands a good chance of being rewarded.

2. Hensoldt AG (XETR:HAG): a bet on the speed of order-book execution.

The company spans several layers of electronics (radar, optronics, electronic warfare). What we're looking for is self-evident: recurring revenue from upgrades and MRO over long life cycles, layers of software and, of course, analytics.

But what interests me about Hensoldt is its conversion speed. Case in point: book-to-bill jumped to 3x last quarter (guidance is closer to 1.5-2x). Orders are coming in far faster than they turn into revenue, so the whole thesis hinges on that conversion. The 2026 adjusted FCF guidance was raised to 50% of EBITDA, up from 40%. That bump also comes from higher customer prepayments, so it's as much a demand signal as a real conversion gain (assuming the company executes).

The real catch is the price. At 25x EV/EBITDA (18-19x on 2026 TTM), it's objectively still too expensive despite the pullback. But no doubt Mr. Market will hand us better entries.

3. Kratos Defense & Security Solutions, Inc. (NASDAQ:KTOS): a riskier bet, on execution and margins at once.

Here the starting point is different. The company makes drones, unmanned systems, and propulsion, among other things. It has a clear technological edge. The thesis, if it works, rests on its ability to convert that edge into an installed base, and to turn that base into high-margin recurring revenue.

You can split the company into its two reported segments. Unmanned Systems, about 22% of revenue, home to Valkyrie and the target drones: these are designed to be attritable. Durable recurring-like revenue, if it comes, sits more in parts of the remaining 78%: mainly the ground software that flies the satellites, sold as “software/systems/support,” even if Kratos stays a grab-bag of many other things (and there really are a lot of other things in there; it’s complexity that should widen the margin of safety you demand.)

On the propulsion side, it's a bit trickier: drones and munitions get used up, so they get rebought continuously; once a Kratos engine is qualified on a program, it gets reordered on every production lot for the life of the program. Basically recurring revenue through reorders, just a bit lumpier. For now the EBITDA margin runs around 10%, far from what you'd hope for once the model is at cruising speed. And the cash isn't there yet: management is still guiding to negative FCF of $85-105M in 2026, on $155-165M of CapEx. The potential is enormous, but so are the valuations. I won't even give you the numbers. They're that indecent. Just know that they don't justify risking money on that potential, even factoring in the decade-long tailwind the sector enjoys. But once again, I trust Mr. Market. In the meantime, I'm keeping a close eye on it.

I'll head off the question: I don't see ISSC as a way to play this sector thesis for now. Its military revenue is only about 40% of the total, and its upside in this sector is limited (the F-16 fleet isn't going to double in 10 years). My view could change with the next acquisitions and/or new contracts. Subscribers will be the first to know.

Sector 2

You work at some company... fine, let’s make it a European bank. It’s time for the year-end budget meeting, and neither you nor your colleagues are exactly thrilled. The outlook is grim, and everyone around the table knows cuts are coming.

Every line item goes under the microscope. Cloud migration, pushed back. The two planned hires, frozen. Marketing, it has to come down. Call it 30%. You keep going, line by line, until you hit the IT Security and Compliance lines. Relief, all around. No need to think hard. There’s not much wiggle room here. There are laws, deadlines, audits, and, for good measure, the name of the person sitting at the head of the table on a regulatory filing the day it all breaks. There’s a floor in these lines that’s simply untouchable.

Let’s put some numbers on it anyway.

The global cybersecurity market is $213B in 2025 (the number shifts depending on how you slice it). I find that almost laughably small given the stakes, but reality always gets the last word. It grants me one point, though: double-digit growth, with $322B expected by 2029.8 No surprise there. Attack surfaces multiplying faster than lily pads on a pond, because of (or thanks to) the cloud explosion. Ransomware going industrial. And above all: AI.

AI’s impact on the sector is like handing every primate on earth an AK-47. Suddenly it’s very easy to do serious damage to just about anyone, as long as you’ve got two opposable thumbs.

From an investor’s standpoint: it creates an incentive to attack you, and drastically lowers the barrier to entry (less fun when you put it that way). All it takes is an internet connection and a subscription that runs a few dozen dollars a month. You no longer need to know C/C++ to write polymorphic malware. Learn English to run credible phishing campaigns? Not even necessary anymore. You can even uncover a zero-day without knowing what the term means. A total democratization of offensive cyber, from the total amateur to the totalitarian state (or the less totalitarian one).

This isn’t just theory. Just last year, Anthropic disclosed a large-scale espionage operation (probably Chinese) in which almost the entire attack chain was run by AI, from reconnaissance to exfiltration.9 Nobody knows all the rules of the game, not even governments. To be fair to them, the rules change with every new Claude update.

Let’s put some numbers on AI security: $49B in 2025. That’s peanuts. Though in fairness, that's an easy thing to say a year on, in a world that moves this fast. By 2029, the figure is projected at $160B.10 More than a 3x. Worth noting: it’s easy to slap an “AI” label on a product, and properly counting the cyber tools that are genuinely AI-powered isn’t so simple.

My opening scene “showed” it: you don't slash a cybersecurity line. But there are two ways to bring it down (finer trade-offs aside):

• Push out an upgrade.

• Cut cyber headcount.

In 2025, 36% of companies reported cuts to their cyber budgets, and a big chunk of those came from cutting headcount (about 2/3 cut their cyber teams).11

AI will hit both of these variables:

• As noted, the rapid refresh of AI-enabled offensive capabilities forces an equally rapid refresh of defensive tooling. Those refreshes become recurring revenue, probably high-margin, for the cyber companies.

• AI should (the conditional matters here) help shrink companies’ in-house cyber headcount (all else equal), but also, maybe above all, shift the skill profile they hire for. If AI automates the attack, it automates part of the defense, so it calls for a skill set suited to that offense/defense automation. That reduction and that shift should free up more budget for software and infrastructure, most of which ends up in the cyber companies’ P&L. More on that later.

Like defense, cyber is in for major growth. But there are three big differences worth flagging.

First, cyber already runs largely like a recurring-revenue SaaS model. There’s infrastructure, sure, but it’s usually not the bulk of a company’s revenue (worst case, stock picking solves that).

Second, contract execution costs tend to run lower than in defense. A cyber vendor adds a little infrastructure, wires the client’s workflow into the existing tools, and trains a few people. After that, it’s maintenance and monitoring, where margins are generally higher.

Third, the barrier to entry for starting a cybersecurity company has dropped sharply, and it’ll probably keep falling with every big new AI model release.

From all this, one variable stands out, the one that, in my view, separates the big winners in this sector: customer stickiness. Secular growth only really compounds for the companies that can hold on to their customers.

Back to the shortage of talent suited to these changes. In 2025, 95% of companies surveyed reported a skills gap across their cyber teams.12 It’s hard to be any good in a field that moves this fast. It’s also a market need that the cyber companies can, and should, fill.

Put it all together, and the profile of the future winners takes shape: a company anchored in compliance, with enormous customer retention, and able to turn the talent shortage to its own advantage.

I dug in. Found 3 that check every box:

• A quasi-monopolistic compounder where stickiness is an understatement.

• A banking hardware→software transition already well underway, with supranational regulation strengthening the tailwind.

• A sovereign mission-critical niche, obviously very sticky, and accelerating.

By going Pro, you get the full post, plus every deep dive, past and future (one to two a month, plus follow-ups). For example:

One undiscovered company at the right price makes the cost of this newsletter trivial.

Become a Pro Member

You can get free access to paid content by sharing this newsletter with your network. Rewards compound:

• 3 new subscribers = 1 month;

• 8 new subscribers = 3 months;

• 15 new subscribers = 6 months.

Refer a friend

Read more

At least one of those three got a reaction out of you. That was the point: to make you catch yourself having an opinion. Got it? Good, because we don’t care. Mine included. Opinions are a last resort for when you have no data.

And with IPOs, we have plenty. So check your emotions at the door, shut down the limbic system, and fire up the cortex.

Let’s hear what the data has to say.

On Paper

Normally, this is where I'd find a clever way to say that before you analyze something, you should define it properly. This time, I'll do it by breaking the fourth wall. It’s done. Pardon my impertinence.

Stripped to the facts, an IPO is:

• The first sale of a company's shares on a public market. The sellers are the company, its existing shareholders, or most often, both.

• A prospectus, written by the company, its lawyers, and its bankers, with a stated purpose: to give investors the information they need to make an informed decision.

• One or more investment banks that run the process for a fee: 7% of the deal in most cases.1

• Shares sold at the offer price, allocated mostly to institutions. Retail investors, by contrast, buys on the open market, at the market price, starting on the first day of trading. The pop is the gap between the offer price and the first closing price.

Humans, stakes, incentives, and imperfect information. Time to put on our game theorist hats.

Three things jump out:

• A timing asymmetry: it's literally the insiders who decide when to go public.

• An information asymmetry: the prospectus that’s supposed to tell investors everything they need to know is written by people whose incentive is to sell at the highest possible price.

• A power asymmetry: insiders have been shareholders since day one. VCs have been shareholders for 8 to 10 years or so. Most employees have been shareholders for several years. The institutions get to buy at the offer price, just before listing. Retail almost always buys after everyone else.

"Show me the incentive and I will show you the outcome," as dear old Charlie would say. But I find him too realist cynical for this post. I don't need to take selfishness as an axiom. I have data. Plenty of it.

If you like Charlie Munger, irony, microcaps too good to stay microcaps, data-driven analysis, Charlie Munger, or irony, this newsletter is for you.

Subscribe now

The Raw Reality

When?

First things first: timing. Here's the number of IPOs plotted against market valuation (market-to-book ratio).

A neutral eye sees three things:

• IPOs cluster mainly at valuation peaks.

• The sharper the recent rally, the more of them flood in.

• When the market falls, locally or globally, they drop off sharply, or vanish outright.

All three boil down to one number: on average, major IPO waves come after a market run-up of about +31% annualized over the two preceding quarters.2

The timing asymmetry clearly favors insiders and VCs.

Whether it's deliberate or not isn't our concern. Their gains don't necessarily mean a loss for retail. An equilibrium where everyone wins is still theoretically possible: insiders, institutions, VCs, and IPO buyers.

Let's test that against reality.

Who Gets What?

Unfortunately, I had trouble finding data that breaks down each player's performance in detail. But there's enough to sketch most of the picture.

Insiders and VCs are usually locked up for six months. When the lockup expires, average trading volume rises 40%, with a -1.5% abnormal return over three days.3 Both the volume and the negative abnormal return are larger when VCs are involved.4

But jumping from that to "VC = predator" is a bit hasty.

First, VC-backed IPOs perform better than smaller, non-VC-backed ones.5 Second, their alpha is highest while VCs stays invested (1.2pp/month) and drops once they exits (0.65pp/month).6 Their contribution to the "community" is real, as long as they're still around.

On insider performance, I found nothing worth noting. As for VCs, their returns tend to track the market, just far more volatile and amplified (beta of 1.9).7

Fortunately, the data is richer for institutions.

Institutions get 3.3x more shares than retail. So the lion's share. No surprise, they're the bigger fish. But the split is asymmetric: institutions receive more shares in underpriced IPOs than in overpriced ones (75% vs 68%), that is, when the pop is fat rather than thin.8

That's the short term. The longer term matters too (a few months out). IPOs with high institutional allocation post a +4pp/month alpha; the low-allocation ones, nothing.9 Yes, per month. That’s not a typo. To say institutions are better served, serve themselves more, or both, on the best deals, is still a neutral read.

But money doesn't wait. Which is why institutions are far less patient than retail: within two days of the IPO, they've flipped 41.5% of their allocation, versus 18.6% for retail.10

The numbers are clear: institutions, too, cash in on their power asymmetry. But it still says nothing about how retail performs.

The Final Problem, Our Final Problem

You’re retail, with retail problems and retail questions. But above all, you’re neutral. So you figure none of the data above gives you even the faintest lead. Bravo, lovely mindset. Especially on a subject this divisive, in a period that’s even more so.

Let’s hear the data instead.

Not Short Term

Here's the 3-year performance of more than 9,000 IPOs.

A few key numbers11:

• 60% of IPOs are in the red at three years.

• 16.1% are multibaggers after three years.

• On average, IPOs return 19.1% over three years, versus 39.6% for the market.

• That average is heavily flattered by the 0.5% of firms with exceptional returns: the median falls to -25.7%.

• Performance comes down to two variables: size (large = >$100M in sales) and profitability. Here's their average underperformance versus the market over three years:

  • Small and unprofitable: -40.9%

  • Small but profitable: -25.9%

  • Large but unprofitable: -2.7%

  • Large and profitable: -3.4%

To put it coldly: the 3-year IPO is a losing game on average, worse at the median, and worse still for small companies.

I spent almost as much time writing this piece as I did making these visuals and the ones that follow. Sharing helps spread the knowledge (and makes the time spent on these damn charts worth it too, I'm not gonna lie).

Share

Three years is a long time. Some of you may want a shorter horizon.

Short Term

Institutions are better served on the best deals. The consequence: the retail buyers who do get an allocation get it more often, and in larger size, on the bad ones. The phenomenon is so common it has a name: the winner's curse12 (I didn't pick the name, promise. No judgment, remember).

Fine: retail investors who get an allocation are disadvantaged from the start. But they’re only a small minority of retail : the vast majority buy on the open market.

For a very short-term view, here are the numbers on the biggest IPOs over the past 15 years:

Roughly, it's a coin flip: a little more than half of IPOs are in the green early on, a little less after a year. The turning point comes around six months: the share of companies still positive drops from 57% to 43%, and the median return falls with it. That lines up with the end of the insider and VC lockup, and the hype cooling off.13

The last piece of the puzzle only comes into view when you look at the biggest IPOs of the last 15 years one by one.

In its first year, every one of these IPOs ran a max drawdown of dozens of percent. Every one.

Enough numbers. We have what we need to tell the story.

The Complete Statistical Story

Everyone but retail benefits from their respective asymmetries. Plenty of studies show how the IPO process and the incentives feed each other, making the game even more "biased," but each adds a layer of interpretation, and that's what kept them out of this post.14 Just in case: everything pointed to a structural, sustained disadvantage for retail.

That doesn't mean retail investors can't win. Short term, it's theoretically possible. You'd have to adjust raw data for fees, spreads, taxes, and above all for self-sabotaging psychology (still an objective fact, no?). But on paper, it's possible.

For long-term investing, the numbers leave no doubt: buying an IPO almost certainly means sitting through a drawdown in year one.

Yes, wonderful companies can go public. Yes, nearly every one of the world’s finest public companies went through an IPO. But there’s almost always a far better buying opportunity than the IPO itself. Usually within the first year.

If you're not chasing statistically rare short-term returns (days to a few weeks), there's no statistical reason to buy at IPO.

I don't think I can go further without interpretation.

“Those who do not know history are condemned to repeat it.” - George Santayana

Great opportunities exist outside IPOs too. Here's a great one.

Take care,
Flo

Refer friends and earn free months of paid access.
Zero downside (rare in this market).

Refer a friend

So I'm putting 15,000 words behind that conviction. All of it: arguments, numbers and assumptions. Nothing held back.

May you have as much fun reading it as I had writing it.

Take a highly regulated, highly fragmented industry, so interdependent that one large player relies on dozens of others, each of which relies on dozens more.

Add critical, certified products, installed for years across hundreds of platforms, with a customer base that would be an understatement to call “conservative.”

Let that run for a few decades and you end up with large industrial groups carrying hundreds of product lines. Somewhere in that pile of parts, some become too small to remain part of their “internal priorities.” But they cannot disappear. Hundreds of machines depend on them.

When capitalism does its job well, the smaller company steps in to arbitrage the constraint. That’s the better owner principle.

It buys mature product lines the larger players want to shed, takes over the rights and the customers, then runs them through an optimized cost structure.

Call it industrial recycling, with better margins. Many companies have created a lot of value this way. TransDigm did it in proprietary aerospace components. HEICO did it in replacement parts and approved repair.

Today’s company has turned this mechanism into a business model, one its legacy business sharpens: each side feeds the other.

Revenue up nearly 80% in a year. Margins expanded. Management still guides the company to nearly triple revenue again by 2029, while keeping the kind of margins most industrial businesses would politely call unreasonable.

And yet the stock is down by half because Mr. Market had one of his recurring mood disorders. A quarter came in with little growth, a pause management had been flagging for quarters. But the market decided that a timing issue was a thesis problem. Helpful, as always.

He did the same thing a year ago. It worked out nicely for whoever was buying.

Here's Innovative Aerosystems (ISSC).

Table of Contents

1. Short-form Thesis

2. The Business Model

• Where It Sits in the Industry

• The Proprietary Cockpit Stack

  • ThrustSense Autothrottle

  • The Future

• The Acquisition Engine: Industrial Recycling

  • The Seller

  • The Product

  • The Buyer

  • The Precedents

• The Synergies Between the Two Models

  • Cross-Selling

  • Portfolio Completion

  • More Cash Flow to Finance R&D

3. Acquisitions

• Wave 1, First Steps

• Wave 2, Great Leap Forward

• Wave 3, Preparing the Future

• The Patterns

4. The Operational Trajectory

• The Origin of Growth

• Margins

• The May 14, 2026 Sell-Off and Mr. Market’s Mistake

• The Math Behind IA Next 2029

5. The Capital Structure

6. Governance and Alignment

• The Engineer Who Allocates

• Alignment

• The Board and the Shareholder Base

• The Real “Governance” Risk

7. Valuation

• Qualitative Valuation

  • The Discount

  • The Risks

  • Discount vs Risks

• The Assumptions

  • Ground Rules

  • The Scenarios

  • The Inputs

• Verdict

8. Conclusion

1. Short-form Thesis

The setup is fairly simple.

• After the May 14 sell-off, the market values ISSC at around $300M.

• Management is targeting $250M in annual revenue by 2029, at a 25-30% Adj EBITDA margin, which works out to $63-$75M of Adj EBITDA versus $25M in FY2025.

• At the current 12.3x EV/EBITDA multiple, that gives a 2029 EV between $775M and $920M.

• Net of debt, the equity is then worth around 2.5x. No re-rating needed, despite trading 50%+ below peers. No significant change in the business. Just the same old business.

• Proportionally, the downside is sharply limited. Even if the company misses guidance. Even if the company is "frozen."

Still, whether the guidance is credible is the question that matters. The past says yes.

The acquisition strategy already has a three-year track record:

• In 2025, ISSC almost doubled the footprint of its Exton facility, from 45k to 85k sqft, and more than tripled production capacity, for a total cost of (only) $6M.

• Since June 2023, six asset purchase agreements have been signed, three of them in the February-March 2026 window alone, totaling $87.7M in cash deployed. Most of its growth comes from acquiring product lines that the giants no longer want.

• Revenue also almost doubled between FY2024 and FY2025 (tripled since FY2022), and net margin moved from 14.8% to 18.7% over the same period (20.6% to 24% on operating margin).

The pipeline is set to widen further. Honeywell will spin off its Aerospace segment as HONA in 2026. That’s a $17.4B revenue business about to become a standalone pure-play, with the portfolio discipline that comes with it, and therefore more legacy lines to divest to credible, established buyers. ISSC has already proven that credential to Honeywell, more than once. Either way, Honeywell is only one seller out of five or six. Collins, GE Aerospace, Garmin, Moog, and Elbit each have their own legacy lines to rationalize.

The remaining job is “simply” to keep going. Acquire and integrate another $130-$140M of revenue by 2029.

If everything looks so rosy, why the Q2 FY2026 sell-off?

First, the price was already stretched after a 3.8x in four months. Fair to say the market had gotten well ahead of itself, and the smallest hiccup was going to cool it down. Despite my own conviction, I had trimmed 25% of my position at that point.

Second, management had been guiding to a weaker Q2 FY2026 for several quarters, flagging an F-16 trough between two production cycles tied to the line transition from Honeywell to ISSC. Q2 FY2026 came in at $22.4M of revenue, +2.0% YoY, and the market treated it more or less like a broken thesis.

Nothing new. The stock fell from $20 to $8 between August and November 2025 for the same reasons: a revenue air pocket tied to a line transition. The stock then ran almost 4x over the next four months.

The May sell-off changes nothing about the thesis. It's just Mr. Market being Mr. Market. At least it handed me a discount I was glad to take.

All that’s left is to show this reading holds up to the facts and the math.

One click away from not missing the next undiscovered compounder.

Subscribe now

2. The Business Model

Even knowing the company very well, I have to say that the description of its business model in the 10-K is fairly vague, to put it politely. The investor presentation is a much better starting point:

At the center: the physical structure of the cockpit. Around it: all the electronic systems bolted on top to make the cockpit work. ISSC designs these systems, manufactures them, installs them, and supports them across their full lifecycle. That is complete vertical integration, all under one roof in Exton, Pennsylvania.

In a field where every component requires FAA (Federal Aviation Administration) certification and where the slightest modification triggers a long and often painful recertification cycle, that vertical integration is clearly a competitive advantage.

It obviously didn’t come from nowhere.

This moat was built over 37 years of work and relationships, which today allows ISSC to equip aircraft under OEM contracts with the sector’s biggest players. The legacy business has been generating cash for decades (with ups and downs) and accounted for all of the P&L through FY2022. The sector itself acts as a barrier to entry. An operator does not have the luxury of choosing its supplier. It buys what is approved for its aircraft. That’s the rule. And often only one company has gone through the long certification process for its product on that specific aircraft. When Lockheed Martin builds or upgrades an F-16, it has to buy the display generator and flight control computer produced by ISSC. No one else has the right to manufacture them (under Honeywell licensing since September 2024), and no one else wants to do it at that scale.

The second component is what kept that business model alive for 37 years: innovation. The ThrustSense® Autothrottle, the first FAA-certified turboprop autothrottle (which literally saves lives), has been generating cash flow since at least 2019 and will continue to do so for years, if not decades. The current pipeline is a loaded one, with UMS2 and Liberty Flight Deck at the front, and will not contribute meaningfully to revenue before 2027-2028, subject to a regulatory schedule that is notoriously hard to pin down. Behind these innovations, ISSC has a clear vision: autonomous flight. But one thing at a time.

ISSC keeps the cockpits of yesterday in service and is figuring out what the cockpits of tomorrow look like.

That leaves the third engine: acquisitions. That engine kills three birds with one stone:

• It adds certified installed base to the mature stack.

• It gives ISSC the technical building blocks it was missing in the forward-looking pipeline (things that would have taken years to develop internally).

• And it comes with a customer base ready to cross-sell into.

I'll break them down one by one. But first, let's put ISSC back in context.

2.1 Where It Sits in the Industry

I would have liked to write a full section on the industry and where ISSC fits into it. I’m going to skip it. Three reasons:

• We have little visibility into ISSC’s specifics: what it sells, at what price, and to whom. “Professional secrecy.”

• The industry is genuinely complex, with a lot of players and a lot of moving parts.

• And it’s honestly not relevant to the thesis.

For your viewing pleasure, here’s a snapshot of the sector anyway. Don’t bother thanking me.

But we still need a general overview.

ISSC sells products that run from a replacement seal worth a few dozen dollars to the ThrustSense autothrottle, somewhere between $70k and $100k per aircraft by my estimate. And services too: from overhauling a cockpit component for a few thousand dollars to an Engineering Development Contract (EDC) that can run into the millions.

The product/service split is the only breakdown management fully discloses. That answers the what. Not the who, and not the how.

Start with the how. There’s OEM (the new stuff, gear fitted to an aircraft still in production) and there’s the aftermarket, which is everything else. Roughly, aftermarket = MRO + retrofit.

• Maintenance, Repair & Overhaul (MRO) is everything that keeps already-installed equipment running. It’s recurring revenue, and it isn’t optional: an aircraft has to go through inspections at intervals the regulator sets, or it stays on the ground. No recession in this corner of the market. Just the same old business.

• Retrofit means fitting new or upgraded equipment onto an aircraft already in service. Higher margin, and strongly countercyclical: when there’s no budget for a new aircraft, you modernize the old one.

Here’s the part management doesn’t push hard enough, in my opinion. On most of the products ISSC builds (or whose production lines and patents it buys), it is the only provider cleared to service them and supply parts.1 No competitor is allowed to touch a box that ISSC alone is certified to maintain.

If the aftermarket is so attractive, why bother with OEM at all? Why supply low-margin gear to Pilatus, Boeing, Textron? Because today’s OEM is tomorrow’s aftermarket. Every unit fitted on the production line creates an installed base that throws off MRO and parts revenue for decades.

Management gives no number, but my guess is revenue runs around 80% aftermarket (20% OEM), a good chunk of it MRO. My main argument is the way ISSC’s customer base has shifted, and its tilt toward the military, which carries relatively little OEM.

The military, precisely, that’s one answer to the who. But who are the others? The question matters, because the who drives a lot: type of contract, margin profile, and the rest.

No disclosure here either, just my estimates. They can swing hard from one quarter to the next, even one year to the next. The constant is the weight of the military, which four years ago didn’t account for more than 10%.

Here's what matters about each.

• Military: combat and transport/refueling aircraft flown by armed forces, not necessarily American ones. Usually long-dated contracts with low gross margins, offset by near-zero SG&A. No advertising to pay for, the contract funds the engineering, the customer comes back.

• Commercial air transport: think airlines and cargo. Big customers with big fleets, on frequent and heavily regulated maintenance cycles.

• Business aviation: aircraft or small fleets owned by individuals and companies. Everything is smaller (the aircraft, the customer, etc.), and the market is far more fragmented.

ISSC supplies OEM parts to all three markets. Which means it supplies their MRO and retrofit too.

The big picture is clear enough. Time to get into the details.

Make the time spent on these visuals worthwhile. Share this article with your friends and colleagues. Let's make Undiscovered Compounders a little bit more undiscovered.

Share

2.2 The Proprietary Cockpit Stack

Historically, ISSC’s proprietary business is made up of 6 certified product categories, all manufactured in the one and only factory in Exton, Pennsylvania:

• Sensors, to collect information (altitude, speed, etc.);

• Displays, to show it;

• Communication, to communicate;

• Navigation, to navigate;

• Actuators, to activate the physical flight controls;

• Control Systems, to run it all.

The details matter little. The central idea is that its products cover most of the building blocks of a modern cockpit in the segments where ISSC currently operates, with one exception. An in-house autopilot is missing. Or rather, was. We will come back to that.

Among them, one product is nonetheless worth detailing, because it demonstrates what ISSC intends to replicate in the coming years.

2.2.1 ThrustSense Autothrottle

In 2016, ISSC identified a need the market had not met: no autothrottle for turboprop aircraft.

Very roughly:

• Turboprop = propeller + turbine, slow, short-haul, economical.

• Jet engine = pure jet, fast, long-haul, more expensive to operate.

The autothrottle is what allows engine power to be managed automatically throughout the flight, without the pilot having to touch the throttles manually (takeoff and landing included). That lifts an enormous cognitive load off pilots, and takes away one more chance of a serious, even tragic, mistake.

Every modern commercial jet has one, but not turboprops. Such is capitalism: no manufacturer had ever deemed the market attractive enough to invest in.

ISSC sees the need and, confident in its engineers and its company culture, invests. Three years later, in April 2019, the FAA certified its ThrustSense Autothrottle for retrofit on the King Air.2

It thus became the first autothrottle in the world certified for a PT6 engine (Pratt & Whitney), meaning the engine that powers the vast majority of civil turboprops. In practice, being certified for PT6 means covering most of the global turboprop market. That’s all. To this day, ThrustSense is the only FAA-certified autothrottle for PT6.

After that, foreign certifications follow quickly, and Textron, one of the major players in the sector, signs a multi-year OEM contract to integrate ThrustSense into the newly manufactured King Air 260 and 360. ThrustSense is therefore integrated into older turboprops and new ones alike.

The military spots the opportunity and selects ThrustSense in late 2024 for its militarized version of a civil turboprop.

In reality, the military had no choice. There is currently no other autothrottle for PT6 turboprops. Developing one would take years and a string of certifications.

Here, it is clearly worth acknowledging the prowess of the current CEO, Shahram Askarpour (Vice-President of Engineering until 2012, then President until 2022). He is the one who spotted the need, who had the technical vision and the engineering skills to create this product.

And his transition as CEO has not stopped the company from preparing for the future.

2.2.2 The Future

Two significant products are in the pipeline.

Utility Management System 2 (UMS2), the next generation of the UMS delivered on Pilatus aircraft. Think of the UMS as the brain of the cockpit, monitoring and controlling most of its other components. The main addition in this new version is the integration of a neural network that brings AI to increase the level of cockpit automation. The first UMS2 deliveries to Pilatus for their PC-24 model are expected in the coming weeks. Seen that way, it looks like nothing more than a product upgrade that will at best generate a bit more revenue. But the UMS2 is platform-agnostic. It can also serve business aviation and the military, subject to regulatory approvals that will take time. AI, and a new pool of potential customers. And that is not the biggest innovation.

Liberty Flight Deck (LFD). The “Deck” here is the most important part. Before the LFD, ISSC sold cockpit components here and there, one autothrottle for X, one UMS for Y, and so on. The LFD is literally a complete, turnkey cockpit, with all systems talking to each other natively. And here, ISSC takes a different route. Other competitors already sell complete flight decks (Garmin, Honeywell, Collins), but most remain largely proprietary. You buy the whole ecosystem, or nothing. LFD is fully open-architecture. You can connect competitors’ products to it, you can upgrade it modularly. That is a massive commercial argument, especially when clients (airlines and operators) complain at every opportunity about lock-in by the major integrators.

“The Company believes that its new Liberty Flight Deck (LFD) will be a key driver of its next phase of growth.” 10-K FY2025

First commercialization targeted for 2027 in the business jet aftermarket, and 2030-31 for OEM platforms.

The last comparable phase of innovation dates precisely from the early 2000s. So, why a turnaround 20 years later?

This is going to sound a bit harsh, but I think the main cause is the death of founding CEO Geoffrey Hedrick in early 2022. He had a vision when he created his company, a vision suited to his time, and it worked. He was also an engineering genius, with over 100 patents to his name. But the market evolved, and he failed to adapt. The company struggled from 2006 until his death. His capital allocation was equally questionable, far too conservative for this sector. On several occasions, reading through old analyst commentary, I came across recommendations urging strategic shifts that he dismissed out of conviction.

The arrival of the former VP Engineering, an engineering genius (yes, him too), Shahram Askarpour, as CEO was the triggering event. He sought to reorient the product mix and target military clients, put a strong focus on innovation, and used his network and reputation to make acquisitions. And it shows perfectly in the non-financial data. Engineering headcount has grown by more than 50% in each of the last two fiscal years. The number of patents has more than tripled in 3 years (international & trademarks included). Management and the core engineering team average more than ten years of tenure at the company.

The 30x from the Covid trough to the 2026 high is not a consequence of a 37-year-old business model. It is the reward, earned year after year, of an exceptional management team and its employees who transformed this company from a barely surviving business into a recognized player.

Shahram Askarpour has transformed the company in a few years, and I am convinced he is the primary architect of its success.

That is why the avionics giants chose them when it came to selling their orphan product lines. They needed a credible buyer, and ISSC had become one.

2.3 The Acquisition Engine: Industrial Recycling

Acquiring orphan product lines from sellers for whom the sale is immaterial, and who want quality buyers to avoid alienating their customers.

2.3.1 The Seller

Of 6 acquisitions, Honeywell Aerospace is the source of 5 of them. Honeywell Aerospace is a $17.4B revenue business in 2025, currently being spun off under the ticker HONA, expected by Q3-2026.

For HONA, a product line generating $5M or $10M per year, even a profitable one, is a rounding error. But a rounding error that ties up resources (engineers, certifications, factory space, etc.). For an efficient capital allocator, the equation is relatively simple: these product lines are generally worth more sold than kept.

But that was the old reality for Honeywell. Honeywell Aerospace is being spun off in the coming months. To understand the impact this will have on ISSC, it is better to treat this spin-off as a source of answers rather than questions:

“Honeywell has been over the past couple of years divesting a number of their avionics product lines. The exact nature of their purpose they will not share with us, other than these are the kind of the previous generation of the product lines and now they’re moving on with their latest and greatest.” Conf Call June 2023, CEO.

3 years later, we have the answer. All those divestitures were preparation for this spin-off. And the Form 10 adds even more clarity: HONA intends to focus on its recent innovations. I use the term “innovation” again, but we are objectively talking about a different level of technology than ISSC’s (to be fair, ISSC is a dwarf compared to HONA): a touchscreen cockpit with native cloud, software to navigate in GPS-degraded environments, and even electric propulsion systems. The future, basically.

On the other hand, an autopilot designed 20 years ago installed on tens of thousands of civil aircraft (KFC 325) is not the future, nor Honeywell’s priority. Same goes for military display generators on aging F-16s (still in production 50 years later), or electrical generators on even older F-15s. That is precisely why it has been selling them to ISSC since June 2023.

Once listed, the incentives will be even stronger. HONA will be scrutinized quarter by quarter on its margins and strategic focus, and anything that is not the “future” stands a good chance of becoming sellable.

Obviously the pattern is not specific to Honeywell. It is in fact a classic cycle among the aviation giants: they acquired a line 15 years ago, integrated it efficiently, then moved on to the next generation, which eventually made it irrelevant. It may be awkward to have in the portfolio, but the line exists physically, and the aircraft that carry it and need maintenance or even part replacements also exist, and often for decades. They certainly want to get rid of it, but they do not want to sell it to a company that will manage the line poorly and anger their clients (those clients very often remain clients on other products, it is a very closed ecosystem), and they choose, in their view, the best buyer to protect their reputation.

When Moog sells its S-TEC line (autopilot product line, well played if you have been following) in February 2026, it does not explain why in detail, but this is precisely the mechanism I just described.

2.3.2 The Product

First important point: ISSC only acquires product lines certified by the FAA (USA) or EASA (European Union), or both. Without that, the aircraft has no right to fly.

Second important point: these certifications do not transfer automatically. When Honeywell sells a product line to ISSC, ISSC must obtain its own TSO authorizations (Technical Standard Orders, the FAA’s minimum performance authorizations required to manufacture and sell specific avionics equipment) for each product (still simpler than a full certification from scratch since the design data already exists). Askarpour estimates this at around one year per transaction to migrate all TSOs. Obviously, no newcomer can enter a product market without having paid this entry cost (both time and money), for a market that often represents only a few thousand units per year at the end of the chain.

And these aircraft fly for a long time. A very long time. ISSC’s 10-K states it clearly: its products equip aircraft “in service for up to fifty years.” A King Air rolling off Textron’s production lines in 2026 will need ISSC-certified parts in the 2050s. And for the lines Honeywell manufactures today and will sell off in 15 or 20 years, there will be several more decades of demand after their sale. Concretely, ISSC’s future acquisition pipeline is already on Honeywell’s assembly lines.

To top it all off, the installed base is aging. The average age of the global commercial fleet reached 14.8 years at the end of 2024 (a record), and the curve has been rising for 20 years. Blame the airlines that do not have the balance sheets to renew at the historical rate (or rather a supply chain that extracts maximum value upstream and a hyper-competitive sector). So you keep the aircraft, you replace the parts, and that is exactly where ISSC fits.

2.3.3 The Buyer

On the first deal in June 2023, Honeywell received 7 bidders. Of those 7, ISSC was the one chosen, as it was for the four following acquisitions. ISSC's CEO puts the choice down to one thing: “We had unique capability at IS&S that made us a very attractive buyer.” Fine, but which ones?

First, ISSC manufactures 100% of its products in its Exton factory in Pennsylvania. A major difference, given that nearly the entire sector outsources a significant portion of its subassemblies. Less subcontracting = better margin on products, which is good news for ISSC, and for Honeywell. Obviously, ISSC gets new product lines that bring in a material amount of cash flow, and Honeywell gets rid of lines that are too old (whose customers sometimes complain that they had been neglected by Honeywell, with delays, lower quality, etc.) and finds a player for whom those lines matter and who has every incentive to manage them effectively. It is even one of the main levers that drove ISSC’s Adj EBITDA from $11M in FY2023 to $25M in FY2025.

“Typically Honeywell has outsourced a lot of these manufacturing of their subassembly. And there are opportunities there for us to take a look at some of the outsourcing they do because we have full in-house capability.” Conf Call, June 2023, CEO.

Second, ISSC has room. A lot of room. ISSC prepared the reorientation of its business model toward M&A and optimized the setup of its current factory, doubling the floor space and tripling production capacity in anticipation of this M&A. Honeywell therefore finds itself facing a player that has already prepared the ground and is ready to integrate these lines quickly and efficiently.

Third, they showed they were capable. Once the first transition was completed, then the second, ISSC built a reputation as an efficient acquirer and integrator. Once again, the aerospace world is a fairly closed one, everything gets around quickly, and ISSC’s reputation also reaches the 5-6 other giants ready to get rid of their lines. Moog’s divestiture of its autopilots to ISSC in 2026 is in part a reward for ISSC’s previous successes.

ISSC deserves credit for its execution, but it did not invent the business model.

2.3.4 The Precedents

To understand ISSC’s future potential, you have to look at those who came before.

TransDigm ($80B) and HEICO ($40B) spent decades doing something similar to what ISSC does.3

TransDigm acquires companies that own proprietary and sole-sourced aircraft components while HEICO is best known for creating alternative parts that compete with the OEM. Both exploit a similar asymmetry to ISSC: products that don’t die, margins majors tend to leave on the table, customers who pay for reliability, all in a niche, heavily regulated sector with recurring revenue.

Can ISSC become the “next” TransDigm or HEICO? The current math allows it. My read: very unlikely. Three questions crystallize the uncertainty:

• Does the company have the human capabilities to execute this transformation?

• Will ISSC be able to make enough acquisitions for that?

• Will the acquisitions be sufficient?

The first is too uncertain to attempt an answer. The only thing to say is that so far, management has executed the transformation very well over the last 3 years. But there is a real key personnel risk with CEO Shahram Askarpour, who is 68 years old. HEICO has had and will have multiple generations of Mendelson. More on that later.

For the second question, one argument is worth (re)mentioning. The global fleet is aging, increasing the need for MRO and retrofit of older aircraft, exactly the kind of lines ISSC targets through these acquisitions. So demand for these products has a future. On the supply side, today’s new aircraft will become tomorrow’s old ones, and therefore a new pool of lines for ISSC. Relevant, yes. Sufficient to cross tens of billions in market cap, hard to tell. No doubt it will be more complex than acquiring private companies outright, like TransDigm does.

On the other hand, ISSC has something HEICO and TransDigm don’t have at the same level: a clean-sheet R&D engine on top of its M&A pipeline. Its own proprietary R&D has historically been a source of revenue and cash flow, and the company continues to fill its innovation pipeline for the years ahead with UMS2 and the Liberty Flight Deck.

The third question is the most important but also the most complex. Clearly, ISSC doesn't have TransDigm's pricing power and probably never will. The TransDigm level is, in my view, out of reach for ISSC. On HEICO, I'm less categorical. Obviously, the comparison isn't perfect, particularly on the Parts Manufacturer Approval segment, where ISSC has no presence. But ISSC's model has its own advantages, notably enormous cross-selling opportunities. The “yes” isn't obvious, but neither is the “no”.

All of this remains speculation over several decades, and obviously won’t constitute a bull case, but the pattern was worth mentioning.

2.4 The Synergies Between the Two Models

This is what makes the ISSC opportunity so attractive and, in my view, completely underestimated by the market.

3 major synergies stand out.

2.4.1 Cross-Selling

By acquiring these product lines, ISSC enters into commercial relationships with new clients and can also offer them its other products. In a sector this conservative, it’s a way to expand its customer base without having to spend millions on marketing and customer relations, and above all, a lot of time.

This applies to existing products:

“We benefit from deep customer relationships which have further strengthened following the Honeywell transactions. This gives us the opportunity to cross-sell our broad portfolio of existing products to customers. As an example, some of the radio products recently acquired from Honeywell were coupled nicely with our cockpit offerings in the C-130 platform.” Conf Call Q1-2026, CEO.

But also for future products. When Liberty Flight Deck is certified for the business jet aftermarket, the customer base will already exist and be well established. Same for UMS2, which is platform-agnostic. ISSC innovates where demand meets expertise. My guess is that cross-selling drives more of those R&D decisions than the filings suggest.

2.4.2 Portfolio Completion

Earlier, I told you that ISSC had every product needed to offer an integrated cockpit suite except one: a certified autopilot.

The gap is enormous. The FMS calculates the route, the FPDS displays it, the UMS controls all the subsystems... but no autopilot to fly the aircraft. Not for Part 23, and even less so for Part 25:

• Part 23 is the standard for light aviation (turboprops, light jets, etc.). These aircraft carry fewer passengers (≤19) and are less regulated.

• Part 25 covers large business jets and large commercial transports (A380, Boeing 737, etc.). Extremely regulated, but significantly larger in value than Part 23.

To create a certified Part 25 autopilot from scratch, you are looking at 5 to 7 years from design to first sale, and tens of millions of dollars out of ISSC’s pocket. Or rather, used to be.

In February and March 2026, ISSC acquired two complementary autopilot product lines for approximately $25M cash combined. Together they cover the spectrum from small General Aviation (all civil aviation except airlines) aircraft up to certain Part 25 applications. Among others, they acquired the lines for:

• The KFC 230, a next-generation digital autopilot developed by Honeywell only 4 years ago. It becomes “the workhorse for our own autopilot.”

• The KFC 325, whose autopilot is installed on tens of thousands of aircraft. Aircraft to which ISSC will now be able to sell its UMS2 and Liberty Flight Deck once certified. And above all, an enormous client base for future MRO and upgrades.

For $25M, ISSC bought future clients and everything needed to offer a full cockpit suite to those same customers. According to its CEO, with these acquisitions it have become “probably the largest autopilot supplier right now in the market covering the spectrum of aircraft”.

2.4.3 More Cash Flow to Finance R&D

You make avionics components, you watch the market and its inefficiencies, and you find ideas that could address them and that require R&D. You go to a bank to borrow a few million and explain that you want to borrow to fund R&D that may or may not turn into future cash flow, and if it does, only in a few years (regulation doesn’t rush). Best case, you get a polite “no.”

You want to buy product lines that the giants want to get rid of, often at low prices. You have the expertise and infrastructure to do it. You go to a bank, explain the situation, the quick cash flows, and it lends you $20M over 5 years. Perfect. You buy the line. You integrate it quickly, generate your quick cash flows, to the point where you pay back the debt in a few quarters. You repeat this across several deals, and each time, you allocate part of the cash flows to R&D. In the end, you’ve managed to double your R&D budget in 3 years, and the cherry on top is bigger than the cake: you have more revenue and cash flow even without the R&D needing to pay off.

You are now able to innovate significantly more and create a new generation of products.

The first scenario is obviously fictional, and ISSC did not make these acquisitions in order to fund R&D more freely, but it shows how the acquisitions enhance innovation and therefore ISSC’s historical business.

To temper expectations, this synergy is probably limited at this stage. UMS2 started being funded before the acquisitions. Liberty Flight Deck is probably the only product directly benefiting from this R&D facilitation. What is certain, however, is that ISSC now holds all the cards to focus on R&D when it finds an opportunity worth the time and money.

Now that we have a clear picture of the business model, we can look at the acquisitions.

3. Acquisitions

6 deals in 3 years for a total of $87.7M cash. 5 were done with Honeywell and 1 with Moog.

Analyzing them one by one would be the best way to miss the strategy behind these acquisitions, but a short summary of each doesn’t hurt.

These 6 acquisitions can be split into 3 waves.

3.1 Wave 1, First Steps

Some context. In June 2023, ISSC sat at $34M of revenue, $11M of EBITDA, and a market cap of around $100M. The deal itself was $35.9M. A good third of its market cap. For a microcap that had been going nowhere since 2006, this was objectively a big bet. The risk was partly cushioned by the financing: $20M in term loan and the rest from its own cash.

With that, ISSC bought lines of inertial reference units (IRU, used to calculate trajectory data without GPS) and communications/navigation radios installed on thousands of aircraft (mostly Part 25). Why these lines in particular? For 3 reasons.

First, the basic financial logic is attractive: ISSC had about 50% of its production capacity sitting unused at the time, and these new products require the same infrastructure and the same engineers as the existing ones. So integration costs are low and integration is fast. On top of that, post-integration margins on these lines are similar to ISSC’s margins at the time. The lines can generate meaningful cash, and quickly.

Second, the logic is also strategic: portfolio completion & cross-selling. At the time, ISSC had a “limited capability” in inertial and comm/nav. This acquisition allowed it to complete and optimize the portfolio quickly. And these lines serve thousands of aircraft, which gave ISSC a network of customers it could never have built organically.

One last detail makes these lines particularly attractive for ISSC. An IRU or a comm/nav radio installed on a Boeing 737 never stays in service for 25 years without intervention. They need MRO. If Honeywell wasn’t interested in these lines anymore, it was even less interested in the MRO around them. Honeywell had stopped servicing this market for these lines years ago. An operator had to look for providers certified and authorized by Honeywell for MRO on these lines, and according to customer feedback, the offering was clearly insufficient.

For ISSC and its $100M market cap at the time, this was a windfall. So ISSC bought the exclusive IP rights and is more than happy to service the MRO on these lines. Every time an aircraft needs MRO, ISSC is the only authorized supplier. A perfect example of the better owner principle in M&A.

Theory is clean. Practice too.

A word on Adjusted EBITDA, because I can sense it irritates someone. I'm going to use it throughout. Management uses it for almost everything, and it's the clearest read on the business model once you strip out the noise.

Management's definition is surprisingly honest: net income before interest, taxes, D&A, transaction-related acquisition and integration costs, and non-recurring items. And yes, SBC is left in (Honeywell, for one, strips it out…). Nothing is swept under the rug. It's just the least-bad way to compare ISSC's performance after one acquisition or after three.

In either case the adjusted figure stays reasonably close to reported EBITDA: the spread between adjusted EBITDA and reported EBITDA has ranged from 4% to 14% over the last three fiscal years.

Management was projecting (on a trailing twelve months basis): +40% revenue, +75% adjusted EBITDA, and EPS accretive (one would hope). Based on these numbers and ISSC’s pre-deal reality, you get incremental EBITDA of $8M and therefore a payback in 4.5 years. To sum up, the line was acquired at a multiple of roughly 4.5x EBITDA.

Reality delivered: revenue +54%, adjusted EBITDA +75%, and EPS +28%. Clearly an economic and strategic success.

In July 2024, ISSC put another $4.2M on the table for a complementary comm/nav line. Just a small asset acquisition (no goodwill) that follows exactly the same logic as the first.

At this point ISSC has shown three things:

• It has a strategic vision and is capable of executing it efficiently.

• Even a modest pace of future acquisitions can structurally change the P&L.

• It is ready for wave 2.

Wave is clearly an understatement. The impact looks more like a tidal wave. It radically transformed ISSC.

You can't understand the company as it is today without grasping everything its next acquisition implies.

3.2 Wave 2, Great Leap Forward

This is where it stops for free readers.

Going paid gets you 20+ deep dives a year, the follow-ups that come with them, and every single deep dive I've already published.

I won't do too much marketing here. I'll let my work speak for itself. So I'm giving you one free post. On the house.

Read more

This single sentence says a lot more than it seems.

In case you missed it, here's my full deep dive on Atlas Salt, one of my largest positions.

The thesis fits on one page, but the conviction needed to hold it requires a deep understanding of the company and its market.

Don't trust me. Check my arguments, my hypotheses and my results.

The Detail

Atlas agreed to issue 8,333,400 common shares at C$1.20, for C$10,000,080, on a bought-deal basis co-led by Ventum Financial and Raymond James. The underwriters hold an option on another 1,250,010 shares at the same price (about C$1.5M), exercisable until 48 hours before the closing, which is expected around June 11. [These were the terms as announced. The deal was upsized the next day and closed at 12,628,000 shares for C$15,153,600; see the note above.]

As in October, these shares carry no hold period: they are free-trading from closing. So there is no deferred "wall," that moment, four months later, when a hold expires and everyone can sell at once.

After a 6% cash commission, the C$20,000 corporate-finance fee and roughly C$160,000 of legal and accounting costs, net proceeds come to C$9.22M (about C$10.6M if the option is exercised in full).

In reaction, the stock pulled back within the range of the dilution (7.5%, up to 8.6% if the option is exercised), no further [corrected: actual dilution 11.4% at final size, see note above]. On the face of it, the market reacted well.

But I don't trust the market.

The offering document gives us a good deal of important information. Let's dissect it.

What These Funds Are For

These available funds, about C$13.9M (the net proceeds plus existing working capital), are budgeted for the next 12 months. Here’s the breakdown:

Half of it looks dull. C$3.5M of G&A, C$3.19M of unallocated working capital: together, 48% of the available funds. I can already hear the skeptics: “they raised the money to keep the lights on.” Yes and no.

The C$3.19M isn’t dead money. The LIFE exemption forces the company to certify it has twelve months of liquidity, enough to cover what it needs to keep running, and that unallocated line is the buffer behind it. The rest arms a company preparing a C$589M build (staff, advisors, the groundwork on the commercial agreements, etc.).

The last two lines aren't diligence overkill. In fact, they are the most important ones.

The C$0.25M roughly pays for lender due diligence and the data room behind the construction financing: financing is advancing.

Above all, the C$0.5M. The document is specific: it covers the "advancement of strategic commercial arrangements with key project partners, including Sandvik, Hatch and Continental Conveyor."

More precisely still, it talks about:

• “Advance procurement planning, equipment studies and commercial arrangements for major project packages.”

• “Support long-lead procurement strategies and technical-commercial coordination with strategic project partners.”

In other words, Atlas is developing its commercial arrangements with each of the construction partners with which it has signed a specific MOU: Sandvik, Hatch, Continental. Only Scotwood is missing. That's normal: Scotwood is an offtake partner, relevant closer to production. More than three years out, the opposite would have been surprising.

One last line, among the C$1.2M: “Consolidation of geotechnical and hydrogeological interpretation and field sampling to finalize design parameters.” In plain terms, Atlas is following SLR's advice and deepening the hydrology work to limit one of the major construction risks.

In short: two tiny lines in dollar terms, but this is where the next two catalysts are being built: the senior-debt term sheet and the conversion of the MOUs into binding contracts.

What's below is for paid subscribers.

Going paid also unlocks the full back catalogue: every deep dive and study published so far, including the piece I think holds the most alpha of anything I’ve written. For now.

One undiscovered company at the right price makes the cost of this newsletter trivial.

Read more

Once the junk is cleared out, I look for the 1% worth committing capital to, and under what conditions.

At this level, there’s no filter left. It’s dozens of hours digging through every single filing to find the detail the market missed. If it exists.

Here are 5 companies that made it through all these steps, why they did, and what I’m waiting for before I pull the trigger.

Subscribe now

A Niche Compounder, One Brussels Vote Away From Greatness

I-Tech AB; Ticker: ITECH.ST; Price: SEK 63.3; Market Cap: SEK 760M; Distance from ATH: -50%; Country: Sweden.

A ship needs a very smooth hull to move fast and cheap. Nature, of course, couldn't care less: dozens of organisms have a nasty habit of latching onto the hull. Everything that clings to it is called fouling. A hull full of fouling means up to 40% more fuel and slower voyages. When globalization depends on those voyages, efficiency isn't optional.

I-Tech sells Selektope, a product that targets one of the main sources of fouling: barnacles. It's the only antifouling on the market that is, all at once:

• non-metallic (most antifoulings rely on copper, which is increasingly under fire);

• non-lethal (it overstimulates barnacle larvae and stops them from settling);

• effective at a tiny dose (0.1%), in a global regulatory environment increasingly worried about biocide concentrations in hull paints;

• effective even on a stationary hull, where most antifoulings lose their bite once the ship stops moving.

I-Tech sells Selektope to nearly all the big marine coatings makers, who then resell the paint that contains it. And once Selektope is in a formula, there's no getting it out.

Selektope is on just about 3,000 ships, out of a global fleet of roughly 100,000 (a TAM clearly inflated by management, but the point stands: the room to grow is enormous).

A hull keeps its paint for about 5 years before the next drydocking (yes, that's recurring revenue). Selektope has been through only about two cycles so far. In an ultra-conservative industry, allergic to the slightest change, two successful cycles start to build a track record that can win over even the most reluctant buyers.

Almost all of its revenue comes from Asia, where the paint makers and the shipyards are. Barely 2% from Europe, and yet Europe is the source of I-Tech's biggest problem: regulation.

In late 2025, the European Commission proposed not to renew the approval of medetomidine (Selektope’s active ingredient), suspected of being an endocrine disruptor (which I-Tech disputes). The stock is down about 50% from its peak.

If the EU doesn't renew the approval, two major consequences:

• European approval often carries weight beyond the EU. If Europe says no, other regulators could be tempted to follow.

• Without approval, Selektope could no longer be placed on the EU market, and a ship carrying it would become a headache to sell to a European owner or re-flag in the EU. And ships travel: over 25 to 30 years, they change hands and flags more than once. So it's a huge incentive for every paint maker not to offer Selektope to a client whose ships might one day end up European. And that "might" matters. Even if Europe isn't in the plan, the "just in case" often wins the owner over, and it costs the coatings maker almost nothing to sell a different paint instead.

Approval = thesis largely intact. No approval = a serious setback for the business model and its future. In between, a whole spectrum of outcomes (conditional approval, limited duration, and so on), each with different consequences for I-Tech. For now, the business is holding up.

The company is refocusing on R&D to diversify. But that’s too many “ifs” to build a thesis on today.

The final decision was expected around mid-2026, after a member-state vote in the first half of the year. But the approval has just been extended to 31 December 2026, because the assessment couldn’t be finalized in time. In the meeting minutes, some heavyweight member states come out flatly opposed. That said, this latest extension may be a sign that consensus is hard to reach, and that “no” is less of a foregone conclusion than it looks. What’s almost certain: if approval comes, it will be time-limited (say, 4 to 8 years) and conditional.

Trigger: a non-approval is a dealbreaker for me, and right now it's the most likely scenario. If approval comes, I still need to see the duration and the conditions. The only certainty today is that we’re dealing with a capable management team and an exceptional product, with the makings of a niche compounder that could grow into a very nice mid cap. If the EU is willing, of course.

Got something out of this? Do one thing: hit like and forward it to one investor who'd want to read it. That's how the undiscovered get discovered.

Share

A Toy Store With a Moat That Beat Nvidia

Build-A-Bear; Ticker: BBW.NYSE; Price: $37.21; Market Cap: $468M; Distance from ATH: -50%; Country: USA.

Originally, Build-A-Bear sold build-your-own stuffed animals through a ceremony that mimics a birth. After serious trouble tied to a very asset-heavy balance sheet and a former CEO’s strategy that had grown ill-suited to the market, Build-A-Bear pulled off one of the most value-creating moves there is: going from asset-heavy to asset-light. From the COVID bottom to its late-2025 ATH, the stock did a 75x in five years, outperforming Nvidia itself.

I owned Build-A-Bear myself for a good part of that run, and it’s my biggest gain to date. I’ve been out of the stock for over a year, and that could change soon.

To understand Build-A-Bear, look at its moat: it sells an experience. You come in either to build your own bear or to immerse yourself in a world, that’s the reason 80% of store visits are planned.

This moat comes with three big advantages.

First, when mall traffic collapses, an impulse store dies with it, while a destination store like Build-A-Bear holds up far better. During the FY 2024-2025 period, BBW’s traffic was down less than 1% versus 5% for a national benchmark. Of course, mall owners know it and actively seek out a Build-A-Bear for their mall, which obviously gets BBW ideal lease terms.

Second, the moat strengthens itself. The thirty-somethings who built their bear twenty years ago now bring their own kids, and buy one for themselves. This Kidult segment has taken off and now accounts for 40% of revenue. Good luck to the competitor who wants to rebuild thirty years of memories.

Finally, what retail customers love, companies want. That moat has opened partner-operated locations and franchises on nearly every continent (38 countries vs. 19 two years ago). CapEx is limited, margins are far higher, and it has been the growth engine of the past few years. The same argument works for the wholesale segment, where Build-A-Bear sells its products in more than 1,500 Walmarts with double-digit growth.

Financially, the balance sheet is clean. The company generated around $40M of free cash flow last year and used part of it to deliver an 8% shareholder yield. It learned from past mistakes and conflicts with former shareholders, and has bought back 25% of its shares since 2019. It's safe to call it shareholder-friendly. My only gripe: I'd like to see it ease off the buybacks when the stock runs hot.

Two reasons explain this -50%:

• Tariffs cost $4M in 2025, with $10M guided for 2026.

• US consumer confidence (the bulk of revenue) is at rock bottom. During Q1-26, domestic traffic fell 7% and e-commerce demand fell 26%. Almost every retailer has been punished over the past few months.

On tariffs, management has offset part of it with selective price increases. The macro backdrop limits how far that lever can go for now, so those increases haven't fully covered the tariffs.

On the consumer, I think the market overestimates the impact for Build-A-Bear:

• Build-A-Bear's customer base is more economically stable and diverse than the average retailer's. So even with traffic down 7%, a higher dollar per transaction held net retail sales to around -5%.

• The e-commerce drop mostly comes down to Google's AI search problems. It's a known issue, and management has hired and lined up partnerships to fix it.

• The capital-light features (franchise, royalties, and so on) offset the weakness in direct-to-consumer. That's why it printed a solid Q1-26 despite all the domestic headwinds (still down 2% versus the record Q1-25, but holding). Very few of its publicly traded competitors can say the same.

It's not all rosy right now, and the company is guiding for continued DTC pressure through the rest of the year. But even after cutting its guidance for the year, it's still aiming for its best year ever in terms of revenue.

We’re potentially looking at a cyclical trough, offering a nice buying opportunity for a company that holds every card to weather a longer-than-expected cycle and to thrive for years.

Trigger:

That's two of the five, and you already have the exact trigger on each.

The other three are the ones I rate highest:

• An exceptional company, and my way to play the biggest sector conviction I hold for the coming decade.

• A monopoly compounding in plain sight for years, run by one of the best management teams I've ever seen.

• A Japanese compounder that created its own sector, now at the cheapest multiple in its history while it prints record quarter after record quarter.

Each gets the same treatment as the first two: why it survived the filter, and the precise price or event I'm waiting for before I buy.

All of it is for paid subscribers.

I don't think this post alone should convince you to become a paid subscriber. Take your time, check my work. It's my best marketing tool.

There's a deep dive on Build-A-Bear right above, but here's another one if Build-A-Bear doesn't interest you:

Read more

Too bad for us, they’re a constant. Even with perfect foresight, you'd still face 76%. The only variable left is you.

You open your broker app, your portfolio is down too many percent. Red pixels bombard your retinas, your finger hovers over the sell button and… calm down.

Because our good old friend Science is already on it. She named our flaws one by one, and handed us the patches.

Subscribe now

From Red Pixels to Capital Loss

Evolution has had exactly one goal for the last few million years: to get us through the next few seconds, over and over, until we reproduce.1

But when you apply tools custom-built for the next few seconds to decade-long objectives like investing, the damage is inevitable.

There are three culprits. Or rather, two accomplices and one executioner.

Accomplice #1

It’s just an accomplice, but an accomplice in almost every crime linked to bad decisions.

You’re bound to know it, you experience it dozens of times a day. The aptly named: anchoring bias.

What your brain records is a deviation from a reference point, never an absolute price.2 And with every new high, that reference point resets.

Here is the experiment through which Kahneman & Tversky established it in Prospect Theory (the work that destroyed the homo economicus hypothesis, no less).

They showed that perceiving a situation as a loss feels psychologically twice as painful as perceiving it as a gain, even when the expected value is the same: $1,500 in every case.3

“That’s just psychology students making bad decisions. I’m an investor. Being rational is literally my job.” Okay.

You buy a stock. At random, Atlas Salt. You buy it at $1, it goes to $2. Great, you made a 2x. But you’re a long-term investor (or greedy, your call), so you don’t sell. The stock falls back to $1. So what, you’re flat, right? You sure? Not a single part of you would say: damn, I could have sold at 2x?

Of course you would. This dependence on the reference point is coded into our genes. This simple mechanism is probably responsible for trillions in losses over the years. (I have no proof, it’s just a deep conviction.)

In everyday life, it tends to fly solo. But in investing, it has a very close friend. And that’s our second accomplice.

Accomplice #2

Its name is myopic loss aversion.

If the first one creates the pain, this one multiplies it.

Loss aversion is our natural tendency to suffer more from a loss of X than to enjoy a gain of X (another gift from Kahneman & Tversky, they earned that Nobel Prize).

The myopic part emerges when evaluation frequency gets too high. The effect plays out in two steps: the more you check, the more frequently the reference point shifts, the more the pain multiplies.

With markets trading continuously for 6+ hours a day, five days a week, the myopic part has everything it needs to thrive.

Take a negative annual return of -30% and show it to two groups.

• Low Frequency group (LF): sees the performance once a year.

• High Frequency group (HF): sees it once a month.

Both groups are subject to loss aversion (they’re all human). The only variable is frequency, and therefore the impact of myopia. Result: the LF group was willing to invest 69.6% in stocks vs 40.9% for the HF group. A risk aversion almost twice as high, just from seeing the same -30% more frequently.4

And that was the beginner version of the experiment. In reality, investors can check their portfolio multiple times a day, and more importantly, they can buy or sell. That’s where the real culprit steps in.

The Culprit

The executioner finally takes the stage. Its name: the disposition effect.

No need for imaginary skeptical investors here. It has been documented across tens of thousands of real investors, from the beginner who barely knows where to click to the institutional manager running a billion-dollar portfolio.

The deeper investors were in drawdown, the more they acted, and those actions concentrated on keeping or buying more losers and selling winners to compensate.5

If you have any doubt about the quality of that strategy, let me remove it: it is the worst of all. I simulated millions of portfolios to analyze the evolution of opportunity cost, and it is by far the worst strategy. As for the best one, you’ll find the full post here.

One last finding. You have to appreciate the irony: selling randomly would have outperformed their actual sell decisions. A finding that also came up in my own simulation.

Investors spend hundreds, sometimes thousands, of hours a year analyzing companies, and yet a random strategy would have outperformed them on average.6

But it is what it is. Losers tend to keep losing, and winners tend to keep winning.

Between 3 and 4% of annual returns evaporated (or rather, transferred) because of the disposition effect.

The anchoring bias and myopic loss aversion are the two sidekicks pushing toward violence, but it is the disposition effect that pulls the trigger.

Less trivially:

• Anchoring bias: We observe prices and performance from a reference point that shifts with every new comparison.

• Myopic loss aversion: The more we compare, the more any given move carries a disproportionate impact.

• The disposition effect: The deeper the drawdown, the more we tend toward the most value-destructive behaviors.

Diagnosis complete. On to the cure.

The Patches

It’s psychology, there’s no magic trick on its own. Just a few tricks that become magic with a good dose of discipline.

Patch #1

Let’s start with Kahneman’s own favorite method, “invented” by Gary Klein: the pre-mortem.7

Before a buy, you assume you will eventually sell, with certainty. Then you list as many scenarios as possible and sort them into two piles:

• The broken thesis pile

• The noise pile

As soon as a scenario unfolds, you check, and based on which pile it’s in, you sell or you hold.

What, sounds too basic? So what? Human psychology tends to be too.

Two groups of participants are asked to imagine causes for a future event (e.g., why would I sell?), with only one change in the prompt:

• Group 1: the event might happen.

• Group 2: the event will happen, with certainty.

Just by changing the framing from a possible event to a certain event, Group 2 generated 30% more reasons.8

The trick is simple and effective. But it’s clearly not enough.

Patch #2

It’s my second favorite patch.

People are asked to manage a random project with an allocated budget. Everyone goes through 3 phases: positive feedback, then mixed, then downright negative. The idea: build their confidence, then discourage them, to see how far they’ll take the investment project.

Small twist that changes everything, the base instruction varies (again):

• Group 1: “We want to judge the project as neutral observers, regardless of past decisions.”

• Group 2: “If we are about to make an investment decision, then we will judge the project as neutral observers, regardless of past decisions.”

Group 1 sinks into escalation of commitment despite the negative feedback. Only Group 2 escapes it.9

Yet the only thing that changes is the wording, from vague goal to if + then. And that’s the entire power of implementation intention.

By placing the environment inside a specific "if" and tying it to a precise action through "then", you close the door on dozens of cognitive biases.

It’s prompt engineering, but for biological intelligence.

Patch #3

Myopic loss aversion comes from feedback frequency. So if a cause leads to a negative consequence, just remove the cause.

That’s exactly the advice from Benartzi & Thaler (from accomplice #2): focus on information design. Build an information system where feedback access is limited.

It can be many things: deleting apps with live price quotes, logging off once you’re done, etc.

But this only goes so far. The financial industry is literally designed to bury us in information and noisy feedback.

This is where the final patch comes in.

Patch #4

No study here, it’s a personal (and my favorite) trick backed by my experience and nothing else. But it works incredibly well (at least on me).

The 3 previous solutions share one goal: reduce feedback frequency and its impact on our decisions.

Here, the idea is to flip the feedback’s signal: negative becomes positive.

When I’m building a position, I always leave 5 to 15% of my target final position as a limit order at an absurdly low price.

So every time the price drops, the message shifts: from “I’m losing more money” to “I might get to buy this company at an absurd price.”

Of course, you have to have done the work and have confidence in your thesis. Nothing’s free in investing.

But combine this trick with the previous three, and the drawdown is no longer a danger signal, but rather a boring opportunity you saw coming, or noise until something ticks a box on your thesis-breaker list.

Here's the full list of tricks to put in place:

• Pre-mortem: Assume you’ll sell with certainty, and sort as many facts as you can into two piles: noise vs broken thesis.

• Implementation intention: When building those piles, write each intention explicitly as if + then.

• Information design: Minimize feedback exposure to defuse the myopic loss aversion effect.

• Opportunistic limit order: Leave a few limit orders sitting at extremely low prices to flip the drawdown signal from loss to opportunity.

And here's how the system works:

There’s bound to be someone in your contacts who needs this. Be the friend who shares it.

Share

Et voilà, the science of not giving a f*ck about drawdowns.

Some would say the ultimate patch against bad decisions during a drawdown is experience. I think that gets it backwards.

Experience is the accumulation of data. Wisdom is the ability to extract useful rules from it.

Of course, don't be dumb and learn from your mistakes. But more than that, be smart and learn from others’.

Our good old friend Science is just the most efficient and standardized method to do exactly that.

To keep learning and improving as an investor through science:

Subscribe now

One quick last thing: thanks a lot for your feedback on my first two investing theses published in this newsletter. All your messages and comments made the work worth it.

You can find them here.

More coming very soon.

Right back to it.

Take care,
Flo

Leave a comment

Share

Updated August 9, 2026: market data as of August 7, balance sheet figures as of March 31, 2026. Originally published May 16, 2026.

I found a way to get exposure to the Atlas Salt opportunity with a better risk/reward ratio. Of course, I put my money where my mouth is on this thesis; it’s now about 10% of my portfolio (ex-Atlas Salt itself).

It’s just basic math drowned in obscurity. Almost nobody who knows Atlas Salt even knows this other company exists.

Of course, Great Atlantic Salt still has to be built.

It came to me just as I was falling asleep. The idea felt like a eureka moment. But it quickly felt so obvious that it became suspicious.

I spent half the night checking whether I was missing something. And no, I was not. I had found a better way to play the same setup.

This is the joy of microcaps. Alpha is common. It is just extremely well hidden.

This time, it will not take 40,000 words. Just 8,000. The setup rests on Atlas Salt, with one structural subtlety that changes everything.

The alpha here is Vulcan Minerals, listed as VUL on the TSX Venture Exchange and VULMF on the OTC market in the United States.

Yes, this is the company that owns a 3% net production royalty on Great Atlantic Salt.

The setup is simple.

Vulcan is a C$57M market cap company. The present value of its royalty is about C$80M under the UFS (Updated Feasibility Study) assumptions.

Not enough?

Vulcan has no debt and about C$4.09M of cash.

Still not enough?

Vulcan owns about 24% of Atlas Salt, worth C$46M at today’s market price.

The equation is almost offensive:

Vulcan’s C$57M market cap = C$46M Atlas stake + C$4.09M cash + a royalty worth about C$80M under UFS assumptions, and likely more if my Atlas thesis is right.

In situations like this, there is always noise. And in microcaps, noise has a habit of becoming the economics.

Subscribe now

1. What Vulcan is

Since its first listing in 2001, Vulcan Minerals has never generated positive annual free cash flow, and it has reported positive net income only once. In 2021, when Atlas went public, Vulcan lost accounting control of Atlas and recognized a large non-cash gain on deconsolidation.

The accounting looks messy, but that’s baked into the business model.

If you read the annual financial statements, Vulcan Minerals is an exploration company “engaged in mineral exploration on properties in Newfoundland and Labrador.”

Its model is fairly simple: acquire and explore mineral ground, then try to move the asset forward through a partner, an option, a sale, or, in Atlas’s case, a dedicated vehicle in which it kept both equity and a royalty.

It is a prospect generator with a habit of leaving itself a back-end claim.

Atlas Salt is the only real success of that model. And to be fair, it is a good one. This type of serial explorer only needs one real success to compensate for decades of value destruction.

You can see it on the first real slide of Vulcan’s investor presentation. The company mentions its own name, then immediately Atlas Salt.

Atlas Salt is trying to build the first new salt mine in North America in a generation. Vulcan is the company that explored it, kept a large equity stake, and retained a 3% royalty on it. Simple, but not easy.

2. The Rest of the Box

Before getting to the two assets that actually matter, it is worth emptying the rest of the box.

As of March 31, 2026, Vulcan had C$41.51M of total assets. The Atlas Salt investment accounted for C$34.02M of that. Remove it, and the remaining balance sheet is C$7.49M:

• C$4.09M of cash,

• C$52k of receivables,

• C$15k of prepaid expenses,

• C$420k of investments,

• C$2.9M of exploration and evaluation assets, and

• C$20k of equipment.

Against that, Vulcan had C$107k of current liabilities and a C$2.47M deferred tax liability.

I keep it simple: cash gets full credit; non-cash current liabilities net of non-cash current assets are deducted; and every non-cash asset outside Atlas is assigned 0 value.

Some of these assets may be worth something. I simply do not want to stack assumptions for a few million dollars of possible Net Asset Value (NAV).

The Easy Part

Start with liabilities.

At March 31, 2026, Vulcan had no long-term debt and no off-balance-sheet arrangements. The only live operating liability was C$107k of accounts payable and accrued liabilities. Against that, Vulcan had C$52k of receivables and C$15k of prepaid expenses, leaving C$40k of negative non-cash working capital (the only advantage of having no revenue).

The only large liability is the C$2.47M deferred tax liability. That belongs in the NAV and realization math. The tax note points mainly to the accounting/tax gap on the Atlas associate investment, partly offset by tax attributes such as losses and exploration pools. The cash drag should mostly appear when Vulcan realizes value, especially through a sale of Atlas shares or other taxable income. Until then, it is a tax shadow over value realization.

One old overhang also disappeared. The Geophysical Service Incorporated claim, a legacy lawsuit started in 2014 over geophysical data, was discontinued in 2025. Vulcan had always said the case had no basis or merit. Whether or not the market ever assigned value to that risk, it is now gone. So the liability stack is basically payables + deferred tax.

Then there is cash.

Vulcan had C$4.09M of cash at March 31, 2026. That is the real asset outside Atlas. It reduces the probability that the parent company has to finance itself at a bad price while Atlas is still moving through financing and development.

Having no revenue creates a lot of problems, especially when it comes to funding the company itself (surprise, surprise), and Vulcan remains an exploration company with public-company costs. The C$4.09M of cash will be used to pay for all of that.

That is the clean part of the box. Cash is cash, payables are payables, and deferred tax is a future claim on realized value. You do not need much imagination to mark any of them.

The rest requires some.

The Harder Part

From here on, I assume everything is worth 0 (I have no imagination, or I'm just very conservative, your choice).

Investments

The investment line is small enough to ignore, but messy enough to mention.

At March 31, 2026, Vulcan carried C$420k of investments. Only C$28k was disclosed as public-company securities. The rest, C$392k, was Triple Point Resources, a private company that also sits in the Atlas family tree.

The Triple Point position is a private stake in Fischell's Brook, a salt dome that could be used for underground storage one day. That may become interesting if the storage market becomes financeable. But “may become interesting” is not a valuation method. Vulcan no longer accounts for Triple Point as an associate and concluded that it no longer had significant influence as of November 30, 2024.

The small securities bucket is not going to move the valuation either, but the Gander/Sassy arrangement shows how Vulcan tries to monetize ground outside Atlas: option the asset, take some cash or shares, keep a residual royalty where possible, and let someone else fund the work.

Vulcan optioned its Gander Belt Gold property to Sassy, which later assigned the option to Gander Gold. In 2025, the option payment was amended: instead of making part of the original cash-and-share payment, Gander Gold was supposed to issue Vulcan 1M shares by November 11, 2025. The August 2025 release said that, after receiving those shares, Vulcan would own 1.17M Gander Gold shares, plus 0.8M Sassy shares. As of today, Vulcan says the 2025 Gander Gold shares still have not been received, and discloses only 0.2M Sassy shares and 0.18M Gander Gold shares received under the option, plus another 0.6M Sassy shares from a separate purchase-and-sale agreement.1

This is exactly why I do not want to take any of these assets into account for the valuation. In microcap exploration partnerships, there are hidden “ifs” in every corner, ready to destroy value at the slightest opportunity.

The back-end economics matter, too. After Sassy earns 100% of the Gander property, Vulcan can retain a 3% NSR (Net Smelter Return), and Sassy has the right to buy back half of it for C$2M in cash and 0.5M shares within one year after delivering a feasibility report on any deposit advanced on the property.

Then there is the exploration portfolio.

Exploration Portfolio

The balance sheet carries C$2.90M of exploration and evaluation assets. It is just accounting value. Vulcan had 24 active mineral licenses in good standing at March 31, 2026, covering 973 claims, with 4 licenses and 165 claims optioned to Sassy / Gander Gold.

During 2025, Vulcan added C$255k to exploration and evaluation assets, received a C$137k Junior Exploration Assistance grant that was deducted from the relevant project’s carrying value, and wrote down C$40k of costs where management decided no further exploration would be undertaken.

There is real geological smoke in the portfolio (but I still give it 0 value):

• Colchester/Springdale is the only non-Atlas exploration asset worth more than a footnote. It has old copper workings, a historical copper estimate that is not NI 43-101 compliant, and recent Vulcan drilling that returned real copper-gold intercepts at McNeilly.2 That is enough to keep it on the radar, and probably makes it the harshest 0 in the residual portfolio.

• Red Cross Lake is earlier-stage. It is a nickel-copper-cobalt-gold target in central Newfoundland, built around an intrusive body that could, in theory, host sulphide mineralization.3 In 2025, the work was still preparatory: checking the ground, improving access, and cutting trails for a future drill program. In plain English: there are only targets.

• South Voisey’s Bay is simpler from a cash-budget perspective. It is a nickel-copper-cobalt project in Labrador, across three licenses and 30 claims. No work was carried out in 2025. For now, it is inventory.

• Carbonear is the project that eats cash. It is a sediment-hosted zinc-lead target with additional gold and metallic geochemical anomalies. Vulcan completed a soil program in 2023, a gravity survey over the Gull Pond area in 2024, and field work in 2025. In early 2026, the company completed one hole of a planned 6-hole drill program before deferring the rest because of weather-related access issues. Drilling resumed in June and the remaining five holes were completed in July, 1,087m in total (visual mineralization in every hole, assays pending). That makes Carbonear relevant for cash burn, even if it remains irrelevant to NAV for now.

These assets come with required work obligations, but the amounts remain modest. The latest MD&A estimates C$490k of required expenditures over the next nine years to maintain the current licenses, including C$152k in 2026 and C$133k in 2027. That is manageable against C$4.09M of cash.

Then comes the spending that management actually controls. Here, the numbers start to become more meaningful. The entire expense side of the thesis will be built from small cost lines like this, which eventually translate into the microcap investor’s best friend: dilution.

So this is the rest of the box. It is not worthless, but it is easier to treat it as if it were.

Beyond that, I am happy to be surprised.

Until then, the cost of carrying that optionality is the price of admission.

3. The Cost of the Wrapper

The residual assets can be valued at 0 in the NAV, but they still sit inside a public exploration company.

Vulcan pays people, keeps claims alive, attends conferences, files financial statements, pays auditors and lawyers, carries an office, markets the story (or tries to), and grants options.

Once every non-cash asset outside Atlas is marked at 0, the analysis becomes simpler: Vulcan is a cash bridge.

Two questions matter before getting to the Atlas stake and the royalty.

• First, how much cash does Vulcan need before the Atlas bet can start paying?

• Second, how does it fund that gap, and at what cost to shareholders?

From there, we can move directly to its Atlas exposure, and then to Vulcan’s valuation.

The Cash Cost of Survival

With basically no revenue, the P&L is a bad place to stop, but a useful one to start.

Vulcan reported a C$2.57M net loss in 2025, but that number included a C$1.29M dilution loss on Atlas and a C$1.00M equity-accounted loss from Atlas. This is the accounting cost of Vulcan’s Atlas exposure outside the royalty.

It is useful, but it does not answer the question I care about here: how much money leaves Vulcan before the royalty starts paying?

Here's the cash spend by the management since 2023.

Excluding Atlas share purchases, Vulcan’s actual cash cost of staying alive has been about C$0.8M to C$1.0M per year.

The 2025 improvement came mostly from lower exploration spending, but it was partly offset by a higher cost base. G&A and directors’ fees rose from C$601k in 2024 to C$791k in 2025, before C$180k of expensed share-based compensation. Salaries and cost recoveries, professional fees, travel, marketing, and directors’ fees all moved higher.

Then 2026 came. In the first quarter alone, Vulcan's cash balance fell C$416k. Strip out the C$43k of option proceeds and the cash cost was C$459k, already net of a C$48k grant. Exploration took C$252k of it. Exploration costs have historically been very lumpy within a given fiscal year, and a single quarter will typically carry the bulk of the year's drilling costs. Not this time. That C$252k was one hole. Vulcan drilled the first one at Carbonear over the winter, stopped when the access road closed, and went back in June for the remaining five: 1,087m, finished July 21. Those bills land in Q2 and Q3. At the C$295/m4 McNeilly ran in December 2024, Carbonear’s remaining 1,087m put 2026 at C$1.2M to C$1.5M with overhead, half again to double 2025.

For a C$57M market cap company without revenue, this is not small. This brings us to the main problem with this setup: this is not a shareholder-friendly wrapper.

Beyond his salary and share-based compensation, Patrick Laracy, Vulcan’s President and CEO, is paid by the company in two ways:

• First, his services are provided through Triassic Properties Ltd., a corporation he controls, under a management consulting agreement at C$165k per year. The older employment agreement also contains a C$330k termination or change-of-control payment.

• Second, Vulcan pays rent to a corporation controlled by the CEO. That rent rose from C$36k in 2024 to C$48k in 2025.

There is no hidden accusation here. The arrangements are disclosed, and they may be perfectly normal for a microcap explorer. The issue is more basic: the President and CEO of the company whose shares I may own for years, perhaps decades, has both the incentives and the ability to capture some shareholder value along the way. Some through cash, some through dilution, as we will see just below. Either way, it belongs in the valuation.

Two small cash offsets soften the burn.

The first is interest income. Vulcan earned C$275k of interest in 2024 and C$127k in 2025. That is not trivial, but it is self-liquidating. As the cash is spent, used to defend the Atlas stake, or reinvested in exploration, the subsidy fades. Rates can change the picture, of course, in both directions.

The second is the small non-Atlas income bucket. Vulcan recorded C$64.8k of option-payment income in 2024, 0 in 2025, and C$3.2k of royalty income in 2025. There are also potential back-end royalties on assets like Gander, including a possible 3% NSR if the option conditions are eventually met. For now, this category is trivial.

Vulcan needs between C$0.8M-C$1.6M per year to stay alive. With no operating business, every funding source ultimately comes from the same place: shareholder value.

The Shareholder Cost of Survival

Vulcan’s share count tells the usual junior-explorer story, at least until 2021.

• Sep. 30, 2019: 61.4M basic shares outstanding.

• July 2021: 105.6M.
  The dilution mostly came from the way Vulcan had to stay funded: repeated low-priced treasury issuances, then the 2021 Sprott financing, which added 10.0M shares at C$0.20 for C$2.0M (with C$0.30 warrants attached).

• Oct. 31, 2025: 129.3M.

• April 2026: 129.9M.

Yes, Vulcan diluted shareholders. At a C$57M market cap, that is usually the business model.

As of the April 2026 MD&A, Vulcan had 129.9M voting shares outstanding, 7.45M options outstanding, 6.175M options already vested and exercisable, and no warrants. The current options represent about 5.7% of the basic share count. Full exercise would bring in roughly C$1.6M before costs, enough to fund maybe 1-2 years of survival under the current burn profile.

That cash arrives in the winning case, but so does the economic cost.

Vulcan’s option plan is a rolling 10% plan: at any grant date, the option reserve cannot exceed 10% of the issued and outstanding share count.5 It allows terms of up to ten years, and management has used that full duration before. Director options issued in January 2022 at C$0.30 expire in January 2032. The April 2026 grant follows the same logic: 2.55M options at C$0.23, again with a 10-year term. That ties up “scarce” option capacity for longer. Management still chose the 10-year claim: fewer future grants available, but today’s strike locked across Atlas financing, construction, ramp-up, and the first royalty window.6

In plain English, management seems to expect Atlas to reprice Vulcan, and chose to lock in part of its option capacity now, before that repricing happens.

If management is right about Atlas, and I think they are, those options could become very valuable. Good for them. But what about shareholders?

Options can be fine when they buy competence. A charitable reading is that Vulcan’s option grants can be seen as payment for past exploration work, the original Atlas spin-out, and the fact that management did leave Vulcan with both a large Atlas stake and a royalty. That work created value.

The less charitable version is uglier. Management is reserving a larger claim on the future Vulcan value before minority shareholders get paid. This is another friction between shareholders and insiders.

The last problem is when Vulcan tries to do more than survive.

The Shareholder Cost of Doing More Than Surviving

At the start of 2021, Vulcan owned 31.891M Red Moon shares, or 63.03% of what later became Atlas. It then sold 3.0M Red Moon shares at C$0.12 for C$360k, reducing its stake to 57.1% (it needed cash to bridge the gap until a likely dilutive financing). By the end of 2021, after Atlas equity financings, Vulcan still owned 28.891M Atlas shares, but only 36.80% of the company. As of today, it is 23.6%.

In 2025, Vulcan spent C$331.6k buying Atlas shares, including a C$300k participation in Atlas’s LIFE financing. Atlas issued 10.9M shares at C$0.80. Vulcan bought 375k of them.

Here, management learned a lesson. The 3.0M shares sold in early 2021 would have been worth 10x as much six months later, 37x as much two and a half years later, and still around 13x as much today. All because they failed to optimize Vulcan’s cash position and had to raise liquidity before the financing that closed a few months later. I imagine management had some long nights with that one.

The lesson is clear. Atlas Salt is the crown jewel. Even more so now that site work has started. They cannot sell it unless valuations become absurd.

That is one reason the 2025 purchase matters in my opinion. Vulcan put approximately the same amount of cash back into Atlas shares as it had raised by selling them in 2021, only at a much higher share price and therefore for 1/8 as many shares. Even at that price, management apparently thought the shares were worth owning.

But the repurchase also shows something else. Even when Vulcan spends cash, Atlas issues shares faster than Vulcan can follow. And that becomes more true as Atlas moves from study-stage capital to construction-stage capital.

Atlas shares are the crown jewel, but can Vulcan defend its stake?

At 23.6%, a C$100M Atlas equity raise would require Vulcan to invest about C$24M to hold its percentage. A C$200M raise would require C$48M. Vulcan has C$4.09M of cash, and that cash already has a job: keeping the wrapper alive until Atlas can pay the royalty. With the balance sheet it has today, Vulcan cannot defend 23.6% through a serious Atlas construction financing.

Vulcan could raise money.

Of course, but a large Vulcan equity raise to defend the Atlas stake would dilute every Vulcan shareholder’s claim on the royalty. The only way that makes sense is if Atlas is clearly mispriced after accounting for the dilution of Vulcan’s royalty exposure per share, and if the Vulcan raise itself is not too punitive. That is a high and very uncertain bar.

My take is that, if Vulcan wants to create more per-share value, it should let its Atlas percentage fall. That is literally “sit on your ass” investing. Carry the royalty, spend little outside Atlas, and avoid doing something clever with cheap paper.

I do not think the 2025 Atlas purchase contradicts that. C$300k in a C$8.7M Atlas financing is not an attempt to defend control. It looks more like a participation designed to signal continued support for the project at an attractive price. Vulcan’s decision not to participate in Atlas's recent financings is an additional clear sign of that.

So I model Vulcan with a fixed Atlas share position. The underlying assumption is simple: management remembers the lesson, stays relatively disciplined, does not overfund non-Atlas exploration (Q1-26 isn't cut from that cloth), does not issue cheap Vulcan equity to protect a headline stake, and lets the royalty do most of the work.

That may be too charitable, but it is the message management has been sending since 2021.

To assess the likelihood of this outcome, the right place to look is the incentive structure.

Subscribe now

4. Who Controls the Wrapper

Let’s summarize the situation simply.

The good version of Vulcan requires doing nothing. Junior explorers exist to explore. Management teams are human. Humans find it hard to do nothing. Are insiders really paid to wait?

This is mostly a Patrick Laracy question. He beneficially owns or controls 11.69M Vulcan shares, including 515,984 through Triassic Properties, against 129.9M shares outstanding. Call it 9% of the company. At a C$57M market cap, that is approximately C$5.1M of stock. A C$0.10 move in Vulcan is about C$1.1M for him.

The check matters, as discussed above, but the stock matters more. This is the first useful fact.

The second one is less clean: no person or company controls 10% or more of the votes. Board ownership is limited: Fraser Edison disclosed 1.2M shares, Carson Noel 275k, and Peter Mercer nil. In a scenario where management simply waits, this is not really a problem. But it also means that there is no obvious outside shareholder with enough voting power to force the company into the boring version of the thesis.

In practice, Vulcan is controlled less by formal ownership than by habit, history, and management inertia. That is often how it works in microcaps led by the same CEO for thirty years. That is why the company is still behaving like a prospect generator.

For the options, fine. Management is trying to maximize its own economic interest. But the rest does not really make much economic sense.

Vulcan diluted heavily in the 2019-2021 window. The share count went from 61.4M at September 30, 2019 to 105.6M by July 2021. The denominator was reset before Atlas became the asset that matters. Since then, I do not see the same pattern of repeated outside equity issuance at the Vulcan level. The company has mostly lived off the cash raised in that earlier period, helped by interest income, small option payments, government grants, and occasional warrant or option exercises.

Most of the dilution between 2021 and 2025 came from the exercise of warrants issued in 2021 or earlier. The remaining dilution during that period was much smaller and came from SBC.

There are currently no warrants outstanding; the last remaining warrants were exercised in 2025.

Since 2021, management has shown little appetite for keeping the wrapper alive through cheap external equity. My read is that they are waiting for Great Atlantic to be built with at least as much impatience as I am, and that they will try to preserve the parent without issuing too much stock at the wrong price.

They will use whatever cash sources are available. Option exercises are part of that bridge, with one specific point worth spelling out. Vulcan pays salaries and fees to management and directors. Management and directors receive options. When those options are exercised, cash goes back into the company. For Vulcan, it is a form of cash recycling. For management, it increases ownership at a “low” cost.

Yes, it is value leakage, but I can live with that. Option dilution has been modest, and the current cash balance, plus whatever future inflows the company can generate, should help limit external financings at stupid prices until Great Atlantic produces its first cash flow.

The cleanest outcome would be an activist with a few tens of millions.

The job would be simple enough: use the fragmented shareholder base to build a position slowly (for months if necessary), force the corporate structure to match the math, and turn Vulcan into the holdco it already wants to be on paper.

Cut the unnecessary spend: consulting fees, expensive office space, exploration budgets that no longer deserve priority. Sell or option the exploration portfolio. Keep the Atlas shares. Keep the royalty. Then sit there for decades, collecting royalty cash and Vulcan’s look-through share of Atlas cash flows once Great Atlantic is in production. It does not even need an office. A registered address, basic corporate services, and the occasional rented meeting room would probably be enough.

(If only I had the few tens of millions lying around and the desire to pick a fight…)

For now, shareholders are stuck on a boat that has been sailing in the same direction for five years:

• slow dilution helping fund expenses that are mostly unnecessary;

• a CEO who owns enough stock for the share price to matter, and who has created real value before (Atlas did not appear by accident, and neither did the royalty);

• a very low probability that the situation changes.

This is obviously not the most minority-shareholder-friendly setup. But it could be much worse.

In short, the asset creation deserves credit, but the wrapper deserves a discount.

Before applying that discount, let’s look at the assets in detail.

5. The Atlas Stake

Atlas owns a feasibility-stage salt project with a 2025 Updated Feasibility Study showing a C$920M after-tax NPV at an 8% discount rate, based on a 4.0 Mtpa operation and a 24.25-year initial mine life.

At C$1.58 per Atlas share, Vulcan’s 29.4M-share stake is worth about C$46.4M. But that is the market price of the exposure; it is not my estimate of value.

If you remember the Atlas deep dive, you know where my shareholder view diverges from the UFS. It left out, or did not fully capture, several costs and risks that matter to equity holders: financing, working capital, residual corporate overhead, dilution, and the friction of getting from feasibility study to operating mine.

If I use the assumptions from my Atlas DCF, which I am reproducing here, Vulcan’s look-through PV of the Atlas stake would be:

• Bull case: C$311.5M.

• Base case: C$96.7M.

• Bear case: C$17.6M.

These are September 2025 present values, discounted at 8%. Their value today is slightly higher, all else equal, because time has passed and the discount period has shortened.

These numbers already include the dilution I expect Atlas to take in order to finance Great Atlantic. In my Atlas model, the dilution scenarios were not gentle. On a fully diluted basis, the bull case assumed approximately 50% additional dilution. The base case assumed more than a 2x in the share count. The bear case assumed more than a 4x increase.

If Atlas works, Vulcan still owns meaningful upside through the equity stake, even after a harsh financing path. But it also owns the royalty, and the royalty is a very different animal: it is indifferent to many of the things that can hurt Atlas shareholders along the way.

6. Royalty

This is the reason the setup exists.

Vulcan owns a 3% net production royalty on the Great Atlantic Salt mineral licenses.

A small disclaimer before the fun part: I have not found the full original royalty agreement. Vulcan and Atlas disclose the royalty repeatedly, and the UFS “relied on royalty and title information disclosed by Atlas in its public filings and has not independently verified the underlying royalty agreements or title instruments.”

The 2025 UFS models Vulcan’s royalty as 3% of gross revenue after deducting port charges, processing costs, and the Newfoundland and Labrador Mining Tax.7

On paper, that is clear enough, but in practice, it is always messier.

The Uncertainties

Take the Scotwood MOU.

Atlas has a non-binding MOU for a 50/50 joint venture with Scotwood. The structure, as disclosed, would have Atlas sell salt to the JV, while the JV would use Scotwood’s Canadian retail distribution network and split profits between the two partners.

Does Vulcan’s royalty stop at the price at which Atlas sells salt into the JV, or does it also capture Atlas’s share of the downstream JV profit? I do not know. The royalty documents I have are not precise enough to answer that. My strong guess is no, but I have no proof.8

If the answer is yes, great. Vulcan would also capture part of the product-mix upside generated through the Scotwood JV. If the answer is no, the royalty would stop earlier in the chain and would not fully participate in some additional margin Atlas may generate downstream.

There is also a third possibility: the original royalty agreement may already deal with this type of structure and use a different method to calculate the royalty when salt is sold to an affiliate, joint venture, or downstream vehicle. But again, that is a maybe.

In any case, the impact should still be limited. Vulcan should still benefit from higher realized pricing on the volumes Atlas sells directly, locally, or on the spot market, to the extent those sales sit inside the royalty base. And if some of the downstream margin is captured outside the royalty base, Vulcan still has 29M Atlas shares to participate in it.

What is clear, and probably most important, is that the royalty remains exposed to the salt price itself. And that is still the largest source of upside versus the UFS.

As discussed in the Atlas deep dive, the UFS assumes 2% annual price growth after construction, while deicing salt prices have compounded at approximately 4% over the last four decades. It also leaves out the catalysts currently pushing pricing above its historical path: more volatile winters as climate patterns become less predictable, tighter local supply, higher oil prices reducing the economic trucking radius, etc.

This makes the royalty a valuable asset.

• It gives Vulcan exposure to many of the same upside drivers as Atlas: higher salt prices, better realized pricing, faster production, mine-life extensions, and the conversion of resources not included in the initial reserve case.

• It is also less exposed to some of the frictions that sit in Atlas equity: future equity dilution, expensive project financing, construction budget overruns, and so on, at Atlas.

But how valuable?

How Much Is the Royalty Worth?

Under the UFS assumptions, this royalty would transfer about C$266M from Atlas to Vulcan over the life of mine, before Vulcan-level tax. Discounted at 8%, I get a present value of about C$80M.9

But I just spent 40,000 words challenging the UFS assumptions and adding the risks the study could not, or did not fully, capture.

The advantage here is simplicity.

The UFS gives the annual revenue stream. It also gives the deductions that matter for the formula: processing and material handling, which include port costs, and the NL Mining Tax. Great, I do not need to rebuild the full mine equity case to value Vulcan’s royalty.

The formula is: 3% of gross revenue, less port charges, processing costs, and the NL Mining Tax.

Using the UFS numbers, the “royaltyable” base is approximately 82.7% of gross revenue.10

That becomes my base factor. I keep it constant across scenarios.

From there, only three assumptions really matter for Vulcan: delay, salt price, and the part of the sales mix that actually flows through the royalty.

For the delay and the salt price, I simply reuse my Atlas assumptions. For the mix, I had to adjust.

In the Atlas DCF, I gave credit for a modest improvement in realized price because the UFS compresses geography, channel, contract structure, and product format into a single C$81.67/t FOB Turf Point price.11 That was already a restrained way to model a much larger possible upside.

For Vulcan, I apply a 50% haircut to that upside.

If Atlas sells salt at a better mine-level / FOB-equivalent price, the royalty should benefit. If the extra economics are captured downstream in a JV, I should not give the royalty that profit. So for the base and bull cases, I take the incremental price/mix uplift from my Atlas DCF and cut it in half for the royalty (there is none in the bear case to begin with). It is a blunt haircut, but one that keeps me from building a fake channel model from documents I do not have.

Given the impact of mix on revenues, this is probably a pretty conservative choice.

Under those assumptions, the royalty has the following present values:

• Bull case: C$126.9M.

• Base case: C$97.1M.

• Bear case: C$83.9M.

Great: at least 47% above the market cap. Long live microcaps.

True, in a tax-free world. In the real one, Newfoundland and Labrador and the federal government both take their cut before Vulcan shareholders see the value.12

Tax

In Newfoundland and Labrador, there is a Mineral Rights Tax equal to 20% of royalty receipts above C$200k, after certain deductions.

After that, there is Canadian corporate income tax on profits, at 30% all-in.

So, in the rough version, each C$1.00 of royalty revenue becomes:
C$1.00 x 0.80 x 0.70 = C$0.56 of after-tax profit.

From there, two things complicate the calculation:

• the deductibility of expenses;

• the tax shields Vulcan will have accumulated by the time the royalty starts generating revenue.

Not all expenses have the same tax treatment, and I cannot know what Vulcan’s future tax shields will be when the first royalty dollar arrives. So I directly fold the uncertainty into the scenario tax rates:

• Bull case: 40% tax.

• Base case: 42% tax.

• Bear case: 44% tax.
  No expense deductions, no tax shields, and taxation applied to every dollar of royalty revenue.

In all cases, the tax rate has a limited impact on the final NAV, so the cost of being wrong here is not very high.

After applying these adjustments, we get the following royalty present values by scenario:

• Bull case: C$76.1M.

• Base case: C$56.3M.

• Bear case: C$47.0M.

Share

That is less impressive than the pre-tax figure, but it is still roughly equal to or above Vulcan’s current market cap in the bull and base case. The bear case is 18% below Vulcan’s market cap but that also means you’re paying for Vulcan’s C$46.4M stake in Atlas for only C$10.2M (or the other way around, whichever you prefer). And that’s without counting all of Atlas’s own upside. Not bad, in my humble opinion.

Note that the range is much tighter than for Atlas equity, which is exactly the point. The royalty sits much closer to the revenue line. It does not fund the mine, issue shares, or require Atlas to capture every dollar of downstream margin. It only needs tonnes sold and a mine-level revenue base.

That is the power of the royalty, and the core strength of Vulcan. Without it, there would be little reason not to own Atlas directly instead.

We now have the important pieces to value Vulcan Minerals.

7. Valuation

We have already done most of the work.

• Royalty PV → Check.

• Atlas stake PV → Check.

• Asset/liability value → Check.
  I assume everything outside Atlas is worth 0, except for cash, the deferred tax liability, and the small negative non-cash working capital position.

That leaves two variables to model:

• Wrapper costs

• Dilution

I lay out the full assumptions in the DCF.

Download The Valuation Model

Here, I will only show the ones that matter for understanding the valuation.

Wrapper Costs

For the annual wrapper cost, I use:

• Bull case: C$0.8M, which assumes an optimistic continuation of the 2025 expense level.

• Base case: C$1.2M, which reflects the last two years plus 2026 annualized from what the first quarter already spent.

• Bear case: C$1.6M, which assumes management keeps drilling every year the way it drilled Carbonear.

I escalate each scenario at 2% per year over the full mine life, although the impact of this escalation is minimal.

I treat these as net wrapper costs. In other words, I already include potential cash inflows such as interest income on cash, non-Atlas royalties, option exercises, grants, and other small receipts inside that number.13

It is conservative, but mostly, it is simpler.

Dilution

I split dilution into two buckets.

Survival Funding

In the bull case, Vulcan’s cash and small sources of flexibility carry the company to the first royalty cash flow. In the base and bear cases, the wait is longer, wrapper costs keep compounding, and the cash balance is eventually exhausted. So I force one equity raise to fund the gap:

• The base case runs about C$1.3M short and issues 3.7M shares at C$0.35.

• The bear case runs C$5.4M short and issues 21.6M shares at C$0.25.

Once the royalty starts paying, I stop modeling survival dilution. From that point, the royalty pays for the wrapper.

Management Dilution

Vulcan’s historical share-count growth is not a clean proxy for that. It needs to be decomposed.

Most of the dilution since July 2021 came from old low-priced warrants. Those warrants came from past financing dilution, at a time when Atlas was not yet the project that matters and management seemed less careful with dilution, while mostly living off the public-company G&A machine. The last financing of that kind took place in 2021. Since then, Vulcan has been more careful with its cash, precisely to avoid further dilution, and today it actually has more cash than it had in 2021. The last warrants were exercised or expired in 2025 and will no longer contribute to dilution.

I therefore base the dilution assumption on SBC alone, on top of the financing dilution already modeled in the base and bear cases.

The bull case is anchored in realized option dilution. Since July 2021, Vulcan has issued about 1.225M shares from option exercises, against a starting share count of 105.6M. That is only 1.16% total dilution, or about 0.24% annualized, so I round it to 0.25% per year.

The base case is anchored in today’s option overhang. Vulcan has 7.45M options outstanding against 129.9M shares, equal to 5.74% of the current share count. Spread over ten years, that is about 0.57% per year, so 0.50% per year is a reasonable normalized cost of the existing option book.

The bear case is the option machine that keeps moving. Vulcan granted 3.75M options in February 2025 and another 2.55M in April 2026, or 6.3M options in fourteen months, equal to 4.85% of the current share count. The April 2026 options have a ten-year life, which reduces the risk of an immediate recycling loop. Fine. But bear cases are not built around management restraint.

If the stock works, the options become valuable and exercisable. If they are exercised, Vulcan gets cash and insiders get more shares. If the board refills the pool, the leakage repeats.

That is why I use 1.00% per year in the bear case. A long-dated option loop that helps fund the wrapper while moving more ownership to insiders along the way. The 2025-2026 grant pace is too lumpy to annualize mechanically, but it is enough to justify charging the structure for recurring leakage in the bear case.

To sum up, over the 28-year model period, I use:

• Bull case: 0.25%/year (7.2% dilution over 28 years).

• Base case: 0.50%/year (15% dilution over 28 years, excluding financing dilution).

• Bear case: 1.00%/year (32% dilution over 28 years, excluding financing dilution).

In all three cases, the 7.45M options currently outstanding are added on top of everything else.

Final Valuation

At a very high level, I value Vulcan as follows:

(PV of post-tax royalty + PV of Atlas stake - PV of wrapper costs + cash - deferred tax liabilities + non-cash working capital) / diluted share count

Here is the final table:

At a C$0.44 share price, the bear case implies 51% downside, the base case is almost a 2x, and the bull case is around 5.9x.

4 important reminders:

• These are not future price targets. These are present values: the value of the business today under my assumptions and certain UFS assumptions, after discounting the future value of the assets and liabilities. All else equal, those values should mathematically increase over time as the discount period shortens. That does not mean the stock price cannot trade above or below them.

• The probability asymmetry is also very pronounced. I estimate the probability of a bear case, or worse, at less than 10%. Obviously, this is an instinctive number. No serious calculation can realistically get you there. But I laid out all my arguments in the conclusion of the Atlas Salt valuation section.

• I haven't updated my Atlas Salt model. That said, there has been a lot of news since then (Export Development Canada letters, a few favorable financings, a new CFO, etc.). None of them individually moves the needle, but cumulatively, I think they tilt the value and the probabilities slightly in our favor. I'll update the model once we have more concrete information.

• At no point do I include the 288 Mt of additional Indicated Resources or the 868 Mt of Inferred Resources. All scenarios are based only on the reserves currently bankable under the mine plan. Over time, in a highly homogeneous salt deposit like this one, additional resources are likely to be converted into reserves as production advances.

That is what I call asymmetry and optionality. This is why I have roughly 10% of my portfolio in it (a bit more than 20% in total with Atlas Salt).

8. Vulcan vs Atlas

The risk/reward ratio is very attractive, to say the least, for both Atlas and Vulcan, but it is obviously more attractive for Vulcan. That is just arithmetic.

Vulcan carries the same upside, minus the risks that come with funding a mine: construction budget overruns, and the massive dilution that pays for them. The royalty is indifferent to both.

For anyone wondering: yes, the assumptions are the same in both cases where they overlap (except for the sales-mix haircut, which is more conservative for Vulcan).

But this is a purely mathematical analysis. The shape of the value matters too, and so does the way the risks are transformed.

If we assume that the C$4.09M of cash will be used to carry Vulcan to 2030, then Vulcan can be viewed as an Atlas stake trading at a 23.3% premium.14 That stake is exposed to almost exactly the same risks as owning Atlas directly. So this premium has to buy something.

And it does:

• First, a royalty that is probably worth more than Vulcan’s current market cap, and that carries far less risk than Atlas equity, especially the most important risk in the whole setup: financing conditions.

• Second, the holdco layer: additional dilution, capital allocation and governance risks at Vulcan.

My view is simple: the extra risks are more than offset by the royalty and its quality. Vulcan remains the better way to play Atlas on the numbers.

In market reality, two other problems show up.

Liquidity. Vulcan is clearly less liquid than Atlas, and there is no market maker “supporting” the stock the way Atlas has one. If you want to invest a few hundred thousand dollars, you will need several days at least to build the position without giving up too much to slippage. I think this problem should shrink if Atlas succeeds and the Vulcan setup becomes more visible. But if something goes wrong, getting out of Vulcan will be harder than getting out of Atlas.

M&A. As I explained in the Atlas deep dive, Vulcan may have an incentive to protect its royalty in a takeout situation and use its influence over Atlas to maximize the value of its Atlas stake plus royalty, rather than only maximizing the cleanest outcome for Atlas shareholders. Owning Vulcan is also a form of protection against that scenario.

So, Vulcan or Atlas?

The best choice, in my opinion, and the one I chose, is to own both. They have different risk profiles, and I prefer to spread that risk by owning both. Vulcan is more convex on the numbers. Atlas is cleaner. Each investor has to decide which set of problems they would rather own.

For relative position sizing, I think the right approach is to compare one against the other.

Here is Atlas Salt’s market cap in green, and Vulcan’s in purple.

If you look closely at the 2022-2026 period, the two market caps do not always move together. Sometimes one asset moves 15-20% in a few days while the other barely reacts, even though the underlying Great Atlantic exposure has not changed. The Atlas/Vulcan chart shows this well (careful, the dilution distorts the numbers).

These inefficiencies are normal at this size. They can also create opportunities to build a position in one name rather than the other, or even to reallocate between them. I compare these curves regularly, and twice I have swapped a bit from one to the other when the pricing gap became too obvious (just be careful with fees, spreads, slippage, and… taxes).

The real asset is Great Atlantic Salt. Atlas and Vulcan are two different ways to own exposure to that asset.

If you are most worried that Atlas will be forced to dilute massively at a low price, Vulcan is probably the better choice. If you are convinced that Laracy will spend too much of the royalty cash flow on value-destructive exploration, Atlas is probably the better play.

If you do not have a strong view on which risk matters more, owning both is the best answer. That’s mine.

Share this widely enough and maybe one of you will show up with the tens of millions it would take.

Share

9. Conclusion

That is why Great Atlantic Salt is probably the best risk/reward I have seen in my career so far (I use “probably” out of epistemic honesty). They're both up roughly 40% since my deep dive in May, and my view hasn't changed. I didn't sell a single share.

That does not mean there is no risk. Far from it.

A serious accident during mine construction. A once-in-a-century earthquake like the 1929 Newfoundland event. An armed conflict. Pick your black swan.

These two deep dives were written to show how exceptional the potential upside is relative to the risk being taken. But that should never obscure the fact that the risks are real, and that capital can be permanently lost.

Personally, I have taken that bet, and I would take it again.

Take care,
Flo

At first glance, the company is easy to dismiss. It operates in a corner of the economy most investors do not even know exists, with no close publicly listed peer, and almost no natural public-market investor base.

The fact that the project exists at all is unusual: North America has not seen a comparable new entrant emerge in a generation. In this market, a company needs a rare combination of scale, cost advantage, logistics, customer relevance, and timing before it can even matter. This company has all of them.

Disclosure: I have skin in the game. I’ve been a shareholder for months, and I bought more during the writing of this investment thesis. That creates an obvious bias. The only honest way to handle it is to make the machinery visible: every assumption, every number, every calculation, and every place where the conclusion could be wrong. Do not take my word for it. Check the work. Then check it again.

The aim of this piece is to make the entire chain explicit: the upside, the downside, the assumptions, and the asymmetry between them. I spent over 500 hours on this thesis, condensed into 40,000 words, because the full answer emerges when you grasp the whole picture.

That level of work is deliberate. I spend 90 hours a week analyzing businesses, writing, reading filings, rebuilding models. This newsletter is where I publish the work, and I want it held to the same standard.

If this is the kind of work you want more of, join the list.

Subscribe now

The valuation section includes the full DCF, with every figure and assumption made explicit.

Table of Contents

01. An Overview of the Company

02. The Salt Market

• The $26 Billion Picture

• Not All Salt Is Created Equal

• One Molecule, Three Jobs

  • TINA

  • The Implication for an Upstream Producer

• How the Market Actually Works

  • The Buyer Side

  • The Seller Side

• Demand That Does Not Bend

• A Supply Base Frozen Since 2001

• Import Dependence, Weeks Away

• Pricing Without a Market

• Where the Margins Are

• Climate & Environment: The Objections That Do Not Stick

• Where This Leaves Us

03. Atlas Salt’s Mine: Great Atlantic Salt

• The Deposit

• Transport Logistics as a Moat

• The Operational Design

• The Social Component

  • Permitting

  • Jurisdiction and Community

  • Environmental Profile

04. The Business Model

• The Product: One Feedstock, Several Commercial Formats

• Where Atlas Plans to Sell

• How It Actually Sells: CIF, DAP, FOB

• The Contracts

  • Scotwood Industries: Commercial Offtake + Canadian JV

  • Sandvik Mining: Equipment + Vendor Financing

  • Hatch Ltd.: Lead Engineering & IPD Partner

  • Continental Conveyor: Material Handling + IPD Partner

05. Management and Alignment

• The Operators

  • Patrick Laracy: The Founder Who Stayed

  • Rowland Howe: The Man Who Knows the Competition

  • Nolan Peterson: The Executor

• The Construction Bench

• The Cap Table

• The Royalty

• What Laracy’s Alignment Means for Shareholders

• The Net Read

06. A Quick Look at the Financial Statements

• The Wrong Place to Look for Value

• Balance Sheet Snapshot

• The Cash Question

• The Next Step

07. Financing

• What Atlas Needs to Fund

• The Debt Architecture

• The Cost of Financing

• What to Watch

08. The NPV: Decoding the C$920M Headline

• The Revenue Side

  • The Price

  • The Volume Schedule

  • The Escalation

• The Cost Side

  • The Build

  • Sustaining Capital

  • Operating Cost

  • The Royalty and the Taxes

• The Sensitivity Analysis

• The Discounting Mechanics

  • Why 8%?

  • My Choice of Discount Rate

09. Risks

• The Financing

• The Build

• The Commercial Ramp

• The Salt Price

• The Permitting Tail

• The Port Agreement

10. Valuation

• The Ratio and Its Double Reading

• Qualitative Valuation

• Quantitative Valuation

  • Financing Costs

  • Working Capital and Corporate Overhead

  • Residual Risk Reserve

  • Execution

  • Salt Price

  • Sales Mix

  • Dilution

  • How Much Is Atlas Salt Worth?

11. Repricing

• Recognition

  • The Critical Gates

  • The Variables

  • The External Catalysts

  • Timing and Sequence

• Sit-on-your-Ass

• Someone Else Pays

  • Why M&A

  • When M&A

  • Management’s Take

  • Signals to Watch

12. My Final Take

13. A Better Way to Play the Setup

01. An Overview of the Company

The company is Atlas Salt Inc., listed under the ticker SALT on the TSX Venture Exchange and SALQF on the OTC market in the United States. Its market cap is around C$120 million (US$88M).

The project, Great Atlantic Salt, is a planned underground salt mine in Newfoundland and Labrador (NL), Canada, wholly owned by Atlas Salt.

If that sentence made you want to stop reading, that is exactly why I think it is so cheap.

The 2025 Updated Feasibility Study (UFS) gives Atlas a 24.25-year reserve-case mine plan:

• 95.0 Mt of Probable Reserves,

• C$272 million of average annual pre-tax cash flow,

• and a C$920 million after-tax NPV at an 8% discount rate.

That is only the reserve case. The 95 Mt sits inside a much larger resource envelope: 288 Mt of additional Indicated Resources + 868 Mt of Inferred Resources, to which the UFS assigns no value.

Atlas trades at a P/NPV of approximately 0.13x,1 about 1/2 to 1/3 of the level at which comparable feasibility-stage mine developers typically trade, even though many of them arguably carry more risk than Great Atlantic Salt.

In the last eighteen months, Atlas has rebuilt the team around construction. A new CEO was hired. The project now has tier-one names attached across the build chain: non-binding MOUs on equipment, financing support, distribution, and engineering/delivery. Early works have started on site. The project-financing process is now underway.

The stock still trades as if almost none of that progress has been priced in.

And that is before the main point of the thesis.

My conviction is that the UFS materially understates Great Atlantic’s earning power. Most of the next 40,000 words are the work behind that sentence.

The first step is the salt market itself. A large part of the story, and a large part of the mispricing, sits there.

02. The Salt Market

In January 2026, for the second consecutive year, parts of Ontario ran out of contracted deicing salt supply and turned to the spot market. Emergency spot prices reached nearly C$300 per tonne, more than 4x normal bulk pricing, in a matter of weeks.2

More than a hundred trucks lined up outside Goderich, the world’s largest underground salt mine, waiting for product the mine couldn’t deliver fast enough. Municipalities were rationing supply, spreading salt at just a quarter of normal rates, while private contractors scrambled for every tonne of deicing salt they could find. They were paying nearly C$300 per tonne for what little they could get. They had to. And in a hard winter, they will again.

This is what an undersupplied and fragmented market looks like when winter hits. And yet, the salt market has no exchange ticker, no ETF, and no CNBC segment. A product as old as civilization, a use case nobody finds exciting, and a US$26 billion market that Wall Street can’t be bothered to cover.3 Precisely the kind of sector worth digging into.

The $26 Billion Picture

The global salt market is worth around US$26 billion, with annual production of approximately 270 million tonnes. Deicing is only about 13% of global salt demand, and even that demand is not spread evenly across the map. It sits where winter actually forces governments to act: North America and Northern Europe. Northern Europe matters, but it is approximately one-quarter the size of North America’s deicing market. So we start here.

And even North America is too broad. Road salt is not really a continent-wide market. It is a logistics market wearing a commodity label. The core sits in the East, where winter is frequent and unforgiving. For Atlas, the relevant map is therefore much smaller: eastern Canada and the U.S. East Coast.

The eastern North American deicing market consumes between 30 and 36 million tonnes of salt in any given year, depending on winter severity.4 This is, for all practical purposes, one cross-border physical market, even if it remains fragmented commercially and logistically. Canadian rock salt moves into American cities, and American production moves north into Canada.5 In plain terms, the border is broadly administrative.

The U.S. accounts for the majority of this consumption and provides by far the most granular public data. The analysis that follows relies on U.S. data because that’s where the data is, but in practice, market dynamics are similar enough on both sides of the border for the U.S. data to be useful.

In most commodity markets, a persistent supply gap attracts capital, but not in eastern North American salt. When a product can cost more to ship than to produce, every additional kilometer is a burden. That is where the economics are decided, including Atlas Salt’s future economics.

Not All Salt Is Created Equal

Walk through any Department of Transportation (DOT)6 yard in February and you’ll see the same molecule in three or four different forms. A dome of dry rock salt. Tanker trucks of clear liquid brine. A second pile, slightly damp and faintly brown, coated with magnesium chloride and an organic additive. At the spreader, an operator sprays more brine onto the dry salt as it leaves the spinner.

At the center of all of it is sodium chloride. The variations are delivery systems around one feedstock.

The standard taxonomy sorts salt by production method and hides the chain underneath.

1. Rock salt comes straight out of a mine.

2. Brine is rock salt dissolved in water (in winter maintenance).

3. Vacuum pan salt is produced by evaporating purified brine back into crystals.

4. Only solar salt takes a different path: seawater or lake water evaporated by the sun in coastal ponds.

All four are sodium chloride and all four can melt ice.7 The market is still not a four-way contest. Vacuum pan salt does not belong on highways, it is too expensive for that. Its markets are food, pharma, and industrial applications. Brine reaches the road as a liquid for anti-icing and pre-wetting, and the brine DOTs spray is almost always made on-site by dissolving rock salt in water. Collapse the chain, and the deicing market has two real upstream inputs: locally mined rock salt and imported rock and solar salt from coastal evaporation ponds.8 Everything between the mine and the road is downstream chemistry around one of those two.

One Molecule, Three Jobs

Winter maintenance runs on three operations.

Anti-icing happens before snow falls or freezing rain begins. Crews spray a liquid brine onto dry pavement hours ahead of the event, typically a 23.3% NaCl solution, the eutectic concentration9 of sodium chloride in water. The brine leaves a thin salt film on the pavement and stops ice from bonding to the asphalt. When the storm arrives, plows can clear the road. Anti-icing has become standard practice across most snow-state DOTs because it prevents snow and ice from bonding to the pavement, making plowing more effective and materially reducing the amount of chemical needed versus reactive deicing alone.10

Deicing is the reactive application during and after the storm: dry or pre-wetted salt spread onto bonded snow and ice to break the bond so plows can scrape it off. This is what most people picture when they think of “salting the roads.”

Pre-wetting is the in-between move. Liquid brine is sprayed onto the dry salt at the spinner, just before it hits the pavement. Wet salt sticks to the road instead of bouncing into the ditch. It dissolves faster and starts working sooner. Agencies that use pre-wetting properly report 20% to 30% reductions in total salt usage at the same level of service.11

Less rock salt, then? Atlas is going to mine rock salt. Bad for business, right? No. Where brine is used, it is usually made on site by dissolving rock salt in water. Treated salt is rock salt coated with small amounts of chloride brine and, in some cases, organic additives. The dry salt is rock salt.

A single tonne mined in Goderich or Cayuga, or eventually in the future Atlas Salt’s Great Atlantic Salt, can leave the mine and end up on the road as dry crystals, as 23% brine, or as treated salt with a chemical coating. The downstream chemistry is the buyer’s local decision. The upstream supply is the same molecule from the same kind of mine.

TINA

Calcium chloride and magnesium chloride sometimes get called “alternatives” to road salt, which misreads what they do. Their economics, at 3 to 5 times the cost per tonne, rule them out as primary deicers. Their job is narrower: extending the working temperature range of a rock salt program.

Sodium chloride starts losing effectiveness around -10°C. Calcium chloride works down to approximately -25°C, magnesium chloride down to about -15°C. When pavement temperatures collapse, agencies do not switch to calcium chloride across the board. They spike a calcium chloride solution into their brine, or coat their rock salt with it, and get the extra temperature range without paying full alternative-chloride prices on every tonne. Most treated salt is some version of this: rock salt with a brine coating that extends its useful range by 10 to 15 degrees.

Acetates and formates are a separate category. Potassium acetate, sodium acetate, calcium magnesium acetate, and related products do not contain chloride; they are therefore generally less corrosive. That matters mainly in two places. Airports, where chloride contact with aircraft is a non-starter and deicing products must meet specific aviation standards. And a small set of bridges and parking structures where corrosion risk can justify paying 10 to 20 times the cost of rock salt.12 Everywhere else, acetates and formates remain a rounding error in volume.

Sand and abrasives are also used, but they play a different game. They do not melt anything; they provide traction on top of ice when temperatures drop below what any chemical can fix, or on rural roads where the cost-benefit math does not justify a chemical program at all. Most agencies that use sand still mix in 5% to 10% salt to keep the stockpile from freezing solid.13

I will leave the final word to the United States Geological Survey14:

The Implication for an Upstream Producer

A rock salt mine sits upstream of the entire deicing system. Dry tonnes for direct application. Brine feedstock for anti-icing programs. Base material for treated salt and pre-wetted formulations. Almost every operation, every chemistry trick, every “smart-salting” innovation municipalities have adopted over the last two decades starts with rock salt.

Atlas sits at that input point: dry salt for direct use, feedstock for brine and treated products, and emergency tonnage when winter breaks the procurement plan. It is building a feedstock source for a system already under strain, and in hard winters, one setback away from breaking.

Let’s focus on the strain.

How the Market Actually Works

The Buyer Side

The buyers split into three tiers.

The first is government: state and provincial Departments of Transportation (DOT), counties, municipalities, the people who keep highways open or get blamed if they do not. Together they account for approximately 70 to 80% of North American deicing rock salt consumption (a national average; the split varies meaningfully by state and by density of development).15

The second tier is commercial. Professional snow management companies servicing parking lots, warehouses, hospitals, and corporate campuses under seasonal contracts, plus the facility managers of large retailers and logistics operators who buy directly. About 15% to 20% of volume.

The third is retail. Homeowners buying 10 kg bags at Home Depot, Walmart, grocery stores, and hardware stores. The remaining 5% to 10%, give or take, is an entirely different economic creature. The buyer here is the person who slipped on their own porch last February and decided never again.

The three tiers buy on completely different terms.

Government buys most of its deicing salt on the calendar. A state DOT or provincial transportation agency signs its contracts months ahead of winter, usually between May and August, for a season that won’t start until November. The salt is usually delivered in September and October, stored in the salt domes and sheds you see along the highways, and used through April. The public buyer usually does not buy from a single national supplier; it carves its territory into zones (e.g., Massachusetts has eighteen contract zones) and each zone is awarded to a single vendor for the entire season. The contract sets a volume band: the buyer commits to a minimum, usually 60-80% of estimated quantities, and the vendor commits to a maximum of 110-140%.16 Anything above that contracted band falls outside the contract economics. Once the protected tonnage is gone, buyers are competing for uncommitted supply at market prices. Call it spot, emergency supply, or panic procurement; the mechanism is the same. Ontario in January 2026 was just the mechanism in public view. Counties and municipalities buy in smaller quantities, sometimes through their own contracts, more often through pooled bids run by a state or provincial agency to capture the volume discount.17

Commercial buyers play a harder game. Their contracts are usually signed between a property owner or facility manager and a snow management contractor, not directly with a salt supplier. The pricing model determines who carries the weather risk:

• In a per-push contract, the contractor gets paid each time crews service the site.

• In a per-inch contract, the price rises with snowfall depth.

• In time-and-materials contracts, labor, equipment hours, salt, and other materials are billed as used.

• In a seasonal flat-rate contract, the customer pays one fixed fee for the winter, regardless of how often it snows.

Each model transfers weather risk differently, but all of them end in the same place: the contractor has to source the salt.18 There are approximately 22,000 commercial snow contractors in the United States.19 Most of them buy from a regional bulk distributor. That layer exists precisely because integrated producers won’t build a commercial team, a dispatch desk, and a fleet of trucks to chase 22,000 individual contractors, many of them ordering a few hundred tonnes. The math only works at public-bid scale.

Retail does not really buy; it panics (yes, I’m still talking about deicing salt). When a storm is forecast, shelves empty. Season-to-season demand is more volatile than the two bulk tiers, but the long-run floor is the same: across much of the Northeast, homeowners and property managers are expected, and often legally required, to keep sidewalks and walkways clear. A slip-and-fall claim20 turns a ten-dollar bag of salt into cheap insurance.

Everyone needs salt before the storm. But not everyone can get it on the same terms. That is why a 10 kg bag of salt can end up costing several times more per-tonne than the bulk road salt purchased by a DOT, even before a storm hits.

The Seller Side

On the seller side, 4 structurally different types of players exist. Each controls a different bottleneck.

The first is the integrated producer. Compass Minerals owns the Goderich mine, operates the mine-site loading infrastructure, controls an extensive network of regional storage depots, contracts for bulk vessels, barges, rail, and trucking, and sells directly into public-sector contracts. This end-to-end integration took decades to build, but it allows the producer to invoice on a delivered-price basis. Almost all major North American rock salt producers are integrated this way, and that is part of what makes the segment effectively closed to direct competition.

The second is the port operator. Take Eastern Salt. It does not need to own the mine. It owns or controls the local bottleneck: bulk vessel unloading, storage, trucking, and relationships with public buyers. Eastern Salt buys imported salt from foreign producers, brings it into its terminal, stores it, and resells it at a delivered price to the same public buyers. Without this intermediary layer, distant importers would struggle to sell physical tonnes into the United States.

The third is the regional bulk distributor. Companies like Midwest Salt, Ninja De-Icer, or SISCU. They do not mine anything. They buy in volume from integrated producers, from port operators, and sometimes directly from exporters, then they store the salt in their own depots and resell it in flexible truckloads to the thousands of contractors no integrated producer wants to serve directly. They run 24/7 in season and deliver next-day, sometimes same day. They’re the lubricant that lets an already-tight system function, until the salt itself becomes the constraint. Lately, that has meant many winters (more on the causes and consequences later). Without them, the commercial tier collapses.

The fourth is the distant importer, mainly Chilean, Egyptian, and Moroccan producers. They load a bulk carrier at the port of origin, cross the ocean in two to four weeks, and unload dockside in Boston, Philadelphia, or Baltimore. They own nothing in the United States. They sell to a port operator or to a distributor with import capacity, and that buyer takes it from there. What makes their model economically viable despite the distance is their extremely low FOB port (Free on Board, the price at the port before shipping costs): dry climates that allow solar evaporation, lower labor costs, open-pit mines with no underground CapEx. Even after 2-3 weeks of ocean freight, the dockside cost in Boston can still remain competitive.

Integrated North American producer, port operator, regional distributor and distant importer. There is one box with no relevant occupant today: a North American producer that sells dockside without downstream integration, the way an exporter would, but a few days from the market instead of a few weeks. No great mystery here: no new rock salt mine has opened in North America in nearly twenty-five years, and the existing mines are almost all integrated by historical inheritance. Atlas Salt could therefore establish a new category, unless it is acquired first (a scenario we’ll cover later).

In short, here is the structure of the deicing rock salt market:

Demand That Does Not Bend

About 70% of U.S. roads are in snowy regions, and nearly 70% of the population lives there. Most deicing salt consumption is concentrated east of the Mississippi. Highway deicing alone accounts for about 41% of total U.S. salt consumption.21 That single application ties year-to-year deicing demand to one variable above all else: how harsh the winter is. “Harsh” here means the frequency of precipitation events that cross the freezing threshold, not temperature alone. Several low-snowfall icing events can consume more salt than a single large storm, which is “good” news for suppliers, because those events occur more often.22 Even during normal winters, North America still imports 8 to 10 Mt of deicing salt, leaving a clear source of demand for a new local supplier able to deliver more reliably.23 Atlas Salt is the only new local greenfield supply candidate positioned to fill that role. The only one in decades.

On the downside, demand has a built-in floor. State DOTs and provincial transportation agencies operate with minimum stockpile targets regardless of how mild the previous winter was. A warm season spreads its demand reduction across subsequent years as agencies draw down reserves before reordering. Private contractors operate much the same way, although they are often the last to be supplied and the first to return to the market when winter conditions turn severe. This asymmetric structure is what makes the market so attractive: demand can surge in any given year, but it rarely drifts far below its baseline. And it shows in the price. Over the past 40 years, U.S. rock salt PPI has compounded at 4% per year versus 2.8% for CPI, and the largest cumulative price decline never exceeded 10%, outside of post-spike corrections. On a chart, it looks like a compounder on a log scale.

This chart is one of the most important in this investment thesis. Salt price is the single most important variable in the investment case, by far. The BLS index tracks the prices received by domestic rock salt producers on their first commercial sale, whether to a DOT or through an intermediary. National, weighted, and smoothed by design. It hides three things.

First, geographic price dispersion: a DOT hundreds of miles inland pays far more per tonne than one next to a lake port, because salt’s low value-to-weight ratio means logistics dominate the delivered cost. The national average erases that gap. Second, downstream markups: the index captures what the producer receives, not what the distributor charges a private contractor three layers down. Between the mine gate and the parking lot, margins expand significantly. Third, and most importantly, spot pricing: the channel where both DOTs and private contractors scramble for every available tonne when winter runs harder than the procurement contracts anticipated.

That spot market is where crisis prices live. Ontario’s 2025-26 winter is the extreme case, but the same mechanism plays out at a smaller scale in many harsh winters, invisible in the headline index. The chart is therefore a proxy for producer pricing with the upside spikes flattened. It hides the operating leverage of every salt producer. In the valuation section, this chart will be our base case. Note that spot-price spikes tend to flow through into the next contracting season, as buyers and suppliers reset expectations around scarcity and replacement cost. Given the spot-price spikes observed in several regions during the 2026 winter season, contracted salt prices are likely to move higher in the next tender cycle.

To understand why public buyers end up paying several times the normal price,24 you have to think like them. As always, it comes down to incentives. Research on a 30-mile Iowa highway segment found accidents increased 1,300% when maintenance was disrupted during severe winter weather events, even as traffic volume fell.25 That 14x multiplier translates into lawsuits, emergency response strain, and political cost for whoever made the call. No politician on either side of the border wants to be the person who let roads go unplowed. Cities apply salt at the first sign of freezing precipitation, and keep applying as long as the supply holds. Private snow-removal contractors operate on the same arithmetic, through a different legal regime. A slip on an unsalted parking lot can become a lawsuit against the property owner. The incentive structure is identical; only the courtroom changes. For both public and private entities, when salt is needed, the money is found.

But sometimes there is simply no salt. The 2008-09 season was the clearest example: depleted stockpiles from the prior winter + Mississippi flooding + hurricane-damaged Louisiana mining equipment = a perfect storm. Iowa DOT opened bids for 31 locations and received… zero offers. The first time in the state’s history. I’ll let you imagine what that meant for the municipalities on the other end of those bids.26

And the public buyer still gets the “better” end of the deal. Private contractors sit below municipalities in every producer’s allocation queue. In early 2026, major suppliers like Morton Salt suspended commercial accounts across several states to prioritize public-sector contracts.27 Contractors received a single-line email from their distributors: “Everybody’s cut off. No more salt.”28 Around the same time, American Rock Salt raised its commercial price by $25 per ton in New York.

Demand is locked by public safety mandates and commercial liability, floored by winter stockpile behavior, prone to spike when winter runs harder than planned, with no viable substitute at any point in the cycle. Whatever goes wrong in this market, it won’t be on the demand side.

A Supply Base Frozen Since 2001

U.S. rock salt production in 2019 was 19.1 million tonnes.29 North of the border, the eastern supply base is heavily anchored by one asset: Compass Minerals’ Goderich mine, which produces approximately 7.25 million tonnes per year from 1,800 feet beneath Lake Huron. But rock salt does not only serve deicing; it also feeds chemical, industrial, and agricultural markets. Once those volumes are accounted for, domestic North American rock salt production falls well short of deicing demand alone. North America must import 8 to 10 million tonnes of deicing salt every year just to keep roads clear.

That gap is the persistent consequence of a supply base that has not expanded in over two decades while demand has quietly grown with population, road networks, vehicle miles, and service-level expectations.30

At the mine gate, U.S. rock salt is a low-value commodity, with benchmark prices around $50-$60 per tonne.31 Not an operating cost, but a producer-level selling price. But that mine-gate price can become an afterthought. In some lanes, transportation costs can exceed the value of the salt itself. This is why geography is the moat: the mine closest on the efficient shipping route usually wins. Short-sea shipping moves the same tonne for a fraction of what trucking costs. The mine closest to a deepwater32 port can serve the largest addressable market. A mine without port access serves whatever fits inside its trucking radius.33 Keep that in mind.

The last new rock salt mine built in North America began production around 2001. Before that, there had been no new mine for more than fifty years. The reason is purely economic: salt deposits are vast and widely distributed, but the revenue per unit is too low to justify the capital expenditure of a greenfield operation unless the logistics are exceptional.34 Even expanding existing mines is difficult: regulatory agencies may require complete facility upgrades to current standards if operators seek permits for new sections, creating compliance costs that are simply prohibitive.35

The major producing base is aging.36 Many of these mines sit hundreds of meters underground, beneath lakes or urban areas, with environmental and permitting constraints. Goderich is the clearest illustration. The mine experienced geological movements that disrupted production in 2017, followed by a three-month worker strike in 2018 that tightened supply across both sides of the border.37

And it is getting worse. Since 2021, North America has lost about 1.5-2.5 Mtpa (million tonnes per annum) of domestic capacity with the permanent closure of Cargill’s Avery Island mine (operating since the mid-1800s). Another 5.5 Mtpa remains in uncertain hands: Cargill’s remaining U.S. assets, both beneath lakes, have been listed for sale since 2023 with no buyer in sight.38 Compass Minerals, operator of the continent’s largest mine at Goderich, is under financial pressure: cutting production in 2024-2025 to pay down debt after a failed lithium diversification, a class action settlement, and a particularly damaging food-grade recall.39 Stone Canyon, North America’s largest salt producer by total volume, is digesting $5.2 billion in acquisitions.40 The rest are either absorbed in integrating existing assets or too regional to move the needle.

In this aging and saturated landscape, one name emerges: Atlas Salt, the only greenfield rock salt mine project in North America in 25 years with a completed feasibility study and early works underway. That scarcity is not a coincidence. The project has structural advantages that make the math work where it has not worked for anyone else in a generation. But that is only one project. For now, the continent still has to fill the gap somewhere else. It imports.

Import Dependence, Weeks Away

Deicing is the most important end use for imported salt. The question is where those 8-10 Mtpa come from, and what constraints they face.

After Canada, Mexico is the closest large foreign source for U.S. salt imports. But Mexican solar salt serves the U.S. West Coast and Asia almost exclusively. For the North American East Coast market, Mexican salt is out of the picture.41 Chilean and Egyptian solar/rock salt reaches the East Coast by ocean, a journey of 14 days or more. It is produced cheaply thanks to low labor costs, low extraction costs, and proximity to deepwater ports.42

When everything lines up, this system works. Imported salt can arrive at the same price, sometimes even cheaper, than the domestic product. It takes lower foreign production costs, favorable exchange rates, and ocean freight rates that undercut overland rail and trucking.43 But that cost advantage rests on everything going right, and it can reverse the moment one variable moves against them. Right now, with Brent having doubled in four months, nothing lines up for foreign salt exporters. And even if fuel prices collapsed tomorrow, cost is only the first problem; the others are time, priority, quality, and recourse.

Even when the salt arrives, it is not necessarily reliable. Distance means less visibility and less control over the counterparty. When a buyer sources foreign salt, it may not know whether it is dealing with a broker or the mine owner, whether its order will be prioritized, whether the product delivered will match the quality purchased, how long the vessel will wait before loading, or what dispute process exists if something goes wrong. Available salt can still become a reliability problem. And even if everything goes well, salt sitting two weeks away by boat cannot help a municipality that runs out of inventory in the middle of a storm. Only a few local producers can.

The policy environment has become less… comfortable with distant imports too. Non-USMCA sources like Chile, Egypt, and Morocco have also faced Section 122 surcharges of 10% following the February 2026 Supreme Court decision.44 USMCA-compliant Canadian salt enters duty-free.45 So far, Canadian road salt has remained outside the main U.S.–Canada tariff pressure points. Canadian rock salt qualifies as USMCA-originating by default, and USMCA-compliant goods remain exempt from the IEEPA tariffs hitting most other Canadian exports. Even if that changed tomorrow, the math would not move much. The deficit is too structural for buyers to source elsewhere, the end customer is a municipality that has to keep the road open, and most North American producers run on margins too thin to absorb a tariff themselves.46 The cost would land where it often lands in tight markets: on the U.S. consumer.

The U.S. has tried to promote domestic salt production, but policy cannot override physical reality. New York’s Buy American Salt Act (December 2022) allows state agencies to prefer U.S.-mined salt when the bid is within 10% of the lowest offer.47 But in practice, the policy mainly exposed how fragile domestic logistics still are. After the 2024-25 shortage, New York officials stepped back from applying the preference to the next road salt solicitation, explicitly stating that the Buy American Salt Act preference would not apply to the 2025-26 replacement contracts.48 They still need imports, and they need them nearby.

A demand floor with sharp spikes, and a supply base that shrinks to local producers when winter bites. In the middle of all this, Great Atlantic Salt is 2 km from a deepwater port on the Atlantic seaboard. The project alone cannot close the entire gap, but its economics are built around this market structure.

What’s left is where supply and demand meet: the price.

Pricing Without a Market

Salt has no liquid futures contract, no exchange-traded benchmark, and no transparent spot market. Pricing happens through negotiated physical contracts, awarded months before winter, or through emergency quotes when everyone is chasing the same tonne of salt.

Contract prices tend to move gradually upward, while emergency prices can reset violently when inventories run short. The buying logic is starting to move from pure lowest-cost purchasing toward security of supply, especially after the shortages of 2025 and 2026. That shift structurally favors domestic producers with proximity to end markets and the ability to respond within days. Atlas Salt can deliver salt in under three days to the nearest East Coast cities (Boston, New York, etc.).49

There is another slice of the deicing market I’ve barely touched so far. It is small in volume, but it plays by completely different economics.

Where the Margins Are

In sales to public buyers, producer margins are thin. The counterparty is a government entity with a public-safety obligation to keep roads usable, and sometimes a legal obligation to do so. This is the floor.

Now take that same salt, bag it, slap a brand on it, and put it on a shelf at Home Depot. At the producer level, packaged salt already sells for almost 2x bulk. By the time it reaches retail, the multiple expands to approximately 6x. Make no mistake: it is the same salt: same NaCl, same purity, same mine. The markup is mostly a function of packaging, distribution, and channel access.

This is why evaluating a salt producer on average revenue per tonne is misleading. What matters is which channels they can reach: packaging lines, logistics networks, and shelf space at major retailers. Few producers have all three. Atlas Salt does not, and probably never will, at least not directly. That said, the company is already trying to capture part of those higher-margin segments through partnerships.

Climate & Environment: The Objections That Do Not Stick

Two objections come up every time someone looks at this market, and they deserve answers before we move to the setup. Both sound reasonable, but neither survives contact with the data.

The first: climate change will kill winter salt demand. No data points in that direction. Instead, climate change is changing the shape of demand. Total precipitation across the Great Lakes states has increased 14% since 1951, while the heaviest 1% of storms have become both wetter and more frequent.50 Warmer average temperatures shift more of that precipitation from snow to rain, and specifically to freezing rain, which can consume more salt per event than snowfall. Researchers studying county-level records from NOAA’s Storm Events Database from 1996 to 2025 found that freezing rain events are not declining. Instead, they are shifting in location and timing, moving later in winter and expanding into regions less prepared for them.51 As one research team put it52:

Climate change makes deicing salt demand less predictable, more concentrated in peak events, and more punishing when supply cannot respond. That, once again, favors producers with proximity.

The second: environmental regulation will restrict road salt. Here the picture is more nuanced. The pressure is real: chloride accumulates in freshwater systems, and the EPA sets toxicity thresholds at 230 mg/L for chronic exposure. Several states have tightened water quality standards. But the regulatory response on the East Coast has been the same: optimize application, do not eliminate it. Pre-treatment brining, calibrated spreaders, GPS-controlled application, anti-icing strategies, and operator training are already part of the toolkit across public winter-maintenance programs.53 No relevant jurisdiction has moved to ban or materially restrict road salt use. The reason is the same one that makes this market what it is: there is nothing else that works at scale for the price.54

Climate change raises the peaks and, in some regions, the floor. Every efficiency gain of the last fifteen years has been absorbed by rising demand. Rock salt consumption is higher today than it was before chloride in rivers became a water-quality concern.

Where This Leaves Us

Every structural feature of this market shows up in a producer’s income statement: revenues, margins, cash flows.

Six structural facts, each compounding the next:

• Demand that does not bend, locked by public safety mandates and by slip-and-fall liability on the commercial side, floored by winter stockpile behavior, compounding at nearly 4% a year for four decades in the floor case.

• A supply base that has not expanded in twenty-five years and is showing stress: mines closing, producers under financial pressure, and capacity for sale in a market that still needs more local supply.

• An 8-to-10-million-tonne structural import gap filled by salt that takes two to three weeks to cross an ocean.

• No economic substitute at scale.

• Climate change making demand less predictable and more peak-loaded.

• A procurement model shifting from lowest-cost to security-of-supply, rewarding exactly what distant imports cannot offer: proximity and reliability.

This is a market that has already broken, visibly, and repeatedly. The structural conditions that caused the break are getting worse. Ontario in January 2026 was the system behaving exactly as designed, with no slack left.

Into this market, one company is building the only new salt mine in North America in over a generation. A shallow, high-purity deposit. Two kilometers from a deepwater port. An Atlantic port open while the Great Lakes are closed. Completed feasibility study. Environmental approval secured. Early works underway. And a valuation that prices in almost none of it. Even less of the market it sits in.

That brings us to Great Atlantic itself.

03. Atlas Salt’s Mine: Great Atlantic Salt

Atlas Salt is, for underwriting purposes, a single-project company. The asset is the Great Atlantic Salt deposit, in western Newfoundland and Labrador (Canada). Here’s why this is the first greenfield rock salt project in 25 years to clear every barrier that has kept new salt mines out of North America.

The Deposit

Start with the rock itself. Atlas’s September 2025 Updated Feasibility Study (UFS; remember this acronym, because I’ll use it a lot) outlines a 24.25-year mine life based on 95.0 million tonnes of Probable Reserves grading 95.9% NaCl (sodium chloride, common salt). Those reserves sit inside a much larger resource envelope: 383 million tonnes of Indicated Mineral Resources at 96.0% NaCl, inclusive of the reserves, and 868 million tonnes of Inferred Mineral Resources at 95.2% NaCl. In plain English, the current mine plan uses 95 Mt, while approximately 288 Mt of additional Indicated Resources and all 868 Mt of Inferred Resources remain outside the reserve-case mine plan. All three categories exceed the 95% NaCl road-salt specification referenced in the UFS. The salt does not need beneficiation or chemical upgrading. Once crushed and screened to specification, it can be sold as road salt.

Mineral Reserves are the drilled, tested, and engineered part of the deposit, with enough confidence to support a bankable mine plan. Indicated Resources carry less confidence than reserves. Inferred Resources carry less confidence again.

The 24.25-year mine life in the UFS has to be understood on those terms. It is a planning horizon built on the Probable Reserve base. SLR, the firm that prepared the UFS, made that explicit: further drilling during production could convert additional material into reserves, and the expansion case extended mine life to 47.5 years, on partial reserve support and with no commitment that the plan would ever be executed.55 The UFS cannot capitalize those later tonnes in reserve-case cash flow today, while management has little economic reason to spend aggressively to prove year 50 before year one exists.

Bedded salt deposits like Great Atlantic tend to convert inferred resources to reserves at rates metal projects rarely match. The deposit is homogeneous, laterally continuous over kilometers, high purity, and grade appears not to vary meaningfully across the envelope. For a mining project, that is exactly the kind of geology you want: boring, repeatable, and hard to misread. Goderich, now Compass Minerals’ flagship mine, became the world’s largest underground salt mine from a remarkably thin original drilling base: 16 vertical core boreholes drilled in the 1950s, later validated by six decades of mining.56 Atlas has drilled forty-eight holes in a shallow, laterally continuous deposit.57

The UFS cannot price what is not yet a reserve. That’s the rule, and SLR follows it: Approximately 288 Mt of Indicated Resources beyond the reserve base, plus the 868 Mt of Inferred Resources, sit outside the reserve-case mine plan and the NPV. I’m not bound by that rule. I could assign part of that 1,156 Mt base a probability-weighted value, based on the geology and the historical track record of bedded salt deposits. I’m not going to. I’m leaving it as free upside. The 95 million tonnes of Probable Reserves are already enough for the thesis.

But there is more. Management’s view is that, once underground, the mine may be able to keep moving horizontally beyond the current reserve, then beyond the current resource, and eventually into undrilled ground. If true, this could translate into lower sustaining capital, lower development cost, and lower unit operating cost than the reserve-constrained UFS can model today. Think of the mine as a stack of levels. Opening a new level is more expensive than extending development horizontally within an existing one. So if Atlas can recover more tonnes than planned from the same level, the economics improve: the fixed development cost is spread over more salt. But again, the UFS cannot include that in the reserve-case economics today. And again, I leave it out, more for lack of data than out of excessive conservatism.

Transport Logistics as a Moat

Once purity and gradation specs are met, deicing rock salt is largely interchangeable. Geography, and the logistics that flow from it, are the moat. Atlas is built around that.

The project sits two kilometers from Turf Point, a deepwater port on the west coast of Newfoundland, with direct access to Atlantic shipping lanes. The design calls for a conveyor linking the mine directly to a ship-loading facility sized for Handysize and Handymax bulk carriers.58 No trucks, just a modest line item in initial CapEx and ongoing maintenance, and the mine-to-port leg is solved upfront. If built as planned, Great Atlantic would be able to serve Boston, New York, Philadelphia, or Baltimore within a 3‑ to 5‑day round trip, narrowing that gap between import lead time and emergency demand in a way Goderich, Cote Blanche, or Chilean and Egyptian salt simply cannot match on the same winter-response basis.

It gets worse inland. Cargill’s remaining U.S. mines and Compass Minerals’ Goderich rely on the St. Lawrence Seaway and the Welland Canal for waterborne access, which close every January and reopen in late March.59 That is exactly when a harsh winter tears through contracted tonnage: DOTs hit their contracted volume ceilings, private contractors run dry, and the spot market takes over. At that point, price is set by whatever tonne can actually reach the customer. Great Lakes producers go silent for ten weeks every winter, restricted to overland delivery within trucking range. Atlas could keep shipping.

The rest of the year, when the Seaway is open and everyone bids into the same public tenders, Atlas’s positioning should still win on distance to market. A short North Atlantic haul to Boston, New York and Philadelphia beats the Great Lakes detour from Goderich, and beats anything Chilean or Egyptian supply can put on a two-to-three-week ocean route. The geography pays in every season, but it also costs something.

Management has built Atlas’s logistics pitch around this advantage, describing Turf Point as “year-round with no seasonal ice closure.”60 That needs one clarification. Turf Point sits in St. George’s Bay, a sub-basin of the Gulf of St. Lawrence that the Canadian Coast Guard designates as Ice Zone Area 2, with a formal ice season running December 21 to April 15.61 This is not the same as being ice-free. However, the Gulf remains navigable through winter with icebreaker support, while the Seaway physically closes. Commercial vessels over 200 gross tonnage transiting an ice zone during the ice season pay a Canadian Coast Guard icebreaking fee. For the 2025-26 fee, that rate is C$3,818.61 per transit.62 Under plausible assumptions about vessel size, billing mechanics, and how much of Atlas’s shipment calendar falls inside the ice window, that translates into somewhere between C$0.03 and C$0.10 per tonne shipped in direct icebreaking fees.63 Against the UFS LOM operating cost of C$28.17/t, it is a rounding error.64

The logistics advantage is physical, but it is neither free nor fully built. Turf Point is an existing third-party-owned port, already operating as an aggregate export facility. Atlas also ships small amounts of gypsum from its Flat Bay mine through Turf Point.65 The existing trestle, ship loader, and caissons can already handle Handymax vessels. What Atlas plans at Turf Point is a brownfield expansion: a new 47,300-tonne storage building, refurbishment of the trestle and ship loader, reclaim conveyors, dredging, and one additional mooring caisson. The UFS puts the port line at C$58.5 million, about 10% of the C$589 million initial capital. That creates normal construction and logistics risk. A second risk, more upstream: no commercial agreement yet binds Atlas to Turf Point. That is normal at this stage. The port owner has strong reasons to sign, and failure to sign strikes me as very unlikely. But it remains the key binary item of the setup. I treat it in detail in the risks section.

The rest is already in place. The St. George’s substation sits 1.4 kilometers from the project, the Trans-Canada Highway runs alongside it, and a regional airport is within reach. Atlas essentially bolts a shallow, laterally continuous salt deposit onto road, power, and air infrastructure that someone else built decades ago.

The Operational Design

Goderich’s mine sits at 550 meters. Cayuga’s, beyond 700 m. Great Atlantic’s is around 180 m. A third of the depth changes the entire economics of the project.

A deep mine needs a vertical shaft: a cage to lower crews, a hoist to lift ore in skips, several years of construction and several hundred million dollars before the first tonne reaches the surface. Atlas can skip that shaft-and-hoist architecture. At 180 meters, the gradient is gentle enough to drive declines that run from the surface down to the orebody. Trucks, conveyors, and personnel can move through ramps (and gravity) instead of relying on a vertical shaft.

The geometry pays a second time. Once underground, you attack the rock. The chosen method is room-and-pillar: you carve a grid of rooms into the salt, leaving pillars of salt untouched between them. The salt mostly holds itself up; that is the point of leaving pillars behind. The method works best on tabular, thick, homogeneous deposits, which is exactly what Great Atlantic is.

Then you have to break the rock. Most underground hard-rock operations still use drill-and-blast: drill holes, pack explosives, fire, clear debris, ventilate and repeat. Atlas chose a continuous miner, a machine that bites into the rock with a rotating head and drops it onto a haulage system. Salt is soft compared to granite or metallic ore, soft enough that mechanical cutting becomes the lower-cost option.

The rest of the flowsheet is deliberately short. An underground crusher and screens size the salt. A conveyor carries it up the declines. At surface, it passes through a 47,300-tonne storage building. From there, a second conveyor carries it the two kilometers to Turf Point, where a dedicated storage and conveyor-fed shiploading system would load vessels for export. But even simplicity carries risks, and the project’s risk profile gets its own section.

In video form (please forgive the music):

Because the design avoids drill-and-blast and uses an electric underground fleet, the ventilation load falls sharply. Great, because ventilation is one of the largest energy line items in a conventional mine. That allows Atlas to design the site as an electric-based operation: electric cutting equipment, battery-electric underground trucks and loaders, power drawn from Newfoundland’s hydroelectric grid. The ESG case wrote itself, and it happened to align with exactly what the province and the local communities wanted to see from a new mine.

The design is not being invented from scratch. Great Atlantic Salt draws heavily on Irish Salt Mining’s Kilroot mine: an underground soft-rock room-and-pillar mine, with an underground conveyor system, and product moved to port for shipment into the deicing market. Atlas’s senior engineers visited Kilroot, and the Great Atlantic study team included engineers with Kilroot experience, especially on ground support, underground conveyor structure, and drift access design. That does not remove execution risk, but it does show that similar mine architecture has worked before. It is also worth noting that Kilroot was developed in the 1960s.

A shallow deposit, a continuous extraction method, a five-step flowsheet, and a conveyor that ends in a ship’s hold. On paper, the operation is about as simple as a salt mine gets. But on the ground, simplicity tends to end where people begin.

The Social Component

A mining project also stands on its ability to clear permits, hold community support, and present a defensible environmental profile.

This part is not necessarily riveting, but it matters. For those who really do not feel like reading it, let me summarize: on these three fronts, Atlas is not only moving without friction but is actively supported by both local stakeholders and the provincial government.

For the rest of you, let’s dissect.

Permitting

Seven weeks. That’s the time it took Newfoundland and Labrador (NL) to release Atlas from its environmental assessment process, from filing to clearance. For mining projects, that is fast. Full environmental assessment processes can run 12 to 18 months, and full environmental-impact-statement cases can stretch over several years. Those seven weeks tell you at least three things. Two are obvious: the file was clean, and the regulator wanted it through. The third is less obvious, and we will get to it when we talk about who runs this company.

At the federal level, the IAAC (Impact Assessment Agency of Canada) confirmed in writing in December 2023 that the project is not subject to a federal impact assessment. Salt mines are not on the list of designated activities, and 4 Mtpa of production does not trigger a designated-project review. Two levels of government and two green lights.

Several approvals and project plans are already in place66: EA, Early Works Development Plan, Environmental Protection Plan, Waste Management, Water Resource Management, Wetland Conservation, Bat Preventative Measures, Gender/Diversity/Equity/Inclusion Plan, and Benefits Plan. What’s left belongs to the construction phase: a separate Capital Development EPP (Environmental Protection Plan), a Mill Licence before operation, and potentially a Fisheries Act Authorization for the in-water works at Turf Point (dredging and new caisson). That last one is the only potential bottleneck, with an up to 18-month process if triggered, but the four-year execution schedule has room for it (I contacted management on this point, and on several others, but as of today I have not received a response). Incentives are aligned across all parties, which makes it unlikely to become a problem.

Jurisdiction and community

NL ranks third among Canadian jurisdictions for mining investment in the Fraser Institute’s 2025 survey, and fourteenth globally. The regulatory framework is stable, the provincial government actively courts mining capital, and the region around the project has an unbroken mining tradition since the 1950s: gypsum for decades (still shipping today), and a limestone mine still operating. St. George’s is not a town meeting its first mining project.

Three stakeholders have formally voiced support: the Town of St. George’s, the Bay St. George’s Chamber of Commerce, and the First Nations Bands of St. George’s, Flat Bay, and Three Rivers. The Qalipu First Nation67 is still considering its support, but has raised no specific concerns. The project site already sits next to an active gypsum quarry, and Qalipu’s administrative base is approximately 70 km away. The probability of friction appears low.

Environmental profile

Credible ESG arguments are rare in mining, but Atlas ticks several boxes that make the defense unusually easy. The mine is designed to be an electric operation, powered by NL Hydro’s hydroelectricity at C$0.062/kWh. Projected Scope 1 emissions: 79 tonnes CO2 per year (approximately what four 4-person households in Newfoundland emit). No tailings, no chemical processing, and no diesel underground fleet. The project’s carbon footprint is a rounding error on a suburb. That helps explain why the environmental assessment approval was granted in less than two months, compared with approximately 1-1.5 years for other mining projects.

Sensitive-species on site? Only one issue has shown up in surveys so far. Bat acoustic monitoring picked up activity in July 2024, with follow-up results still pending. At this stage, that looks like routine procedural work, more likely a small cost line than a thesis risk. Black ash and raptor surveys turned up nothing. No known archaeological sites either.

Now the best “incentive” part: the projected economic return to the community is enormous: 170+ direct long-term jobs, nearly C$2.8 billion in tax revenue over the mine’s life, and active support for Memorial University’s critical minerals research. St. George’s is a rural community whose population declined 5.3% over five years and it is about to inherit a multi-generational employer.

Aligned incentives are rare in mining. It is one of the reasons even projects with exceptional numbers on paper, technically de-risked, stay undervalued for years. They die in the maze of permitting, opposing incentives, and community opposition. Atlas sits outside that maze. And yet it trades at a deeper discount than many projects stuck inside it.

Subscribe now

04. The Business Model

Compass Minerals, one of the biggest deicing salt producers in North America, owns and controls: mines, packaging plants, distribution centers, brand names, storage depots, vessels, barges, rail and trucking relationships, and direct contractual relationships with virtually every DOT and public-sector buyer East of the Mississippi. Morton Salt and Cargill have built the same end-to-end machine. They mine the salt, screen it, bag it, brand it, ship it, and sell it directly to the entity that will spread it on a road. These are fully integrated supply chains built over 60 to 130 years.

Atlas controls a planned mine project that is expected to start production in 2030.68

This section is an analysis of what that gap means.

One clarification upfront. As of the date of this report, Atlas has not signed a single binding commercial contract with any relevant counterparty. The key commercial relationships disclosed by management so far are non-binding MOUs (Memoranda of Understanding). Early works are underway on site, but remaining construction-phase permits, project debt, and equity gap funding are still ahead. First production is at least four years away under the schedule disclosed by management. This contractual posture is the normal state for a junior developer at this stage, and it is the starting point from which the rest of this section should be read.

The Product: One Feedstock, Several Commercial Formats

Atlas plans to sell essentially one feedstock: crushed rock salt for road deicing. The same raw material comes off the crusher, then gets separated into different specifications and formats depending on the sales channel, but not in equal proportions. Within the bulk road-salt segment, the UFS distinguishes two commercial specifications coming off the same crusher. Public road-salt tenders typically require the ASTM D632-12 standard.69 Commercial products sold through distributors demand tighter “Screened Mediums” specifications: the same 95% NaCl minimum, but lower moisture, narrower gradation, and fewer fines.70 The two streams would be separated at the processing plant: Atlas produces one product and screens it into commercial grades depending on which sales channel it feeds.

A smaller share could go to packaged consumer salt, sold through branded partners such as Scotwood Industries, the segment where Atlas could capture a meaningful piece of the approximately 6x markup that exists between bulk producer pricing and retail shelf pricing. A marginal residual goes to colored salt for specialty markets, irrelevant to this analysis and not contractually committed in any case.71

Where Atlas Plans to Sell

The planned 4.0 Mtpa production splits across four geographic destinations, each with its own logistics profile and pricing dynamics.

The U.S. East Coast represents 56% of planned volume and the bulk of the C$920M NPV. The UFS defines this market as ports along the U.S. East Coast from Maine down to Baltimore, plus inland states reaching as far west as Pennsylvania and West Virginia. This is the region where the 8-10 Mtpa structural import gap concentrates, and where Cargill’s permanent closure of Avery Island in 2022 removed another 2.5 Mtpa of domestic supply from the market. Realistically, 1.5 to 2.5 Mtpa could land somewhere along this coast, displacing the highest delivered-cost suppliers first: Chilean and Egyptian salt with multi-week ocean lead times, but also North American mines whose distance from the East Coast forces freight costs that Atlas’s geographic position simply does not carry.

Québec, at 15.5%, is the defensive line. The St. Lawrence corridor consumes an estimated 3.5 to 5 Mtpa every year, and for ten weeks each winter, Atlas would be one of the few nearby producers with waterborne access into the region while the Seaway is closed. As noted, the 15.5% allocation is not committed. But no new commercial infrastructure is needed to capture it: the corridor is already there, the buyers are already there, and the geography does the rest.

The Maritimes (Nova Scotia, New Brunswick, and Prince Edward Island) at 8.5%, is a smaller target, and a relatively standard one. Atlas would sell salt dockside at ports like Halifax or Saint John, and a local distribution company would handle unloading, storage, and delivery to the final customer. For those provinces, Atlas would become a nearby, reliable source they can tap in volume and during the season. For Atlas, it is a minor share of output with above-average margins. Geography, yet again, tilts the economics.

The 20% allocated to Spot Sales and Commercial is the most economically interesting line in the mix. It captures three distinct flows. First, emergency purchases by municipalities when their contracts run short (the Ontario 2026 dynamic, with spot prices at C$300/tonne). Second, ongoing sales through commercial distributors to private snow contractors, who get pushed to the back of the queue every time municipal demand spikes and end up paying multiples of normal bulk pricing. Third, the local Newfoundland market itself, approximately 0.3-0.35 Mtpa with no internal production, served by truck and short-haul vessel shipments directly to municipalities along the west coast. All three flows pay above tender pricing. And all three favour the producer who can ship year-round.

I’ll say it again: this is a planning allocation by design, and a deliberately conservative one. Atlas appears to want a larger spot-market allocation than the UFS assumes. As a shareholder, I would welcome that (more on that later).

How It Actually Sells: CIF, DAP, FOB

There are three commercial structures, depending on the destination. Each transfers risk and captures margin differently.

CIF (Cost, Insurance, Freight) is the standard structure for the bulk of U.S. East Coast shipments. Atlas would arrange and pay for ocean freight and insurance to the destination port, where a local distribution company manages unloading, storage, trucking to the salt depot, and final sale to the road-spreading entity. Atlas would carry the freight-cost exposure; the distributor would carry everything downstream. This is the dominant structure for the 56% U.S. East Coast allocation.

DAP (Delivered at Place) applies to the western Newfoundland market. Atlas would deliver all the way to the customer (typically a municipality) by truck or short-haul vessel depending on distance. Atlas would control the entire chain to final delivery, but the addressable market is small. Management has not named carriers or routes, but the economics point to contracted trucking rather than an owned fleet (at least in the early production years). At 0.3-0.35 Mtpa, the volume does not justify dedicated vehicle CapEx.

FOB (Free on Board) applies to buyers who arrange their own vessel at Turf Point and take title once the cargo is loaded. Atlas’s risk ends once the salt is loaded aboard the vessel. This is the simplest structure for Atlas operationally, but it can limit pricing power: the buyer’s negotiating leverage is highest when they control the vessel.

I hope you’ve been following closely, because this is where it gets important. Section 19.4.2 of the UFS:

The entire C$920M NPV runs on a uniform C$81.67 per tonne price on an FOB Turf Point basis, regardless of the commercial shipping terms. The assumption is simple and elegant. If anything goes wrong in the long-distance deicing salt chain (and plenty can), Atlas’s 2 km-to-port geography becomes structurally more valuable than the model assumes. Imported Chilean and Egyptian supply loses competitiveness first. Atlas keeps its relative position intact. The FOB-equivalent pricing assumption is therefore potentially a source of additional margin. Like the 288 Mt of additional Indicated Resources and 868 Mt of Inferred Resources earlier, I’m treating it as free upside. The setup still does not need it.

The Contracts

Section 19.5 of the NI 43-101 technical report captures the current situation perfectly:

“Atlas has had preliminary discussions with potential salt purchasers as well as salt brokers. Similarly, the Company has had preliminary discussions with logistics companies and port operators that would be involved in the process of delivering salt to destination markets. As of the effective date of the UFS, Atlas has not entered into any binding commercial contracts with respect to salt marketing or logistics.”

Nothing abnormal here. Early works on the mine have begun, production is at least four years away, and counterparties rarely commit to binding contracts that far ahead of execution. Atlas has already announced four non-binding MOUs (Memoranda of Understanding; statements of intent) with tier-one counterparties.

MOUs serve two purposes, depending on the counterparty: they can reduce the capital Atlas has to raise upfront, and they tell future lenders that a tier-one counterparty has already diligenced the file and stayed at the table. Take Sandvik. If that equipment MOU converts to binding vendor financing, Atlas could pay for the fleet out of post-production cash flow, not out of equity raised today at a depressed valuation. Each MOU has its own specifics, but the logic is the same: less potential dilution now and/or easier financing later.

Atlas currently has four such MOUs with major counterparties. Each is non-binding, but each is a potential milestone. Especially the first one.

Scotwood Industries, Commercial Offtake + Canadian JV (August 20, 2024)

Atlas signed the MOU thirteen months before the UFS was published. It bundles two things into one deal. First, a proposed strategic offtake: Atlas would supply bulk salt to the partnership. Second, a 50/50 joint venture: the JV packages that salt, brands it, and sells it into the Canadian retail market. A combined target of 1.25 to 1.5 Mtpa, with profits split equally after each side recovers its costs. The deal explicitly excludes bulk road deicing salt (the public-tender channel) and excludes the four Atlantic provinces (the market Atlas can reach on its own). With one important exception: Scotwood’s national Canadian retail accounts (think Walmart Canada, Home Depot Canada) can be supplied anywhere in the country, including the excluded provinces. Remember, this is the channel where the same salt can sell for several times the bulk price.

Why this matters.

Scotwood is the largest packaged retail deicing distributor in the United States: Home Depot, Walmart, big-box mass merchandisers, DIY chains, farm and home retailers, grocery chains, and others. When a distributor that size signs even a non-binding MOU with a junior pre-production developer, the file has passed a filter. Chase Wilson, Scotwood’s President, said so at signing:

“Being the largest distributor of packaged retail deicing salt in the United States we are very selective in choosing new business partners. If completed, we believe Atlas Salt’s long life underground salt project would make Atlas Salt an ideal long-term partner for us in Canada.”

Rick LaBelle, then CEO of Atlas Salt, was equally explicit about Atlas’s side of the logic:

“Since I joined the Company, I’ve been focused on opportunities in the retail market, where supply is more predictable, and margins are generally higher.”

This is where the UFS NPV stops being the full picture.

Every tonne in the C$920M NPV starts from a C$81.67 per tonne base price on an FOB Turf Point basis. SLR chose a single blended price across every destination and every product format, effectively anchored on bulk deicing economics (the lowest-margin channel in the deicing market). Walk into a Home Depot today and look at the shelf: a tonne of packaged deicing salt retails around C$500-600. The UFS prices essentially none of this differential.

Most of that retail spread won’t flow back to Atlas: retailer margin, distribution costs, packaging, and the 50/50 JV split with Scotwood all take their cut. What lands on Atlas’s income statement after those steps could be somewhere between C$60 and C$80 /t above the bulk base case (estimate),72 depending on cost structure and negotiation. At 1.5 Mtpa through Scotwood, that’s C$90 to C$120 million of incremental annual pre-tax contribution attributable to Atlas not captured in the NPV. Discounted at 8%, with no salt-price escalation, over the 24-year mine life, the C$60-80/t uplift on 1.5 Mtpa would imply approximately C$945 million to C$1.26 billion of pre-tax present value.

After taxes and leakage, the incremental present value would land closer to C$630 million to C$840 million.

Let me reframe this. With up to 37.5% of annual production routed through packaged/retail-related channels, the incremental value alone could approach the entire NPV built on bulk pricing. Recalculating the NPV with 37.5% flowing through the packaged/retail-related channel (the high end of what the MOU contemplates) would lift project value by 68% to 91%. Cut that estimate in half as a safety margin and the uplift still lands between 34% and 45%. Just from Atlas selling its salt in bags through a partnership. Back-of-envelope, but the direction is unambiguous. And the company trades at a P/NPV of approximately 0.13x against a project NPV that does not count any of it.

That was the picture at 1.25-1.5 Mtpa. Then the project scaled.

The 2023 Feasibility Study modeled 2.5 Mtpa. The 2025 UFS modeled 4.0 Mtpa. Yet the Scotwood target volume did not move. Yet. Nolan Peterson, who took over from LaBelle in June 2025, addressed this directly:

“For Scotwood, that tonnage was set in the original feasibility study, remember when the project was much smaller. So there is the opportunity to expand that offtake to encompass the larger project production, and we are in discussions with Scotwood and other offtake partners about taking the remainder.”

Management has begun the conversation to expand that offtake, and Peterson’s phrasing (“Scotwood and other offtake partners”) telegraphs that the search has widened. If the binding arrangement eventually routes closer to 50% of nameplate through higher-margin packaged/retail-related channels, the economics per tonne attributable to Atlas could reshape the project entirely.

Peterson added one more line worth hearing:

“It does not mean that we need a partnership like this. We can always develop those relationships ourselves and market the salt ourselves, but it’s very helpful. It’s one less thing for us as miners that are familiar with mining and developing projects that we have to worry about.”

The CEO makes the point explicit: the MOU is not on the critical path. Atlas can walk away from it if the terms do not evolve in its favor. But the shortcut is worth taking, in my opinion, even on less favorable terms. Replicating Scotwood’s retail chain would take Atlas years and a lot of capital, starting after first production, with management bandwidth it does not have today and won’t have at the start of production. Owning the full margin sounds better on a slide, but getting half of it, years earlier, without the CapEx, is probably the better trade. At least through the first debt-repayment window, until cash flow stabilizes. In any case, senior lenders would probably reject a go-it-alone retail buildout. When project lenders underwrite a project, they do it against a specific base case, and they are usually reluctant to accept deviations, especially ones that introduce that much commercial risk.

The NPV holds up fine if Scotwood goes nowhere. The stock could reprice materially if it does. That’s the kind of asymmetry a shareholder wants to own. It shows up, to a smaller degree, in the three MOUs that follow.

Sandvik Mining, Equipment + Vendor Financing (September 2024, expanded February 13, 2026)

The Sandvik MOU, expanded in February 2026, is also the kind of thing we want to see as shareholders. The expanded scope raised the total estimated commercial value from C$73 million to up to approximately C$132 million, covering both initial construction and the multi-year ramp-up to 4.0 Mtpa. Sandvik would supply underground mobile mining equipment, technology, and associated services.73 Sandvik is also Atlas’s Integrated Project Delivery (IPD) partner, which means it would be embedded in the mine design, automation planning, and long-term maintenance architecture from the start.

The financing piece is critical (and easy to misread). It contemplates vendor-supported financing for Sandvik capital equipment within the C$132M commercial scope, potentially less than the full amount, subject to “customary due diligence, receipt of Sandvik’s required internal approvals, and negotiation and execution of the definitive agreements.” It is an expression of intent from a tier-one global equipment supplier that, if converted, could materially reduce Atlas’s upfront cash equipment outlay. If not, Atlas would need to find another partner, or fund the equipment from project debt or equity. We come back to this in the financing section.

Hatch Ltd., Lead Engineering & IPD Partner (November 12, 2025)

Hatch is one of the world’s largest engineering firms (10,000 staff, 150+ countries, over seven decades of engineering experience and a track record in underground soft-rock mines), with an established Newfoundland and Labrador presence. The MOU establishes Hatch as Lead Engineering Partner and IPD Partner. In plain English: Hatch would integrate all engineering disciplines under a single delivery framework, from detailed design through construction support.

The value of the Hatch relationship to Atlas is reputational as much as technical. A tier-one engineering firm putting its name on the project would signal to lenders and equity investors that the technical foundation is being built to institutional standards. For a junior developer with limited internal engineering depth, this matters.

Continental Conveyor, Material Handling, IPD Partner (October 29, 2025)

This is the most operationally specific of the four. Continental would design and supply the complete material handling conveyor package: 28 belt conveyors and head chutes spanning underground crushing, surface storage, the 2-kilometer overland conveyor to Turf Point, and the port loading systems. The equipment would be supported by Continental’s Canadian facilities in Thetford Mines, Quebec, and Napanee, Ontario.

The conveyor system is a critical-path item for the project (without it, neither the mine-to-surface flow nor the surface-to-port flow exists), so lining up a credible Canadian supplier early is operationally sensible.

The four MOUs are straightforward:

• Scotwood Industries: Packaged retail offtake + 50/50 Canadian JV. 1.25-1.5 Mtpa. No vendor financing.

• Sandvik Mining: Underground mobile mining equipment, technology, and services (commercial scope up to C$132M). Vendor financing contemplated.

• Hatch Ltd.: Lead Engineering Partner and IPD Partner. No vendor financing.

• Continental Conveyor: 28 belt conveyors, mine-to-port. No vendor financing disclosed.

Reading the four MOUs together leads to three remarks.

First, Atlas has assembled four tier-one counterparties, and Sandvik alone could address up to approximately 22% of the C$589.1 million initial capital cost if signed. This is important for a reason that sits at the core of this thesis: Atlas is, in my view, clearly undervalued today. As long as that discount to intrinsic value persists, minimizing dilution is directly aligned with shareholder interests. Binding contracts are what we want, but at the right time. For now, every MOU signed with a credible partner is a small incremental positive.

Second, the pace tells a story. The MOUs accelerated once the permits came through. Early Works commenced in February 2026 and will likely run at least through year-end. That gives Atlas time to secure additional MOUs and to begin converting project-delivery agreements (Hatch and Continental) into binding contracts.

Third, the four MOUs cover a range of construction-side workstreams, but on the production side, coverage is thin. Only Scotwood addresses offtake, targeting 31-37.5% of planned volume. At least part of the remaining 62.5-69% is actively in play: Peterson confirmed that discussions are underway with Scotwood to expand the original volume and with other retail offtake partners to cover additional production. We are still four years from first production, so there is runway. For the bulk side, finding buyers is the easy part. The market already imports because North American mines can’t produce enough. If Atlas shows up with a year-round source three days from Boston, the sale will probably write itself. The real prize would be a binding retail offtake, even on a limited volume: it would mean Atlas has locked in a channel where every tonne ships at margins not reflected in the current NPV.

Subscribe now

05. Management and Alignment

In junior mining, a perfect deposit mismanaged becomes another orphan on the Lassonde curve. Atlas’s rock looks exceptional. Whether the team can match it matters more than any number in this deep dive.

I look at management in three steps: who the people are, what aligns them with shareholders, and what they have already delivered.

The Operators

Look at the Atlas team and you see two generations of mining executives stitched together on purpose: the older generation built the company, the younger generation is going to build the mine. Knowing where one hands off to the other tells you more about Atlas than any single resume.

Patrick Laracy, The Founder Who Stayed

Start with Laracy, because everything at Atlas eventually routes back to him. He is the founder of Vulcan Minerals (1995), the founder of what became Atlas Salt (2011), and the current Chairman of Atlas Salt and longtime head of Vulcan Minerals. He holds an LL.B. from the University of Calgary and a B.Sc. Hons. in geology from Memorial University. It is a rare combination of lawyer and P.Geo in one person, which means he can speak both the technical and legal languages of the project without translation. Over thirty years, he has leveraged more than C$100M of high-risk exploration spending into actual deposits in the ground, most of it in Newfoundland.

The job title understates what he does. He’s a node in the province’s mining establishment: Vice-Chairman of Mining Industry NL, director of the Newfoundland & Labrador Chamber of Mineral Resources, past president of the NL Explorationists Association. When the province thinks about mining, it thinks about people like him. When it thinks about rock salt specifically, there are not many people ahead of him in the queue.

Remember the EA registration that cleared the provincial process in seven weeks. The file was clean, yes. But a clean file submitted by a man who has been part of Newfoundland’s mining regulatory conversation for three decades was always more likely to get a fair hearing, fast. Those seven weeks are also a byproduct of Laracy’s network.

Rowland Howe, The Man Who Knows the Competition

Howe is the single most overlooked name on the deck, and in my opinion, one I think Atlas should put front and center. The investor presentation lists him as “Director” and “Salt industry veteran.” That word is doing a lot of work.

From 1995 to 2011, Rowland Howe ran Goderich (yes, the same mine I’ve been referencing since the first section). He inherited it, scaled it, and took it from 3.5 Mtpa to a record 7.5 Mtpa, making it the largest underground salt mine in the world.74 He stayed another six years in strategic engineering and project roles before retiring from Compass in 2016.75

In 2021, Laracy brought him out of retirement to serve as President of Atlas. Howe led the feasibility work through its critical years, then stepped back as the project crossed into execution. He is now a director, still close to the project, and still serves as President of the Goderich Port Management Corporation.

This is the person who knows Goderich’s cost structure, its labor dynamics, its vessel logistics, its failure modes, and its customer relationships from the inside, over more than two decades. For a would-be competitor to Goderich to recruit him is an information advantage that is difficult to quantify. But let me try anyway: a man who spent sixteen years running the industry’s flagship mine does not come out of retirement to sit on a board for a project he thinks is mediocre.

There’s an episode in the middle of that history worth naming. Between Howe stepping back and Peterson stepping in, Atlas had another CEO: Rick LaBelle, appointed in August 2023, departed in March 2025 by “mutual agreement,” after nineteen months. LaBelle came from Dumas Mining, where he had been President & CEO from 2017 to 2022; the same Dumas where Andrew Smith, Atlas’s current Project Director, had run the PMO. Smith arrived during that period and stayed when LaBelle left.

As usual with these exits, the full story never made it into the press release. You can read the departure positively or negatively; I won’t speculate. What matters for the thesis is the observable fact: the company has already survived one CEO transition without losing its operating bench. That helps bound the succession risk around Peterson. Speaking of which.

Nolan Peterson, The Executor

The ideal CEO for Atlas at this stage would be someone who has already built and brought a similar salt mine in North America into production. That person probably does not exist, because no comparable salt mine has been built on the continent in decades. The next best thing is someone who has crossed the feasibility-to-production bridge on other mines, understands both the engineering and the financing, and knows what goes wrong in the middle.

Peterson arrived as CEO and director in June 2025. MBA, CFA, P.Eng., PMP, twenty years across finance, operations, and mine development. On paper, the boxes are checked. Now in practice.

Peterson spent seven years at New Gold, straddling the line between engineering and corporate finance. He started on the technical side as Project Engineer on New Afton, a gold-copper underground mine in British Columbia that he helped construct and commission into production. From there, he helped advance both the Rainy River and Blackwater projects through their development stages. Then he crossed to the other side of the house: he joined New Gold’s finance group and led financial planning and analysis for the company, running the numbers behind the same kind of projects he had helped build.76

After New Gold, he joined TMAC Resources in senior finance management, working on the Hope Bay gold project in Nunavut until Agnico Eagle acquired it. Then he took the CEO seat at World Copper in April 2021, a PEA-stage77 copper developer with assets in Chile and Arizona. He grew the story, advanced the technical work, then resigned in November 2023 to “pursue other opportunities”. Nineteen months later, he surfaced at Atlas.

Since arriving in June 2025, Peterson has cleared the year-one checklist fast: the UFS delivered, three construction-side MOUs signed or expanded (Hatch, Continental, expanded Sandvik), C$8.7M raised, OTCQX uplisting secured (higher-tier U.S. OTC market, broader institutional visibility and liquidity for U.S.-based investors), and Early Works physically underway on site. No binding project-financing, offtake, or major construction contracts yet. The next milestones to watch are interim funding for early works, then the definitive project-financing package.

The Construction Bench

Below Peterson sits a team assembled over the last two years, with backgrounds that say one thing: we are here to build.

Robert Booth (VP Engineering & Construction): C$1.5B+ in mine builds at Newmont and Hudbay. Andrew Smith (Project Director & GM): C$500M+ in international mine-building contracts at Dumas. Jeffrey Kilborn (CFO): twenty years in mining finance, former Director & CFO of Canadian Gold Corp. And at the board level, Bob Kelly, appointed to the board in March 2025: 40 years in senior mining roles, construction-management work at Voisey’s Bay Nickel, including the hydromet demonstration plant, GM of Teck’s Duck Pond Mine in central Newfoundland, then VP responsible for health and safety across Teck’s operating sites, projects, exploration activities, and office locations. Kelly is the one who has actually run a producing mine in this province, and he’s a past President of the Canadian Institute of Mining’s Newfoundland Branch.

There is no exploration geologist running the operating team anymore. The people with those skills have done their job. What is left is a builder’s team staffed for a builder’s problem, even if the stock still trades like an exploration-stage one.

The Cap Table

The table below gives the starting denominator: 110.7 million basic shares, about 119.6 million fully diluted shares. The old warrant stack is gone; 2.85 million warrants at C$2.40 expired in 2025. What remains is a small broker-warrant line and a normal stock-comp pool for a junior developer entering the expensive part of the curve.

The existing dilution overhang is visible, but largely manageable. Options, RSUs, PSUs, and DSUs represent approximately 7.5% of the basic share count. Add broker warrants, and you get about 8%. This is almost nothing compared to the dilution risk during the construction financing phase.

On February 12, 2026, directors exercised 2.6 million options, and Atlas granted 3.15 million new options at C$0.98, including 1.6 million to directors according to the 2025 MD&A. Yes, the old options became shares, Atlas received cash, and the option pool increased by only 550,000 net. But this is basic retention, the company disclosed no link to project financing, construction start, first salt, or CapEx discipline. That is not catastrophic, the real incentive is still share ownership. But as a shareholder, I cannot see it as good compensation design, especially when the CEO still owns relatively little stock and the company is about to ask the market for construction capital.

Vulcan Minerals remains the anchor shareholder, with about 27.0% of Atlas at year-end 2025. Vulcan is led by Patrick Laracy and remains the main vehicle through which his influence is exercised. Laracy also holds approximately 5.3 million Atlas shares personally and through Triassic Properties. Put together, Vulcan’s stake and Laracy’s direct/related Atlas ownership represent around one-third of Atlas on a basic-share basis. Atlas has public shareholders, institutions, a board, and a new CEO, but Laracy is still the gravity in the room.

The rest of the board adds another layer of alignment. Edison holds about 1.3 million shares through Dollard Investments. Noel holds about 1.1 million. Howe owns 150,000 shares, plus his incentive package from his time as President. Peterson’s starting equity position was small, but he has bought approximately 50,000 shares in the market in 2026. Also worth noting: in the SEDI window I reviewed, over the last twelve months, I saw only insider purchases filed.

The structure is aligned, but still not entirely clean. Three of six directors (Laracy, Edison, and Noel) also sit on the Vulcan side of the table. That overlap is natural: Atlas came out of Vulcan, and the mineral rights, royalty, relationships, and institutional memory all run through the same corridor. But when the board votes on anything touching Vulcan’s interests, half the table has a seat on both sides.

The counterweight is Howe, Peterson, and Kelly. Howe knows the salt industry better than anyone else in the room, Peterson is the execution CEO, and Kelly brings Newfoundland operating credibility. My read is simple: Laracy controls Atlas in the practical sense. Through Vulcan, through his personal stake, through the board’s history, and through thirty years of weight in Newfoundland’s mining ecosystem. That is powerful when your interests are aligned with his. Less so when they are not. There is one specific case where this could matter: a potential M&A scenario. We will come back to it later.

The October 2025 financing added one powerful signal: Atlas brought in its first strategic investor. Management has not named the party, only described it as an investor whose interest in Atlas and Great Atlantic “aligns with its long-term strategic objectives.” I would not turn that into a takeover thesis, but someone with strategic interest wrote a cheque before project financing was in place. My take is that the strategic investor is more likely someone with a direct interest in securing future tonnes than a mine operator trying to fold Great Atlantic into its existing logistics network. Geography matters here, again. For most incumbent producers, Atlas is not an obvious plug-and-play asset, it is a new Atlantic supply point with its own route to market. The “obvious” candidate is Scotwood, given the MOU already signed with Atlas, but its strategic logic of owning equity is less obvious. The second bucket would be an East Coast port operator/distributor. Eastern Salt is the default name that comes to mind, although there are other possible candidates. Larger salt groups such as Kissner or Stone Canyon are also possible, especially if they want an early position ahead of a future M&A scenario, but a C$8.7 million LIFE financing would be a very small move for them. Possible, just not my base case at the moment.

Atlas has the right people, and enough ownership for alignment to matter. But Vulcan sits everywhere. That is acceptable while the goal is simple: get the mine built. It becomes less comfortable when Vulcan’s interests diverge from those of minority shareholders.

There is one case where this matters. Here it is.

The Royalty

In 2012, when Laracy spun the salt and industrial minerals business out of Vulcan into what was then called Red Moon Potash (later renamed Atlas Salt), the mineral licences were transferred from Vulcan to the new entity. The consideration came in two forms: common shares in the new company and a 3% Net Production Royalty (NPR) payable on production from the mineral licences, calculated on Atlas’s gross revenue less port charges, processing costs, and the NL Mining Tax. That was the price Atlas paid to get its own assets.

SLR has already modeled this royalty in the C$920M NPV, so it does not change the headline number. But it changes who receives each dollar, and the distribution of value is more lopsided than first appears.

Back-of-the-envelope, using the royalty formula in the UFS: the royalty base is approximately C$8.86B over the initial 24.25-year mine plan, or about C$365M per operating year on average. At 3%, that gives Vulcan about C$266M of undiscounted royalty cash flow before Vulcan-level tax. Discounted at 8% from the UFS effective date, with the four-year pre-production period and the UFS production ramp, the stream is worth about C$80M. That is just under 9% of Atlas’s C$920M after-tax project NPV.

But the royalty may have a ceiling, and that ceiling may be Scotwood. If Atlas signs a definitive agreement and the JV operates as contemplated, Atlas would, under the UFS pricing assumption, sell its salt FOB Turf Point at around C$82 per tonne. The JV would package it, distribute it, and sell it at retail. After costs, profits would be split 50/50 between Atlas and Scotwood. Atlas’s share could potentially be classified as investment income from the joint venture, not as revenue from the sale of salt. If so, it could fall outside the royalty base, assuming the final royalty and JV documents do not contain a look-through mechanism. In other words, every tonne routed through Scotwood, or any similar structure, could leave more value with Atlas shareholders than a tonne sold directly into the market.

What Laracy’s Alignment Means for Shareholders

Most of the time, Vulcan’s 27% stake and Laracy’s direct and related 4.8% stake mean his interests are aligned with shareholders’ interests. When Atlas re-rates, his net worth rises in lockstep. But in three specific cases, the interests can diverge.

Financing decisions. Vulcan has been diluted from 29.79% in 2023 to 27.02% at year-end 2025, and will dilute further through construction equity. Every dollar of equity Atlas raises from someone other than Vulcan is a dollar of dilution to his largest asset. His economic incentive is therefore to raise at the highest possible price, as late as possible, and only what is needed.

Acquisition scenarios. If a salt major or private equity player tried to buy Atlas, Laracy has two interests to weigh. Atlas shareholders want the highest takeout price per share. Vulcan wants that too, but also wants to preserve the 3% NPR through any change of control. Laracy’s board seat means the royalty won’t be traded away in a rush. Whether that protects Vulcan at our expense depends on whether preserving the NPR compresses the takeout premium an acquirer is willing to pay.

Spin-offs. In 2022, Atlas spun off Triple Point Resources, which holds Fischell’s Brook salt dome and other mineral claims 15 km south of Great Atlantic. Howe is currently listed by Triple Point as a Special Advisor. The spin-off was marketed as a bonus for Atlas shareholders, and maybe it is. The mechanics, however, follow a recognizable pattern: assets that once sat inside Atlas now sit in a separate entity, with its own cap table, its own management and advisers (Howe), and its own future financings that Atlas shareholders will not automatically participate in. Atlas retains a 14.4% equity stake in Triple Point, but no board representation. The “main” other assets still inside Atlas are the Ace Gypsum Project (historical gypsum mines at Flat Bay, 3 km southwest of Great Atlantic) and the Black Bay Nepheline Syenite Property (a surface occurrence of nepheline syenite in southern Labrador, with historical metallurgical work and minimal 2025 spend). Neither is material to the thesis.

The Net Read

This ownership-and-royalty structure is a double-edged sword, except one edge is razor-sharp and the other is dull.

The sharp edge cuts in shareholders’ favor. Laracy holds approximately a third of Atlas through Vulcan and personal/related holdings. Every re-rating flows directly into his balance sheet. Insider and related-party ownership exceeds 40%. Howe has his salt-industry reputation on the line in a way a detached board director does not. These people are highly incentivized to make Atlas work.

The dull edge comes after Atlas succeeds. The NPR is a life-of-licence claim on value, already netted out of the C$920M NPV and worth about 9% of pre-royalty modeled project value. The board has a Vulcan tilt (three of six directors sit on both boards), so a change-of-control scenario could be negotiated around the royalty as much as around equity.

This is a common setup in junior mining. You get a management team and a board that are genuinely strong for a company this size, and the structure is part of how they get paid for it: the royalty, the overlap, and the concentrated ownership. That’s the price of having Laracy’s network, Howe’s salt expertise, and Peterson’s execution experience.

06. A Quick Look at the Financial Statements

Before opening the statements, one accounting point needs to be killed immediately.

Great Atlantic sits on Atlas’s balance sheet inside C$16.7 million of mineral exploration and evaluation assets. The UFS gives the project a C$920 million after-tax NPV at 8%. Those two numbers look incompatible only if you ask accounting to do valuation work, which you should not do for a junior miner.

The C$16.7 million is mostly the accumulated cost trail of getting the asset portfolio to this point (licenses, geology, feasibility work, engineering, etc.). It generally goes up when Atlas spends more money on the assets, but not when the salt market tightens or when the UFS points to a stronger economic case.

So the financial statements only tell us how clean the balance sheet is, where the cash went, and how much runway Atlas has before the real financing decision arrives.

The Wrong Place to Look for Value

The income statement is the least useful page in Atlas’s accounts.

There is no producing mine yet, no operating revenue, and therefore no margin to analyze. Atlas recorded a net loss of C$3.71 million in 2025, almost identical to the C$3.68 million loss reported in 2024. It may look stable, but it is mostly irrelevant. At this stage, the P&L mostly measures the corporate cost of keeping a listed developer alive while the project moves from feasibility study toward production.

A few items are still worth pulling out:

• Share-based compensation fell sharply, but that was mainly an equity-award cancellation/reversal effect tied to former executives and a former director.

• The C$776,644 gain on Triple Point came from Atlas losing significant influence after Triple Point’s financing and no longer accounting for it as an “associate” under the equity method, which created an accounting gain without bringing any cash into the business.

• On the cost side, the “office, consulting fees, and other” line increased, but the MD&A says nearly C$800,000 of that increase related to financing expenses (the cost of trying to assemble the next layer of capital).

• Management and subcontractor fees rose as well, largely because Atlas paid a termination payment to a former executive.

So yes, Atlas lost money again in 2025. That is the burden of every pre-revenue developer. The balance sheet is already a little more interesting.

Balance Sheet Snapshot

At year-end 2025, Atlas had C$26.3 million of total assets, C$6.0 million of cash, C$6.6 million of positive working capital, and C$1.8 million of total liabilities. There is no project debt, no construction facility, and no off-balance-sheet arrangements. The liabilities are mostly trade payables, a small Business Development Bank of Canada (BDC) loan, a light-duty vehicle loan, a conditionally repayable Atlantic Canada Opportunities Agency (ACOA) contribution whose repayment begins after commercial production, and an asset retirement obligation tied to Ace Gypsum. In other words, nothing that changes the Great Atlantic financing equation.

The company is therefore still simple: Great Atlantic + cash + a few assets around the edges. Those edges are worth identifying, because they can otherwise make Atlas look more diversified than it really is.

The old Fischell’s Brook Salt Dome property was spun out into Triple Point in 2022. Atlas still owns a residual stake, valued at C$1.37 million at year-end, but its ownership fell to about 14.4% after Triple Point’s 2025 financing, and Atlas no longer accounts for it as an associate. Strategically, that asset has left the building.

The gypsum assets are small. Atlas owns the Flat Bay / Ace gypsum assets near St. George’s, including three mining leases, cumulative gypsum sales/proceeds of C$2.1 million, and Atlas has an agreement with Turf Point Resources (the same port group involved in the future rock salt logistics) under which gypsum or anhydrite may be mined and exported in 2026, with Atlas receiving a trivial per-tonne royalty.

Black Bay Nepheline is even further out. Atlas spent C$15,078 on the project in 2025, mentions “encouraging” historical metallurgy, and intends to complete additional work in 2026. That is option value in the strictest sense: present in the filings but absent from the investment case. Bay St. George General is smaller still, with C$48,239 of costs incurred to date and nothing spent in 2025.

So far, none of this does much for the thesis. The same cannot be said of the cash analysis.

The Cash Question

Atlas raised money in 2025 and still ended the year with less cash than it started with. Welcome to the junior mining world.

Cash went from C$8.0 million to C$6.0 million. During the same year, the company raised C$8.6 million in cash through an equity financing, paid C$0.9 million of issuance costs, recorded C$0.6 million of conditionally repayable ACOA financing on the cash-flow statement, and still ended the year with C$2.1 million less cash. The financing basically covered the year, but the split gives us the useful information.

The investing line is the cleanest one. Atlas spent C$4.7 million in cash on mineral exploration and evaluation assets, and the carrying value of those assets increased from C$11.8 million to C$16.7 million. That is the part of the burn shareholders should want to see: money leaving the bank account and showing up in the project. I have seen too many junior miners slowly burn cash into another corporate appendix.

The operating line is less flattering. C$5.3 million of operating cash burn looks high for a company with no revenue. But it is also a dirty number. It includes: the cost of being public, a management transition, preparing a financing package, and cleaning up working capital. Prepaid expenses increased by C$0.46 million, trade payables and accrued liabilities fell by C$0.77 million, and the income statement itself contains non-cash noise from Triple Point and share-based compensation.

So saying “Atlas just burned C$10 million” is lazy at best. In 2025, the burn split into two buckets. Approximately half went into the asset, while the other half paid for the corporate and financing machinery required to move that asset toward construction.

That is what I want to see as a shareholder. A developer can spend very little and look disciplined while the project goes nowhere. Atlas spent real money. Some of it advanced Great Atlantic. Some of it was the toll paid to reach the next capital event.

The Next Step

C$6 million of cash at year-end left Atlas enough room to keep the file moving, but it does not build a mine.

This is why the going-concern language matters, but only if it is read properly. Management prepared the accounts on a going-concern basis, and the auditor did not include a material-uncertainty paragraph. That does not underwrite the future financing of the mine, or even the future salt tonnes themselves. It only says: Atlas has no revenue, burns cash, and had enough liquidity at year-end to keep operating for now. A perfect example of the difference between accounting reality and (economic) reality.

Same with the impairment work. MNP (the auditor) treated the assessment of impairment indicators for the C$16.7 million of mineral exploration and evaluation assets as a key audit matter, and no impairment indicators were identified. Fine. It removes a potential accounting red flag, one that was already obvious to anyone who did the work. But again, it does not bring the C$920 million after-tax NPV8 onto the balance sheet.

One post-year-end event is worth noting. On February 12, 2026, directors exercised 2.6 million stock options, adding cash to the balance sheet. The proceeds were not disclosed, but the year-end option table gives us a decent range. Based on the year-end option table, applying a C$0.79-0.89 weighted-average exercise-price range to 2.6 million exercised options would imply approximately C$2.0-2.3 million of cash, with the upper end probably the better proxy since the exercised options had to be exercisable. For simplicity, I assume the option-exercise proceeds have roughly offset the cash spent since year-end. That leaves Atlas with something close to C$6 million of cash in my current working estimate.

Audited statements make development stories look bad by nature. That is normal, that is their job, and probably for the better. The most positive thing they tell us is that Atlas is clean enough to keep moving for now, but they tell us nothing about how the mine gets financed.

That is the next question.

07. The Financing

No revenue, approximately C$6 million in cash, and hundreds of millions to raise or arrange over four years before the first tonne ships.

How Atlas closes that gap is the single most important question between here and production. It will shape what the existing shareholder actually owns in 2030.

Start with the cash need.

What Atlas needs to fund

Initial CapEx is C$589 million. That’s what SLR engineered and what every investor deck shows. But it is not the full amount of cash Atlas will have to fund between today and the first tonne is sold. Two categories sit at least partially outside the CapEx envelope:

1. Corporate overhead: C$12-16M. This is the cost of running a listed developer for another four years: corporate salaries, audit, listing fees, board, investor relations, etc. Atlas’s 2025 operating cash burn was C$5.3M, but that is too dirty to use as a clean run-rate: it includes management transition costs, financing expenses, and working-capital movements. A normalized figure closer to C$3-4M per year is more defensible. SLR covers approximately C$4.8M of this inside the CapEx envelope as “Corporate Costs,” but I will leave that buffer untouched. It can absorb slippage, extra advisory work, or the usual small costs that never look material until they aggregate. Four years of normalized overhead lands at C$12-16M outside the UFS capital estimate.

2. Working capital: C$10-20M. SLR explicitly excludes this from the CapEx estimate, and the UFS cash-flow schedule does not show a working-capital build either. It is the cash tied up in accounts receivable from customers who pay 30-45 days after delivery, finished product inventory, and operating consumables, net of accounts payable to suppliers. For a mine shipping 1.6 Mt in Year 1, or approximately C$131M of revenue at the Q3 2025 base price, C$15M as a central estimate is a standard order of magnitude. That C$15M should be recoverable at mine closure, but it has to be there as operations start to ramp.

Total funding need before financing costs: range of C$611-C$625M range. I use C$625M for safety.

Financing costs come on top of that, but they’re mechanically linked to how the financing gets structured. To size those, I need to lay out the debt architecture itself.

The Debt Architecture

The rock salt market and Great Atlantic’s features partly simplify the problem:

• 24 years of relatively predictable cash flows. The price of salt rises gently year over year on average, in contrast to gold or even copper, whose prices fluctuate far more.

• An essential commodity with contractable demand.

• Low, stable OpEx. C$28/t modeled LOM operating cost, with no metallurgical processing plant, no tailings, and a simple product.

These three features can unlock capital sources many junior miners can’t touch: infrastructure banks, export credit agencies, and even sovereign wealth funds. These lenders can look at Atlas as a quasi-infrastructure asset with contractable cash flows. Atlas will be financed. The question is how and at what price.

To that end, Atlas appointed Endeavour Financial (a London-based natural-resources financial advisor that has raised over US$4 billion in junior mining debt since 2003) as project finance advisor in December 2024. Endeavour has assembled the lender package: marketability of the salt, engineering, environmental and social diligence, and the UFS model. It is a single due diligence package sent to potential lenders, like a mortgage broker distributing one complete file to the market.78 Since then, Atlas has obtained two Letters of Interest (LOI) in hand from potential debt financiers, including export credit agencies. Nothing new here, the playbook already exists. Nouveau Monde Graphite secured a US$335M senior project-debt commitment letter;79 Canada Nickel, Torngat, Defense Metals, etc., show the same direction of travel80: Canadian resource developers are increasingly using EDC, CIB, ECAs, and strategic/public lenders to bridge the gap between feasibility and construction.

Nothing new, but nothing simple. Debt is a stack of layers, and each layer sits at a different level of risk and pays a different coupon.

Here’s how this capital stack could be structured, according to the CEO:

“60% of our financing [initial CapEx] needs from senior secured lending [...] When we figure that out, then we layer in the subordinate debt, 10 to 20% potentially [...] Also, the possibility of government support, and then of course the traditional equity angle.”

Three sentences that hide a lot of machinery. Let me unpack each layer.

1. Senior secured debt: about 60% of initial CapEx. First lien on the mine’s assets. The lenders would be specialized commercial lenders: they bank airports, power plants, ports, etc., and would be looking at Atlas the same way. Rates typically fall in the 6-8% range.81 Endeavour has been arranging this since December 2024. A common structure in these facilities is to roll the construction-period interest into the loan itself: instead of paying interest in cash during the four years when the mine generates no revenue, Atlas borrows the interest on top of the principal. Importantly, this does not necessarily increase the upfront equity cheque Atlas has to raise. The interest tranche can be structured inside the senior facility itself. It just enlarges the principal Atlas will repay once production begins.

2. Subordinate debt: 10-20% of the package. Second-lien or structurally junior capital, so it gets paid after senior lenders in a liquidation, which means it costs more. A lot more, rates are typically 10-15%. Debt is often structured as PIK (Payment-In-Kind) during construction, meaning interest is added to the principal rather than being paid in cash (with no drain on Atlas’s treasury until production). The natural providers would be resource credit funds that have seen every configuration of this structure: Orion Mine Finance, Nebari, Sprott Resource Lending.

3. Vendor financing: C$132M contemplated under the Sandvik MOU (potentially more). Sandvik signed an expanded MOU in February 2026, covering equipment and services through construction and ramp-up. The financing would likely be tied to Sandvik-supplied equipment, but the exact terms are not disclosed. This is the most clearly defined tranche of Atlas’s stack, but still a non-binding MOU.

Regarding the government support Peterson mentioned, I haven’t found anything concrete. Since I can’t quantify either the probability or the funding itself, I’m setting it aside. But I would be surprised if discussions were not already underway. If the government does end up providing construction financing, it would be really good news (as long as it is not the only lender in the room): interest rates would likely be low, and covenants could be lighter.82

Finally, we have all the axioms we need. We can now do the math.

The Cost of Financing

There’s no free lunch, so you have to pay the people who help you raise the debt and the equity. In the model, I assume senior debt fees of 1-3% of the facility size, subordinated-debt fees of 2.5-7%, equity issuance fees of 5-7%, and separate advisory/legal/closing costs of 1.5-3%. The exact figure depends on the equity residual, which depends on the total funds needed, which depends on the fees. But the final reality is unchanged: financing costs will be the largest cost item outside the UFS, by far. Every assumption and calculation is detailed in the DCF at the end and reflected in the valuation section.

There’s also a structural constraint: senior lenders control what can sit below them. They impose a set of financial covenants that govern how the rest of the financing can be structured:

• Debt Service Coverage Ratio (DSCR). The ratio of cash flow available for debt service to scheduled debt service, commonly around 1.3x to 1.5x for contracted or lower-risk projects,83 with higher requirements when merchant or completion risk is higher. Depending on the facility, the test may apply to senior debt service only or to total debt service. Either way, every additional dollar of cash-pay subordinated debt tightens the equation.

• Loan Life Coverage Ratio (LLCR). The NPV of future cash flows available for debt service over the loan life, divided by the relevant debt balance. Typically required at 1.5x to 2.0x.84 It is a longer-horizon coverage test used alongside DSCR. Same logic: more debt in the stack shrinks this ratio.

• Gearing ratio. The ratio of total debt to total project capital. Usually capped at 70-80% for bankable infrastructure-like projects.85 Above that, the senior lenders can make closing very difficult.

These covenants are contractual: if Atlas tries to add subordinated debt that would push the DSCR below the threshold or the gearing above the cap, the senior lenders can use the intercreditor framework to block or condition the additional layer.

Peterson has said that resource funds have indicated they could push total debt to 70-80% via subordinate layers (vs. 60% senior), implying 10-20% of initial CapEx subordinate on top of the senior facility. The high end of that range is probably out of reach: the senior covenants (DSCR, LLCR, gearing) tighten as subordinate debt stacks up, and pushing to 20% would likely breach one of them. In my opinion, the realistic range is closer to 10-15% of the initial CapEx.

Fortunately for shareholders, Sandvik is different. Vendor financing is usually secured against the equipment itself, rather than sitting as another pure project-finance layer against the same collateral package. Depending on the structure, senior lenders may treat it as permitted equipment financing rather than standard project debt, but they will still care about the rest (repayment burden, liens, intercreditor terms, and cash-flow impact). Senior lenders may prefer vendor financing: it reduces the CapEx they have to cover directly, and it aligns the equipment supplier’s interests with the project.

I am going to spoil the valuation section a little, but it is necessary. In my model, the net construction-equity requirement ranges from C$125M to C$383M. I know the spread is wide, but I’m not going to pretend I can forecast the middle of that range with precision. The point is that the dilution question and the share price at which it happens are critical. Where the needle lands depends on Chairman Laracy’s network, Peterson’s execution, and how the three financing levers above resolve. None of that is in the individual shareholder’s hands. But the signals to watch are.

What to Watch

Here are five signals, in the order cash pressure forces them:

1. A bridge financing in 2026 (a temporary financing used to fund the company until the main financing package closes). In the UFS schedule, the first pre-production year carries C$64.7M of capital, followed by C$100.0M in the second year. Against approximately C$6M of current cash and a corporate burn already running in the hundreds of thousands per month, the conclusion is the same: Atlas will likely need a 2026 bridge before the senior debt package closes. Peterson said so openly at the PDAC (Prospectors & Developers Association of Canada) convention: “[parties] are coming to us earlier… is there an opportunity to accelerate.” Atlas is going to raise again before the senior debt closes. To stay on schedule, this bridge financing should close in the next few months.
   
   Three things to track on the next tranche: the instrument (equity/debt/convertible), the subscriber (an “unnamed strategic” participated in October 2025), and above all the price if it is equity. Every incremental cent above C$1 is capital saved on the construction-equity stack. But the company still looks clearly undervalued to me, and debt, even short-term debt used as a bridge before senior financing, would be preferable to another dilutive equity raise.

2. A senior project-debt term sheet. With two Letters of Interest from potential debt financiers already in hand, the next step is a term sheet. That’s Peterson’s stated 2026 milestone. Its signing moves Atlas from paper project to bankable project, historically one of the moments that can trigger the sharpest re-rating on junior developers. The next update is expected by summer 2026.

3. A binding Turf Point access/use agreement, or clear disclosure that lenders have accepted execution of that agreement as a condition precedent to first drawdown. Turf Point is third-party-owned and embedded in the mine plan; a credible senior financing package should either have this agreement signed or explicitly define its execution as a closing/drawdown condition. If a senior debt package closes on schedule, it would be strong evidence that lenders found no blocking issue in the port relationship.

4. The Sandvik MOU converts to binding. Conversion could move a large and defined equipment package out of the equity-funded portion of the capital stack and tell senior lenders that Sandvik has completed enough diligence to put binding support behind the project. This is most likely to happen alongside or after the senior debt term sheet, not before.

5. Scotwood, or another major offtake partner, signs a binding agreement. An industrial counterparty is unlikely to sign a binding agreement before seeing a term sheet on the lender side, so this signal is correlated with the debt signal above. Note that the thesis survives its absence; demand for this salt exists whether Scotwood signs or not. What a binding contract adds is timing: confirmation that the commercial chain solidified fast enough to support financing and the construction decision.

Five signals with one question underneath each: has the sequence started working, or has it stalled? The shareholder who tracks them has enough to decide whether to hold, add, or exit. What they do not answer is what Atlas is worth if it all goes through. That’s the next question.

08. The NPV: Decoding the C$920M Headline

C$920 million, after tax, discounted at 8%.

That’s the number every Atlas investor deck leads with. It is also the number I’ve spent the most time looking at, because it is the one the entire project valuation rests on.

An NPV is assumptions all the way down: salt price, escalation, CapEx, sustaining capital, discount rate, tax treatment, royalty mechanics. Move any one of them, the output moves. Move several at once, the output can double or be cut in half. SLR picked its assumptions knowing every one of them would be challenged by lenders spending millions on due diligence before wiring a cent. What came out the other end is a number built to survive that review.

Built to survive is not the same as built to be accurate. The shareholder’s job is to challenge each assumption and ask where SLR was conservative, where SLR was generous, and where the real number might sit. Every previous section has been setting up what we need for the sections that follow.

The Revenue Side

Revenue in the model rests on three assumptions: an initial price, a volume schedule, and an escalation rule.

The Price

The base price is C$81.67 per tonne, FOB Turf Point. SLR built it as a weighted average from Q2 2025 pricing estimates across the four destination markets (U.S. East Coast, Québec, Maritimes, Spot & Commercial), weighted by the planned 56/15.5/8.5/20 allocation. A third-party estimate, anchored in observable prices, designed to survive lender due diligence.

It is almost certainly the floor.

The model applies one price uniformly, regardless of commercial structure (CIF, DAP, FOB). SLR first builds the C$81.67/t from the planned market mix (56% U.S. East Coast, 15.5% Québec, 8.5% Maritimes, and 20% Spot Sales & Commercial) using port-level pricing data. Once that blend is set, the NPV does not carry separate realized-price lines for a DOT tonne, a spot/commercial tonne, or a packaged/retail-related tonne. The salt market does not work that way. As set out in the Business Model section, the channel a tonne travels through can determine what it earns by a factor of two to six. SLR anchored its average on the lowest-margin channel: bulk contracts to public buyers. Any dollar Atlas captures above that would sit outside the C$920M: through Scotwood’s retail JV, through the Québec spot market during Seaway closure, through commercial tonnes sold to contractors pushed to the back of the queue.

Management knows this. Asked directly at PDAC 2026 how the C$81.67 compared to a spot market where parts of Ontario were paying C$300, Nolan Peterson answered: “our base case for the general bulk product market also probably does not reflect the pricing increases that we are going to be seeing next year... there’s a lot of upside that’s not built into our model right now.”

There is one more issue with the C$81.67. The Ontario winter of 2025-2026 should influence the spring contract cycle on which Atlas’s base case rests. Peterson expects a 5-10% uplift beyond normal escalation in the next contract round. None of it is in the model either.

For all these reasons, SLR’s C$81.67 looks closer to a bear case than a base case.

The Volume Schedule

The volume schedule follows a three-year ramp: 1.6 Mtpa in Year 1, 2.6 Mtpa in Year 2, 3.7 Mtpa in Year 3, then 4.0 Mtpa steady state from Year 4 through Year 22, tapering over the final 2.25 years. These numbers are the model’s most untestable assumption: I have not found a clean salt-mine precedent for Sandvik MB670 continuous miners paired with a mostly battery-electric underground fleet at this scale. There is no clean comparable operation to benchmark the ramp against. The only thing supporting the curve is the technical work behind the UFS, Sandvik’s proposed equipment package, and SLR’s sign-off.

Grade is held above 95% NaCl, priced uniformly as bulk road deicing salt. Some intervals of the deposit exceed 98% NaCl and may qualify for chemical or food-grade markets at materially higher per-tonne prices. Too many preliminary hypotheses sit between here and that first premium tonne, so SLR left that option unpriced, and so do I.

The Escalation

SLR escalates the C$81.67 base at 4% per year for five years, then 2% per year through the rest of the mine life, citing consistency with other North American rock salt technical reports. The structural backdrop laid out in Part 1 (supply frozen for twenty-five years, incumbents under operational and financial strain, demand becoming less predictable and more event-driven) sits well above that curve, and the historical record backs it up: U.S. rock salt PPI has compounded at 4% per year for four decades, meaning SLR is modelling the next 24 years at a blended rate meaningfully below a trend that has held through multiple cycles.

That’s fine by me. Feasibility studies are paid to produce defensible base cases, not to predict the future. What matters for the shareholder is the probability distribution of what actually happens over 24 years, and that distribution skews meaningfully above SLR’s curve.

The revenue line behind the C$920M rests on three layers of conservatism stacked on top of each other:

• a base price heavily anchored to the lowest-margin channel,

• an escalation curve below the 40-year trend,

• and no credit for any higher-grade product option,

each of which leaves room above the model for Atlas to earn more than SLR assumed.

The Cost Side

Again, three assumptions here: the capital to build, the capital to sustain, and the cost to operate.

The Build

Pre-production capital is C$589 million. SLR classifies the estimate as AACE Class 3 (a standard level of cost-estimate maturity for a feasibility-stage mining project): an accuracy range of -10% to +30%, which implies a practical estimate band of C$530M to C$766M. Against the sector, that range is narrow: McKinsey’s 202486 review of 80 global mining projects found that cost and schedule challenges affected 83% of recent major mining and metals projects, with CapEx overruns of more than 40% and schedule delays of 20% to 30%. Across its sample, mining projects averaged about 40% real cost overruns. A typical project would land meaningfully above that top bound.

But Great Atlantic is not a typical project. It is a shallow deposit, with no metallurgical processing plant (one of the most common sources of commissioning overruns), no tailings, no shaft, and a short flowsheet that ships salt almost as-mined.87 Each of those is a line item that can blow up budgets elsewhere and is materially reduced or absent here. And on top of that simpler base, SLR still assessed contingency line by line across the budget: extra money set aside upfront to absorb cost surprises before they become real budget overruns. Riskier lines like mine development and the boxcut (the open-cut excavation that accesses the decline) were assigned 25%. The average across all lines lands at 15.1%. That’s C$77.5M of buffer already baked into the C$589M. For Atlas to overrun the published budget, overruns would have to consume that cushion first, then exceed it.

Even if they did, the margin absorbs a lot of it. SLR’s sensitivity table puts a 35% CapEx overrun (relative to the C$589M, and therefore on top of the 15% contingency already baked into initial CapEx) at 14% NPV hit, dropping the headline from C$920M to C$794M. It is a threshold that kills most projects but one that Atlas walks through without really moving the undervaluation.

The same cushion that protects the budget also gives Atlas room to finance it intelligently. The clearest piece is Sandvik: C$132M of equipment and services, or 22% of the C$589M pre-production CapEx, with vendor-supported financing contemplated if the MOU converts. Hatch and Continental do not finance Atlas, at least not on disclosed terms, but they still matter to the financing file: engineering and conveyor systems move from anonymous budget lines into named scopes with tier-one counterparties. Again, none are binding yet. Converting Sandvik into definitive vendor financing, and Hatch and Continental into bankable execution packages, is Peterson’s job between now and Final Investment Decision (FID).

Sustaining Capital

Sustaining capital over the 24-year mine life totals C$609M in the economic model, more than pre-production capital itself in nominal dollars.88 The number covers: new mining levels opened as the mine ages, a fleet expansion from one continuous miner to six, and continuous refurbishment of underground and surface infrastructure.

This is the line item where I think SLR is most likely to have underestimated. Underground mines can come in around 30% above their sustaining forecasts, and far higher when the outliers are included.89 Applied mechanically to Atlas, that would push the actual sustaining bill somewhere between C$790M and more than a C$1B.

In a typical mine, a large part of that overrun often comes from categories Atlas largely avoids: metallurgical processing plant refurbishments, re-engineering as ore grades decline, compounding investments to drain water and stabilize ground in deep or under-lake operations.90 The product grade is modeled above 95% for the full mine life, there’s no metallurgical processing plant, the salt deposit is shallow and laterally continuous, and the deposit starts around 180 meters in relatively simple bedded-salt geology. What’s left is fleet expansion, new mining levels, and conveyor maintenance. Those are mechanical line items that track forecasts better than chemical or geological ones.

I’d expect Atlas’s actual sustaining to land 20-30% above plan. A conservative overrun by the sector’s standards, justified by a project profile that removes the lines where most mines lose control.

What saves the math is the discount. A 30% overrun on the C$609M nominal sustaining curve, spread across 24 years and discounted at 8%, translates into C$54M91 of after-tax present value. About 6% of the C$920M NPV (vs 30% nominal). The time value of money, so often an enemy in 24-year projects, is here an ally.

Operating Cost

The UFS operating cost averages C$28.17 per tonne shipped FOB Turf Point, with operating cost items escalated at 2% per year from the Q3 2025 base. The cost advantage is the assumed C$0.062/kWh industrial electricity rate. As laid out in the operational design section, the electric underground design makes that low-cost hydropower especially valuable. Layering sustaining capital on top brings sustaining-inclusive unit cost to C$34.90 per tonne. Against the C$81.67 un-escalated FOB base price, the sustaining-inclusive cash margin lands near C$47 per tonne (57% margin) before royalties and taxes.

The C$28.17 is realistic as a starting point. Where it could drift is mostly in the Mining line. At C$15.00 per tonne, it is the largest of the three buckets and the one most exposed to wage inflation in a region with competing labour demand. The number worth pressure-testing is the 2% annual escalation SLR applies to operating costs, including labour. Labour is consistently one of the largest components of underground mining operating cost in mechanized operations.92 If actual inflation runs closer to 3.25% per year (closer to recent Canadian mining wage trajectories), and labour costs rise accordingly, labour would add C$1.10 to the Mining cost on a discounted life-of-mine basis. At 4.5%, it would add about C$2.44.93 Probability-weighting those scenarios, labour would contribute about C$1.30 per tonne of drift to the Mining line,94 approximately 9% drift on that bucket and 4.6% on total OpEx. Using SLR’s published OpEx sensitivity, where a 17.5% OpEx increase reduces NPV by 8.6%, that drift translates to about C$22M off the C$920M base, approximately 2.4% of NPV.

Atlas’s site connected load starts at 10.7 MW and rises to 14.4 MW in Year 3 and 16.7 MW ultimately, per the UFS, which translates to approximately C$1.30 per tonne (about 5% of total OpEx). SLR escalates that line at 2% per year. The Newfoundland industrial tariff has trended faster than that recently, about 7.5% per year between 2019 and 2025,95 more than three times SLR’s assumption. That pace reflects two specific drivers: a one-time base rate reset following the 2017 General Rate Application, and the introduction in 2022 of a surcharge to start recovering Muskrat Falls costs from industrial customers, neither of which is structural to a 24-year horizon. But even if the rate kept compounding at 7.6% for the full mine life, the NPV only loses 5.5% (at an 8% discount rate), about C$51M. The risk is real, but its impact is modest.

Stack labour and electricity drift together and OpEx would likely land around C$31 per tonne, about C$73M off the C$920M NPV, about 8%, in this stress case.

If you think I am being too lenient with an 8% haircut to the NPV, wait until we get to valuation.

The Royalty and the Taxes

The project carries a 3% net production royalty payable to Vulcan Minerals on the Great Atlantic mineral licences, calculated in the UFS on gross revenue net of port charges, processing costs, and the NL Mining Tax. SLR has already deducted it from the C$920M. Without it, the NPV would land near C$1B: approximately a 9% transfer of pre-royalty project value from Atlas shareholders to Vulcan shareholders, locked in for the life of the mine. Who Vulcan is and why this royalty exists is covered in the governance section.

Newfoundland and Labrador’s Mining Tax is 15% of the mine operator’s net income, separate from federal and provincial corporate income taxes. The combined effective rate over the LOM appears to land at approximately 35-40% of pre-tax cash flow, in line with what other Canadian mining projects of this scale carry. SLR relied on Atlas and its advisers, including KPMG work, for the tax calculations, standard practice for a feasibility study, but worth flagging that the C$920M sits on someone else’s tax math.

One structural exposure worth naming. Atlas is a single-asset company in a single jurisdiction. Any change in NL Mining Tax rates, federal CCA classes for mining equipment, or the provincial royalty regime flows straight to the bottom line with no offset elsewhere. NL has not materially changed mining taxation in over a decade, and the province has every incentive to keep the regime competitive. But the exposure is concentrated in a way it wouldn’t be for a multi-asset major.

The Sensitivity Analysis

SLR runs four variables through its sensitivity table: salt price, operating cost, pre-production capital cost, and production losses. Here are the numbers.

Salt price is the dominant variable, and by a wide margin. A 20% adverse move in the LOM (life-of-mine) salt-price assumption wipes out 41% of the NPV. Every other variable is secondary. OpEx and CapEx matter, but are absorbed by the project’s wide gross margin while production losses are essentially immaterial. This is why the salt market takes 10,000 words.

This is also why Atlas is such a large opportunity. The company is not only priced unusually cheaply for its stage of development, but the most important assumption embedded in that valuation is also, in my view, materially underestimated. It is a growth setup embedded inside a value setup, created by the fact that few investors understand the salt market, including most mining investors.

One thing worth noting: the -20% case (C$65/t FOB) is still about C$30/t above Atlas’s sustaining-inclusive unit cost of C$34.90/t, meaning even at a 20% price haircut, the project generates positive cash margin before royalties and taxes and retains approximately C$544M of NPV. This is the structural consequence of the margin moat discussed above. Few mining projects can say that.

The Discounting Mechanics

The C$920M headline is a present-value figure, base date Q3 2025, discounted at 8%. The model projects 28+ years of cash flow (4 construction + 24.25 operations), then collapses them to a single number using one arithmetic operation: divide each future dollar by (1.08)^years from base date.

Over the 24-year mine life, Atlas is projected to generate C$6.66 billion in pre-tax undiscounted cash flow. After discounting at 8%, that becomes C$1.68 billion pre-tax. After taxes and discounting at 8%, it becomes C$920M. The time value of money eats approximately three-quarters of the nominal cash before taxes: a tonne of salt shipped in 2053 is worth only 15% as much as a tonne shipped in 2030.

The UFS gives three discount-rate cases, and the spread between them is the single most revealing line in the sensitivity table:

Moving from 8% to 10% costs the project C$311M: 1/3 of the headline NPV for two percentage points of discounting.96 That is the largest analytical lever in the model, larger than any operating or capital variable in SLR’s sensitivity table.

And it is exactly where my shareholder reading has to diverge from the model.

Why 8%?

Across 100 NI 43-101 reports filed on SEDAR between 2008 and 2020, the average discount rate was 7.18%,97 with a large cluster at 5% and another around 8%. Gold and precious metals projects skew toward 5%; copper and several industrial or specialty-mineral projects in the sample sat closer to 8%.

So why 8% for Atlas?

The convention varies by commodity, jurisdiction, and stage. AMEC’s reference paper98 on the topic notes that a 2-3% premium is standard practice for juniors with limited development experience, no prior exposure to the project’s commodity, or single-asset exposure. Atlas checks all three: junior, no team in North America has built a salt mine in 25 years and one asset in one jurisdiction. By the AMEC framework, SLR could have justified pushing the rate to 10-11% and stayed within convention.

It did not, and for two reasons we’ve already covered. Salt is a commodity with very low volatility (contract price; spot volatility skews to the upside) and a stable upward trend across decades (4% CAGR over forty years). Also, the mine is structurally simple. The things that tend to go wrong elsewhere have fewer entry points here.

The cash flow profile of Great Atlantic in production looks closer to an infrastructure-style industrial mineral asset than to a typical junior mine. It is long-lived, contractable, essential, and tied to an end market where buyers are forced to secure supply before winter. The risk profile should therefore compress materially if the mine is financed, built, and contracted.

So the math sits on a balance. The junior premium pulls up, while the cash-flow stability pulls down. SLR landed at 8%, which sits near the non-5% mining convention and close to where industrial/base-metal projects tend to cluster. Clearly not generous, reasonably fair today, and definitely conservative once the mine is financed and operating.

My Choice of Discount Rate

A discount rate exists to translate future cash flows into present value.

The more unmodeled risks an interval of time contains, the more aggressively future money should be discounted.

From there, I had three choices.

I could push almost all the risk through the discount rate, then value Atlas Salt by flexing the UFS NPV at different rates. I could do the opposite: identify and quantify every risk directly, apply those deductions to the project value, and discount the remaining cash flows closer to a risk-free rate. Or I could use a mix of both. I chose the third path.

I try to identify, quantify, and charge as many Atlas-specific risks as possible. But I also accept that some risks remain outside my reach. The next two sections do most of that work. They are where I quantify the major risks and apply the relevant haircuts.

That leaves the discount rate itself. SLR used 8%. I will say it now: I am already charging far more risk than SLR could reasonably load into a feasibility-study NPV. There is a reason the next two sections represent about one-third of this entire deep dive. I could probably justify using 5% after explicitly modeling so many risks, but I do not want to.

Two reasons.

First, the probability that 5% would be aggressive is too high. I want to avoid inappropriate conservatism, but I also refuse to build the investment case on aggressive assumptions.

Second, simplicity. SLR built the UFS around an 8% discount rate and only provides sensitivities for selected variables, with discount-rate outputs at 5% and 10%. Keeping the 8% anchor makes the bridge cleaner and easier to audit.

So I keep an 8% discount rate throughout the valuation.

Using 10% would be even less appropriate in my view. It could have been defensible inside the UFS as a broad risk adjustment. But after explicitly charging financing friction, working capital, corporate costs, residual leakage, execution risk, delay risk, ramp-up risk, dilution, and many interactions between those risks, applying a 10% discount rate on top would double-count too much of the same risk.

The preamble is over. Now to the real work: assessing risk.

9. Risks

The Financing

Atlas owns a feasibility-stage project with C$589.1 million of pre-production CapEx, a C$920 million after-tax NPV8, and no operating cash flow yet.

Before the thesis can become real, that financing gap must be closed. The risk is that it closes on terms that dilute away too much of the upside for existing shareholders.

Atlas raised C$8.7 million in October 2025, giving it enough capital to keep advancing early works and engineering, but not enough to build the mine. By year-end, cash was already back to C$6.0 million, which would not be enough to fund even the first half of 2026 now that early works had begun. The UFS allocates C$64.7M to YR-4 (2026), including C$39.3M of mining capital (starting around middle of Q3-2026), before the project steps into a C$100M second pre-production year. Atlas will therefore need to raise additional capital in Q2 2026, or by early Q3 2026 at the latest, if it wants to stay on schedule. Endeavour Financial is leading the project-financing process. But the full project-financing package is unlikely to close before the bridge cash need arrives, making a bridge financing almost certainly required. Sandvik has announced an expanded non-binding MOU covering approximately C$132 million of equipment and services, with vendor-supported financing contemplated if the agreement converts into definitive documentation. Scotwood, Hatch, and Continental Conveyor also reduce the “empty-shell” problem: Atlas is surrounding the plan with a prospective customer/retail partner, engineering, equipment, and material-handling partners.

Some dilution is unavoidable. Even if the company is able, on paper, to raise all the required capital through debt, senior lenders will likely require a minimum equity contribution to align stakeholders and cover items debt will not fund (part of the financing fees, vendor deposits, contingencies, etc.). More than a question of price, it will also be a question of timing. If Atlas does its bridge financing with equity, that dilution will probably happen at a valuation close to today’s. But if dilution is pushed back, for example after a binding offtake or vendor contract or after the announcement of the main senior project-financing package, those milestones should be reflected in the valuation, and the equity raise would make more economic sense. The near-40% insider/related-party ownership has to play its role precisely here.

From here, the checklist is fairly clear:

• Bridge financing in the next few months (by early Q3 at the latest), to keep Early Works moving while Atlas waits for the main project-financing package, ideally through short-term debt or through a favorable convertible structure.

• Conversion of Sandvik’s MOU into binding vendor financing.

• A senior debt term sheet, with an update expected around summer 2026 according to the CEO.

• Additional binding offtake, especially any upgrade to Scotwood’s volumes or terms. This would reassure potential lenders. For shareholders, it would be one of the most important milestones, because it would confirm that a meaningful portion of Atlas’s salt could be sold at better margins than the UFS valuation captures.

• The price and timing of any equity raise. This is the key dilution variable. Equity raised before the next milestones is, in my opinion, expensive equity.

The Build

This is the second major risk for junior miners, alongside financing.

Great Atlantic Salt does not carry the full usual hard-rock mining risk stack. What remains is more practical: can Atlas access the deposit, build around it, and ramp it at the cost and pace assumed in the feasibility study?

It is SLR’s job to quantify all of this in the UFS, and here is its conclusion: after mitigation, the two risks that remain high are failure to achieve the planned decline advance rates and capital cost increases. In plain English, the declines can take longer to drive than planned, and it can cost more to build than planned. Groundwater inflows, grouting, dewatering, contractor productivity, and the rest all flow into the same bucket: time and money before first cash flow.

The problem with delays is that they increase the amount of capital the company needs before the mine starts funding itself. As we just saw, capital raised before the re-rate is the most expensive capital in the story.

Some mitigation is already embedded in the plan. Atlas and its consultants have completed a substantial geotechnical and hydrogeological program, including drill holes, packer tests, vibrating wire piezometers, and laboratory testing. Atlas is also trying to reduce execution risk the right way, along the lines SLR recommends: experienced mine-build management, early engineering work, contractor engagement, and fixed-price packages where possible. It has also put real names around the project: Hatch on engineering, Continental on material handling, and Sandvik on the underground equipment package.

But reducing uncertainty does not eliminate risk: advance rates are only proven once the declines are being driven, and water management is only proven underground.

I do not underwrite those risks as project-killers. As we have seen, cost overruns beyond the contingency already embedded in the UFS have only a moderate impact on NPV, while delays mainly show up as higher pre-production funding needs. Great Atlantic is a simpler build than most mines, and several parts of the design draw on proven salt-mine analogues, including Kilroot, an underground salt mine in Northern Ireland that has been operating since 1965.

I would rather underwrite them as dilution amplifiers. If the declines advance more slowly, if water proves more annoying than expected, or if CapEx drifts higher before the financing package is locked, the asset could still be built. But current shareholders may simply have to pay more of the bill, either through less favorable financing terms, or through direct dilution at valuations that would probably be unattractive.

That is what I will watch:

• Follow-up geotechnical and hydrogeological work on the declines. Atlas has already completed a substantial program, but SLR still recommends additional packer testing, groundwater monitoring, transient modelling, and updated inflow estimates before full construction. If this work comes back clean, it would reduce much of the biggest modeled source of delay and extra CapEx.

• Binding execution packages with Hatch, Continental, and Sandvik. The MOUs help, but they do not yet lock scope, pricing, or schedule.

• Remaining capital-phase permits, Turf Point approvals, and the commercial port-use agreement. Full construction and port execution are not fully de-risked yet. Once they are, one more bottleneck disappears. The port question matters enough to come back later.

• A refreshed CapEx and schedule before FID (Final Investment Decision). The UFS number is only the base case. A more precise cost-and-schedule package would de-risk the project, even if it comes in higher than the current estimate. Less uncertainty on the final bill means less uncertainty on the financing need, and that alone could justify a meaningful re-rating.

• Actual decline performance once excavation starts. This is the central build risk: how many meters Atlas can advance per month, and what that says about other risks (water inflows, grouting, ground conditions, etc.). The first months of decline development will be the best early indicator of whether the schedule is holding.

The Commercial Ramp

Even if Atlas finances the project on acceptable terms and builds the mine within the UFS envelope, it still has to sell the salt.

This is not a trivial point, but it is not the central one either. Great Atlantic is not expected to go from zero to 4.0 Mtpa overnight. The UFS assumes a ramp-up, with saleable production starting at 1.6 Mt in the first year before scaling toward nameplate capacity. That gives Atlas time to place its volumes into the market gradually.

The market context also helps. As discussed earlier, Eastern North America is a seasonal, import-reliant road salt market. Deicing salt is not a new product; Atlas is “only” trying to replace tonnes that already move into the region every winter, often from much farther away. The project sits approximately three days from U.S. East Coast ports, can serve nearby Canadian demand, and should be able to ship year-round. The Scotwood MOU also matters here because it shows that a large packaged-salt distributor is already willing to engage before production.

I would frame this as low relative risk on demand. Given the supply-demand tension in North America, Atlas’s salt is likely to find buyers. The question is at what price.

Even if the Scotwood MOU does not convert, even if Atlas fails to penetrate the retail market quickly, it would still be a potentially lower-delivered-cost source of salt in many target lanes, with better reliability than distant imports and a cost position that would be competitive with legacy local suppliers. And in this market, that is everything: reliability and price. So if needed, it could use part of its cost advantage to win tonnes at lower realized prices. That would not be ideal, but it is very different from having no market.

In the end, the commercial ramp has one job: turn tonnes into the highest possible realized price.

What I will watch here is the sales mix:

• Scotwood conversion, or another major offtake partner. Atlas should try to sell its salt at the highest available margins, but not by becoming overexposed to one channel. In retail and packaged salt, the right path is partnership. The sweet spot, in my view, would be 50-66% of production placed with two or three strong retail/commercial partners, leaving the rest for bulk contracts and spot optionality.

• Contracts outside the basic bulk deicing channel. As a shareholder, I want evidence that the customer book is broadening. I do not want Atlas to enter the final construction year with one dominant channel and a financing case built around a narrow buyer base. Senior lenders will likely think the same way.

• A clear logistics strategy for local and regional sales. If Atlas wants to earn more than bulk port pricing by selling salt itself, it needs to show how those tonnes actually reach the customer. Since Atlas has mentioned the possibility of local salt sales before the first official year of production, it would need to figure out the logistics early and show stakeholders that the system works.

The Salt Price

The single most important variable remains the average realized salt price.

A few mild winters can hurt volumes, spot pricing, and the timing of reorders. That does not bother me much. Road salt demand has always moved with winter severity. What matters is the long-term direction of the market: more disruptive winter events, shifting demand geographies, supply chains built for the old map, and buyers that increasingly care about reliability as much as price.

The real long-term risk is regulatory change or innovation that structurally reduces rock salt usage. For now, I do not see it. Environmental pressure is real, especially around chloride accumulation, but the policy response to date has generally been to adopt better salt management practices while still relying heavily on salt or salt-derived products.

That said, this is a variable I will monitor continuously during the holding period. Road salt has survived decades of environmental scrutiny without a scalable replacement because it remains cheap, available, effective, and easy to deploy. But a new credible substitute that works at scale and at comparable cost would be one of the main paths to permanent capital loss, even if I think it is very unlikely. This risk can only be mitigated by active information monitoring. If I ever come across something credible enough to change my view, I will obviously write about it.

The Permitting Tail

Permitting has already done most of the work it needed to do for the thesis. Great Atlantic Salt was released from the provincial environmental assessment process in April 2024, with conditions, and the federal agency (IAAC) confirmed that the project does not require a federal impact assessment. Since then, Atlas has received early works approval and started moving from paper approval to site preparation.

What remains is the permitting tail that comes with any mine build: translating the environmental release into phase-by-phase approvals, satisfying the conditions attached to the release, and making sure the final construction design stays inside the envelope regulators have already reviewed.

This risk is not fully eliminated. It never is. A missing approval can delay a work package. A regulator can ask for more detail. A design change can force another round of process. But this is no longer the kind of permitting risk that usually deserves a large valuation discount. The big question, whether the project is broadly acceptable in Newfoundland and Labrador, has already been answered.

I will watch whether the remaining approvals move in line with the construction sequence. The project has already cleared the main permitting gate, so the main thing that matters now, in my opinion, is whether conditions, phase approvals, and the Turf Point port work create friction at the wrong moment. If they move with the build, the permitting tail stays a schedule detail. Great. But if they start delaying critical work packages, it will become a real financing problem, with the consequences discussed above.

The Port Agreement

Here is the risk in physical form:

Yes, this is a small port. A very small one, actually. One single ship-loading system. And from that small piece of infrastructure comes a risk that could derail the whole thesis, one Atlas does not fully control: the commercial agreement with the owner of Turf Point. Every number in this deep dive assumes that at some point before first salt shipments, this agreement is signed.

From an operational perspective, every other counterparty can, in principle, be replaced: offtaker, equipment supplier, engineer, lender. The port cannot, at least not without redesigning the full logistics case. The mine plan, the conveyor, the material-flow route, the site layout, and the environmental assessment are all built around this specific port. There is no alternative deepwater port within similar economic distance.99 Atlas is structurally captive to one counterparty.

This is far from a red flag today. Atlas is approximately four years from planned first revenue and has not closed its main financing package. The red flag would appear if financing stalls and the port is cited as the reason.

Several things mitigate the risk:

1. First, Turf Point is not a greenfield port. Atlas, first through its corporate predecessor Red Moon and now under the Atlas Salt name, has long operated in the same district and is already using Turf Point to ship gypsum from the Ace Gypsum quarry since the first shipment in September 2018. The commercial relationship with the port operator has been continuous for years and remains active in 2025, with C$67,000 in gypsum proceeds recorded through the first nine months of the year. Given management’s track record and standing in the region (particularly Laracy), the relationship likely predates the current salt-project negotiations through his various ventures.

2. The incentives are obvious. Atlas plans to spend C$58.5M on upgrades to a port it does not own. By the end of the capital program, the port owner should hold a substantially more valuable asset, without necessarily paying for the upgrade itself.

3. If Atlas completes the upgrade as planned, the existing berth’s loading capacity would rise to 1,400 tonnes per hour. At a 4 Mtpa throughput, this translates into approximately 10 hours of loading per day across 285 operating days per year, every year, for most of the 24-year mine life. Compare that with Turf Point’s current documented export base of approximately 150,000 to 200,000 dry tonnes per year of gypsum. It is as if Amazon approached a small regional airline offering to pay for a new plane fresh off Boeing’s assembly line, and then pay for 24 years to use it.

4. The counterparty is a local family-owned business rather than a large corporate group with extensive capital and legal resources. They could try to leverage their position within this relationship, but any impact on the NPV should not be large enough to break the economics.

The risk also appears to have been worked on by management for years. In the NI 43-101, SLR notes that it relied on a 2021 Stewart McKelvey “Turf Point Property, Report on Title” memo from Justin Hewitt to Patrick Laracy for ownership and title information. Hewitt is a Stewart McKelvey lawyer in St. John’s whose practice covers corporate, commercial and real estate matters, including mining and commercial agreements. That means the legal status of the port had already been identified as a specific due diligence item years before this project-financing process.

That matters because lenders will almost certainly require the same point to be cleaned up before construction debt is committed. The title work is the beginning of that process, and it began several years ago. Before construction debt is funded, lenders will likely require a binding port access/use agreement, plus protections around capacity (tariffs, term, termination rights, remedies, etc.). The relationship must become bankable before construction debt can be drawn.

On operational timing, the UFS gives little concrete detail. The port works are described, but Turf Point does not appear as a separate visible workstream in the summary execution schedule. The only explicit timing disclosed is on the permitting side: in-water works may require Fisheries Act approvals that could take 18 months or more, although SLR believes the current execution schedule leaves enough time to obtain approval before port construction begins. Still, it also means the port remains less transparent on timing than the mine itself.

On paper, this is a symbiotic relationship. Atlas can bring enormous value to a local infrastructure asset; Turf Point, in return, gives Atlas the export route that makes Great Atlantic work. In practice, we cannot know every friction point before the agreement is signed. But both sides have a lot to lose if the contract is not signed. And I suspect the Dolomount family, which controls the port, wants this mine to be built at least as much as I do as a shareholder.

So I see the risk as very low probability, but potentially devastating if it materializes.

I will watch two things here.

• Obviously, whether the port agreement has any impact on the schedule, including through the potential 18-month-plus permitting process for in-water works.

• More importantly, whether the main financing package arrives on time. Senior lenders should have access to far more information than we have on the relationship between Atlas and Turf Point. If the financing package closes on schedule, it would be a strong signal that they found no blocking issue in this potential bottleneck.

The most important risk is the average realized salt price Atlas would achieve. I am willing to accept the other risks materializing: construction delays, CapEx overruns, even financing that is not especially favorable. The mine could be two years late and CapEx could come in 50% above budget; if Atlas signs a binding contract with Scotwood or another distributor to push 50% of production into packaged/retail-related channels, and if salt prices keep following something close to their historical trajectory (which, as a reminder, reflects durable economic forces), the real project value could still end up above the C$920M NPV.

Speaking of valuation, it is finally time to get to the heart of the matter.

Subscribe now

10. Valuation

The Ratio and Its Double Reading

Buying Atlas today costs approximately C$120 million. Building the mine would cost C$589.1 million more. Almost 5x what was just paid to own it. At the end, if the UFS holds, a mine throwing off approximately C$188 million in after-tax cash flow per year, for more than two decades at least.

Put plainly: P/NPV is about 0.13x. For every dollar of value SLR assigns to the project, the market pays thirteen cents. Producing mineral assets typically trade closer to 0.6x to 1.0x of NAV/NPV. Atlas is not producing. But even adjusted for what separates today from 2030, the gap is unusual.

The C$920 million NPV sits at the project level. It is the UFS value, discounted back to Q3-2025, before any adjustment for financing, dilution, execution risk, or the fact that shareholders do not yet own a producing asset. Today’s market cap is a different object. One describes what could be, the other what is.

No simple arithmetic reconciles them. What bridges the two is the set of risks a greenfield mining project still carries. Atlas has cleared most of them. The provincial environmental assessment process was cleared. The reserve and resource base is defined. The local community is onside.

What remains is execution and financing, the two risks that dominate the next four years. Both need to be read in light of what was discussed in Sections 2, 3, 7, 8, and 9: Great Atlantic’s characteristics should make execution risk lower than in a typical mine, and the same features should also reduce financing risk, made easier by the nature of salt pricing: low volatility and a clear upward drift over time.

Two readings of the 0.13x P/NPV coexist.

The first is that the market is pricing those risks honestly. A junior four years from first tonne, with no construction financing closed and no binding offtake, can deserve a steep discount to its modeled NPV.

The second is that the ratio mostly reflects factors outside the project itself: no real analyst coverage, TSXV illiquidity,100 exposure to an industrial segment generalists find boring, and the hangover of the 2022-2025 junior mining cycle.

I spent the rest of this section arguing that the second reading is closer to the truth, and that even if the first reading deserves more weight, the math still works for patient capital.

Before getting there, one more filter has to be named. Even if Atlas delivers exactly what SLR models, the existing shareholder does not keep a 100% economic claim on the full C$920M. Between now and first tonne, Atlas still has to raise a large equity cheque, approximately C$125M to C$383M in my model, depending on the final debt stack, vendor financing, bridge financing, and the full project funding cost, including the items treated outside the headline UFS and initial CapEx (see Section 7 and 8). The price at which that equity is issued determines what’s left in existing hands.

At C$0.50, little. At C$2.50, much more.

Two questions shape what follows. Why does a junior rarely trade at 1x P/NPV, even when everything goes right? And what’s left for the existing shareholder once the project is financed?

Qualitative Valuation

Here, no heavy math, promise. The dirty work, assumption by assumption and discount by discount, comes in the subsections that follow. This first pass is only meant to check that the napkin is not too far from a desirable outcome.

Producing mines rarely trade at 1x P/NAV. Even the best operators, in tier-one jurisdictions with steady cash flow, usually trade at a discount. That is not irrational. A discounted cash flow is the present value of every future year through the mine’s depletion, and every one of those years still has to happen.

The earlier you are, the wider the discount.

Methodological note. Throughout this analysis, I often refer to Atlas through the Great Atlantic Salt’s NPV because, for underwriting purposes, Atlas is effectively a single-asset developer. In mining equity research, valuation comparisons are usually framed as P/NAV. NAV captures the value of the entire company: project NPVs, cash, debt, and other corporate adjustments. NPV is the present value of a specific project. In Atlas’s case, the distinction is mostly semantic: Great Atlantic Salt drives almost all of the company’s current value, while cash and liabilities are small enough that Atlas’s NAV is very close to the C$920M after-tax project NPV calculated in the UFS. I will therefore use NAV or P/NAV when discussing the industry framework and the Lassonde Curve, and NPV or P/NPV when applying that framework to Atlas or referring specifically to project-level value. In this case, for all practical purposes, the concepts are broadly interchangeable.

Pierre Lassonde, co-founder of Franco-Nevada, popularized this pattern in the 1990s.101 A mining project’s market value tends to move through a recognizable curve: a speculative spike at discovery, a long valley through feasibility and permitting, a re-rating at financing, another at first production, then gradual convergence toward 1x NAV as operating history accumulates. The valley in the middle has a name: the orphan phase. Too late for drill-hole speculation, too early for institutions that cannot buy pre-production stories. A strange place to be. Often a very good one.

It is not some kind of model pulled from an average of observed mine valuations. It is a real process governed by economic and psychological laws. Put more simply, it is a discovery rather than an invention. Through this framework, the valuation ranges have become well known, even though the exact bands vary depending on the source, commodity, jurisdiction, and cycle. A first economic study, what the industry calls a PEA (Preliminary Economic Assessment), usually gets a project to a P/NAV range of approximately 0.3x to 0.5x. A completed Feasibility Study, with bankable engineering and a visible permitting path, tightens the band to around 0.5x to 0.6x P/NAV.102 Construction moves the multiple higher again. At first production, a new class of investors becomes eligible to buy, and the stock can start trading less like an option on a mine and more like the mine itself, valuation can then begin to converge toward 1x NAV.103

Here’s the Lassonde Curve:

Let’s transpose it to Atlas’s market cap.

Three phases are visible. The speculative spike from late 2020 through mid-2022, triggered by the PEA launch and delivery, took Atlas from approximately C$5 million of market cap to C$380 million. A 36x move in the stock two years, after dilution (a 75x in market cap). Then came the orphan phase. From 2022 through late 2025, the stock ground lower for three years as the project matured and the market lost interest. This is classic junior mining cruelty: the asset got better while the share price got worse. Since late 2025, the first signs of a re-rating have appeared. Market cap moved off the lows, from approximately C$35 million toward C$120 million, as the UFS landed and Early Works began on site.

Atlas has crossed the dead zone of the Lassonde curve and reached the stage where the market starts asking a different question. For three years, Great Atlantic was a strange industrial mineral project with a study attached. Today, it is a major-permit-cleared project with a completed feasibility study, Early Works underway, and project financing in process.

That alone could form the basis of a thesis: a project leaving the orphan phase and already beginning to re-rate. The problem, or rather the opportunity, is that the curve implies a valuation Atlas has not yet reached. That is the second source of upside.

There is one honest caveat. The 0.3x to 0.6x P/NAV developer range is drawn mostly from gold, copper, and base-metals juniors. Industrial minerals developers are rarer and less familiar to investors. That could justify using the low end of the range for Atlas.

But the caveat cuts both ways. The fact that most developer comps are gold and copper says more about the availability of comparable companies than about the right multiple for an industrial-minerals asset.

In gold or copper, a project trading at a P/NPV of 0.1x on the eve of construction usually says something ugly. The market has seen that movie too many times. It does not trust the project, the numbers, or the team. That is statistical learning at work.

Salt does not have that history in public markets. A listed mining company whose only current business is developing a salt mine, with no operating asset and no precious metal angle, is nonexistent in North America. The closest historical reference I have found is Malagash, Canada’s first underground rock salt mine, developed in Nova Scotia after drilling began around 1917 and the mine opened in 1918. That was another era. Almost nobody in today’s investor base has a mental model for this. Statistical learning is complicated when the sample size is basically zero. Most investors have never had to value a salt mine, think about road salt pricing, analyze the evolution of deicing salt supply and demand, or ask why this project might be simpler than the mines they already know. That is why I decided to spend hundreds of hours of my time on deicing salt.

Put that aside for a moment and look only at the mining and economic characteristics. Deicing salt prices have compounded at about 4% annually for decades, with limited drawdowns and occasional sharp spikes when supply tightens. Gold would envy that pricing profile. The mine itself is structurally simple: shallow deposit, no metallurgical processing, no tailings, short flowsheet. By my reading, Atlas belongs at least in the middle of the 0.3x to 0.6x P/NPV range, maybe near the top.

Seen through the Lassonde curve, Atlas is the archetypal orphan: single asset, single jurisdiction, TSXV listed, industrial mineral, almost no sell-side coverage (the only existing coverage is one month old), and no obvious generalist institutional bid. Most junior mining portfolio managers will not own what they do not know exists. Most small-cap generalists will not touch it because Atlas is pre-revenue. A salt major has little reason to move before the asset is further de-risked. The stock has few natural buyers, except the handful of investors who specifically hunt for this profile.

Two forces point the same way: the curve appears to be turning, and the chart already shows it. Even after that first move, Atlas still trades as if Great Atlantic were several stages behind where it actually sits.

I know, all of this is just valuation by analogy. Every mining project has specifics that can move the real number by thirty percent in either direction, and analogies miss those specifics. The work left is to value Atlas Salt ourselves, as investors who hate losing money do: slowly, with dirty hands.

Quantitative Valuation

The C$920 million headline has already been unpacked. We know what it is: a project NPV, discounted at 8%, built from the UFS mine plan and the costs SLR chose to include.

Shareholders do not really care about what Great Atlantic is worth in the UFS. The final problem, our final problem, is what remains for Atlas shareholders after the missing costs are brought back into the analysis, discounted, financed, and finally divided across the share count that will exist when the mine is funded.

One clarification before the math: the UFS is my starting point, I do not mechanically deduct the same cost line twice. Pre-production CapEx, sustaining capital, operating costs, taxes, royalties, and closure costs are already embedded in the C$920M project NPV. When I adjust them, I do it explicitly as a scenario adjustment: higher pre-production CapEx in bear cases, delay costs, execution leakage, or financing consequences. What follows is therefore the bridge from project NPV to equity value per share: the costs not fully captured by the UFS, the frictions that sit above the project model, and the shareholder-level consequences of funding the mine.

The full mechanics are in the valuation workbook below. I am not going to reproduce every line here. What matters in the text is the bridge: what I deduct, why I deduct it, and what remains before execution, salt price, and dilution.

Here is the full DCF model.

Undiscovered Compounders Atlas Salt Full Dcf

157KB ∙ XLSX file

Download

Download

Subscribe now

I separate the deductions into three layers:

• First, financing friction. Debt does not dilute, but it is not free. Equity dilutes, and the price matters more than the fee. Bridge capital will be needed before the main package closes. None of this is included in the UFS NPV.

• Second, the cash items outside, or not fully captured by, the UFS: working capital, and the corporate costs that continue above the project model (public company costs, board, audit, listing, IR, head office, etc.).

• Third, the bucket that cannot be modeled cleanly: FX drift, future approvals, rights of way, port leakage, drilling, small permitting costs, and the other items that never deserve their own heroic spreadsheet tab but still deserve a haircut.

Only then do we get an adjusted NPV. That adjusted number becomes the base for the next step: execution, delay, salt price and finally dilution.

Methodological point before moving on: every number below (unless explicitly stated) is a present value. The final per-share value is therefore a value today, not a future share price after the mine is built.

To stay consistent with the UFS, I kept the September 30, 2025 valuation anchor. Strictly speaking, time has passed since then. At an 8% annual discount rate, rolling the model forward by six months would add about 4% to the values shown below, all else equal. I am leaving that aside to keep the methodology clean and anchored to the UFS date.

So the numbers that follow are slightly conservative on timing alone.

Start with the cash outflow.

Financing Costs

This is the most expensive missing layer, and the hardest one to model cleanly. So we need to spend a little time on it.

Atlas needs cash to build the mine. I size the financing package around a C$625 million starting funding need: the C$589.1 million UFS pre-production CapEx, plus approximately C$36 million of working capital and corporate cash needs outside the UFS through the construction/start-up window (see Financing section), before scenario-specific CapEx overruns and before the cost of arranging the capital itself. That cash will come with costs attached. I do not deduct the debt raised or the equity contribution as a separate cost, because the capital expenditure they fund is already inside the UFS NPV.

“Costs attached” is the polite version. This is finance. Everyone takes a bite: bridge interest, PIK interest, cash interest, lender fees, equity issuance costs, entry and exit fees, advisory costs, legal costs, and closing costs. Some of those costs are paid in cash. Some are capitalized into the debt. Either way, they reduce what reaches the common shareholder.

The first problem is obvious: Atlas has not published the final financing structure or drawdown calendar. Since those costs have to be discounted, timing matters. A lot actually. The same nominal C$500 million of debt is not the same instrument if it is drawn late, drawn upfront, or allowed to accrue as PIK during construction. Financing is path-dependent, so the model has to make assumptions.

My base framework is simple. Atlas raises bridge financing first, before the main project-finance package closes. The bridge is the toll between early works and the moment senior lenders are finally willing to close. It is repaid once the full package is available. The main package is then split between senior debt, vendor financing from Sandvik, subordinated debt, and equity.

I treat senior debt and vendor financing together for leverage purposes. The economics still differ. Sandvik vendor financing would be equipment-backed, operationally aligned, and probably less punitive than pure project debt. But for the common shareholder, both sit ahead of equity and both absorb part of the project’s debt capacity.

From there, I build three cases: bull, base, and bear. For readers who want the mechanics, the full financing model is in the accompanying file above. I am not going to turn this section into a loan schedule, but here is a summary of the scenarios.

The model requires simplifications. I give Atlas no credit for government funding, no credit for incremental financing-related tax shields, and I do not roll forward the UFS NPV from Q3 2025. I do not try to offset the conservative shortcuts, but I partly offset favorable shortcuts with higher debt costs where appropriate. Again, the assumptions are all in the file.

The financing sheets are built on adjusted CFADS, or Cash Flow Available for Debt Service. That means timing matters twice. First, because earlier cash flows are discounted over fewer years. Second, because those same cash flows determine how quickly Atlas can repay debt. Every additional dollar of revenue pulls repayment forward and reduces financing friction. Every additional dollar of cost increases the funding need, extends repayment, or both.

The result is a present value financing friction of:

• Bull: -C$147.7M

• Base: -C$240.0M

• Bear: -C$325.6M

Yes, that is large: 16% of the UFS NPV in the bull case, 26% in the base case, and 35% in the bear case. Everyone gets paid before the shareholder. In the bull case, the project can still send some cash to shareholders while debt is outstanding. In the bear case, the cash sweep effectively says: lenders first, equity later. This also does not include dilution from the equity raise. That comes later. This section only charges the financing friction attached to funding the mine.

This is, by far, the largest cost layer outside the UFS.

I actually think my bear case is probably too bearish. These look more like the terms lenders would demand from a volatile metals project with a difficult build, not from a simple salt mine selling an essential commodity with something close to a built-in price floor. The MOUs and LOIs are already a strong hint of that. If I were underwriting only what I think is most likely, my bear case would sit only slightly below the current base case, the bull case would stay where it is, still with no credit for government funding, and the new base case would land somewhere in between. But the point of this exercise is to survive being wrong. So I still underwrite the genuinely bad case.

The next deductions are smaller.

Working Capital and Corporate Overhead

That C$625 million number sizes the financing; it does not charge the working-capital balance or the corporate cash items themselves. So if working capital and residual corporate overhead sit outside the UFS economics, they still need to be deducted from the project NPV. That is value shareholders cannot use.

Working Capital

Working capital is cash locked in the business. It comes back at the end, but it loses value on the way.

It is made up of receivables from customers paying after delivery, finished product sitting at the port, consumables, and payables offsetting part of the balance. It therefore grows with revenue.

I model it as a percentage of annual revenue, using 8% in the bull case, 10% in the base case, and 12% in the bear case. The cash impact is only the annual change in that balance. If revenue rises, Atlas has to lock up more cash. If revenue falls, cash is released. At mine closure, the remaining working capital is released back into cash, but with far less value.

Using the UFS revenue schedule as the timing base, with each case’s revenue curve from the DCF, discounted at 8%, this gives:

• Bull: -C$32.6M

• Base: -C$32.4M

• Bear: -C$33.8M

The result looks counterintuitive at first: Base < Bull < Bear, with all three clustered around the same range. The reason is simple: working capital is the toll paid on growth. Each additional percentage point of working capital destroys about C$4.1M in the bull case, C$3.2M in the base case, and C$2.8M in the bear case. Combine those different working-capital rates with different revenue curves, and the order gets messy. In the end, the conclusion is still the same: approximately C$33M of value disappears because cash comes back too late. Who said time is money again?

Now to “real” expenses.

Residual Corporate Overhead

Here I am not adding mine G&A again. The UFS already includes project-level operating G&A and the people needed to operate Great Atlantic. What I am reserving for is the residual listed-company layer above the project: board, audit, listing fees, IR, head office, financing readiness, and the general cost of remaining public while the mine moves from study to construction to first cash flow.

I model that layer at C$3.0M / C$3.5M / C$4.0M per year during construction, in line with the normalized corporate burn discussed in the Financing section. Once the mine is producing, I step it down to C$2.0M / C$2.5M / C$3.0M. Atlas will still be a listed company, with board, audit, listing fees, etc., but it will no longer be a pre-revenue developer trying to finance and build a C$625M project. I escalate those costs at 2% per year (in line with the UFS cost escalation), and discount them at 8%.

That gives:

• Bull: -C$30.0M

• Base: -C$36.7M

• Bear: -C$43.3M

Together, working capital and residual corporate overhead reduce the NPV by C$62.6M in the bull case, C$69.1M in the base case, and C$77.1M in the bear case.

Residual Risk Reserve

Here we are firmly in the artistic part of the exercise. I have to reserve for risks that are, by definition, not precisely estimable.

SLR lists a series of exclusions from the capital cost estimate. Here they are.

Some have already been quantified above. Some represent upside I do not underwrite. Some are either not applicable or already handled elsewhere in the model or by the company. One known item is the geotechnical / hydrogeological program explicitly excluded from the UFS capital estimate. I keep it inside my residual-risk reserve at a fixed C$3.0M (the amount disclosed in the UFS). What remains is the undefined tail.

I am not going to build a tab for each of them. The data is not there, and the precision would be fake.

One item deserves a separate note: currency. It can hurt Atlas, but it can also help. The UFS is built in Canadian dollars, while some equipment, services, shipping, and contractor costs may be exposed to the U.S. dollar. A weaker Canadian dollar can raise those costs. But if part of Atlas’s realized pricing is U.S.-dollar-linked, or set by import-parity pricing against foreign supply, a weaker Canadian dollar can also help the revenue side in Canadian-dollar terms. That is why I do not force a large residual reserve into the bull case.

I apply a valuation reserve directly against the UFS NPV: 1.5% in the bull case, 3.5% in the base case, and 5.5% in the bear case.

This gives:

• Bull: -C$13.8M

• Base: -C$32.2M

• Bear: -C$50.6M

That closes the cost pre-execution bridge.

Put together, the pre-execution cost bridge takes:

• -C$224.1M out of the bull case,

• -C$341.3M out of the base case, and

• -C$453.3M out of the bear case.

The C$920M UFS NPV becomes C$695.9M, C$578.7M, and C$466.7M before execution, the separate salt price / mix adjustment, and dilution.

As a reminder, Atlas Salt trades at approximately C$120M of market cap. We still have room in the bear case. Good, because there is more value left to consume.

Execution

This is the last large project-level risk Atlas Salt will carry until the first tonne reaches the port.

SLR knows it, which is why the UFS budget already contains a cushion. SLR’s pre-production CapEx is C$589.1M, but that number includes C$77.5M of contingency on top of a C$511.6M subtotal. Contingency equals approximately 15% of the pre-contingency estimate. So the first layer of construction disappointment is already inside the C$920M NPV.

I break execution into 5 variables:

• Initial CapEx overrun

• Sustaining CapEx overrun

• OpEx overrun

• Delay

• Ramp-up miss

Case by case.

Bull Case

In the bull case, the contingency does its job. The project is built inside the UFS envelope, on the UFS schedule, and ramps as planned. One could add back part of the contingency to NPV if everything came in perfectly. I do not include it, as I view it as too unlikely.

Execution haircut: C$0M.

Base Case

In the base case, the mine is still built, but the build is messier than the feasibility study assumes. Initial CapEx and Sustaining CapEx come in 15% above the UFS budget, OpEx overruns by 10%, first production slips by one year, and the ramp-up is a little slower than planned.

Initial CapEx overrun. The extra 15% takes initial CapEx from C$589.1M to approximately C$678M. Since the UFS number already contains contingency, this is heavier than it looks: C$678M is approximately 32% above the pre-contingency estimate. Using SLR’s capital-cost sensitivity, I charge C$54.3M against project NPV.

Sustaining CapEx overrun. The UFS carries C$609.1M of sustaining CapEx over the mine life, more than the initial build in nominal dollars. A mine that costs more to build can also cost more to keep running. In the base case, I assume 15% sustaining CapEx drift. Discounted over the mine life, that removes C$27.0M of present value.

OpEx overrun. Operating cost also drifts. The UFS models C$28.17/t shipped FOB Turf Point. That is a good starting point, but as we saw in the risks section, labour, maintenance, contractors, power, port handling, and general underground operating discipline all have room to move. I charge a 10% OpEx overrun, using my interpolation of SLR’s OpEx sensitivity as the conversion table. That removes C$52.7M of NPV.

Delay. I take one year. SLR models a 48-month pre-production period, so one year is a 25% schedule slip. The delay haircut is C$58.3M. I apply the delay after the cost variances, because there is less value left to delay once CapEx, sustaining CapEx, and OpEx have already taken their bite.

Ramp-up. I charge C$20M. The UFS already assumes a three-year ramp to nameplate, so this is only a timing penalty: fewer tonnes early, slower customer conversion and weaker operating rhythm in the first years.

Base execution haircut: -C$212.3M.

That is the normal ugly case, the usual translation from feasibility study to the ground.

Bear Case

This is where I bring in the mining reference class.

McKinsey’s 2024 review of mining project delivery found that, over the past decade, major mining projects suffered average real cost overruns of approximately 40% and average real schedule delays of approximately 25%. Larger projects did worse, and the type of extracted material was one of the variables that mattered most.

As discussed earlier, Great Atlantic sits on the simpler end of the mining spectrum. Its construction file is mostly underground access, material handling, storage, and port logistics, rather than a metallurgical processing plant, a concentrator, a tailings system, or a shaft. The remaining risks are more… prosaic: decline advance rates, water management, contractor productivity, conveyors, commissioning, and the first operating rhythm. They can still cost time and money. They just give the project fewer ways to go wrong than the average mine in McKinsey’s sample.

Still, the bear case has to hurt.

Initial CapEx overrun. I assume a 30% overrun on the UFS initial CapEx number. This is not a direct SLR sensitivity point, so I use SLR’s capital-cost sensitivity as the conversion factor. The overrun takes Initial CapEx from C$589.1M to approximately C$766M, or approximately 50% above the pre-contingency estimate. I charge C$108.7M against NPV.

Sustaining CapEx overrun. I assume 25% drift. The remaining sustaining risk is mechanical: new levels, fleet rebuilds, conveyors, underground infrastructure, and maintenance. Discounted, that removes C$44.9M. Note that the impact is limited because those expenses sit far in the future and are therefore heavily discounted.

OpEx overrun. I assume 15% drift. Labour, maintenance, contractors, power, port handling, and insurance compounding above the UFS line escalation. That removes C$79.1M.

Delay. I take eighteen months. On a four-year construction schedule, that is 37.5% slippage, already worse than McKinsey’s mining average, applied to a project that should be easier than the average mine. After the cost variances already charged, the present-value cost is C$75.0M.

Ramp-up. I charge C$50M. This is the least precise line, so I keep it as a proxy. Against SLR’s production-loss sensitivity, it is equivalent to 2.5 points of additional production-loss effect.

Bear execution haircut: -C$357.7M.

Putting all the cases together, we get this:

• Bull: -C$0M

• Base: -C$212.3M

• Bear: -C$357.7M

Now put execution back into the full pre-price-adjustment bridge. After financing friction, working capital, corporate overhead, residual risk, and execution variance, the C$920M UFS NPV becomes:

• Bull: C$696M

• Base: C$367M

• Bear: C$109M

In the bear case, almost 90% of the UFS NPV is consumed by costs. But there are still two more variables to account for. The two variables that required more than 20,000 words, several hundred hours of work, and that sit at the heart of this opportunity: the price of salt and the product mix.

Salt Price

This is the most important part of the deep dive. So before touching the model, let me restate the market.

The North American rock salt market is structurally short. It has to import 8 to 10 Mtpa of deicing salt, most of it into the East Coast, often from less reliable offshore supply chains. Domestic supply is tight too, with several incumbents already under pressure. Demand has a floor because public agencies must hold minimum inventories. Every winter, municipalities, private contractors, and retail buyers need deicing salt. Consumption can vary with weather, but the need does not disappear. Mild winters mostly shift the timing of replenishment while severe winters expose the shortage immediately.

That shows up in price.

Over the last forty years, U.S. rock salt PPI has compounded at 4% per year. According to Atlas Salt’s investor presentation, the actual historical price CAGR would be approximately 4.2%, with very limited drawdowns, and shallow ones when they happen. That is before looking at the spot market, which can spike violently when winter stress meets tight inventories. During the 2025-26 winter, spot prices tripled in some regions, just as they had the winter before. Those spikes do not vanish when the snow melts. Empty depots have to be refilled, buyers remember who delivered and who did not, and part of the spot pressure flows into the next contracting cycle.

Atlas also has a mix angle. Only about 35% of planned production is currently covered by the Scotwood non-binding MOU, and management wants to push a meaningful portion of volumes into retail through that relationship. Retail is a different market from bulk municipal salt, and Atlas could earn materially higher margins there than in bulk, depending on the final JV economics

Finally, climate change changes the pattern of use. More freeze-thaw cycles, more freezing rain, and more unstable winter events. More salt will be consumed at the wrong time, in the wrong place, and less predictably, in a market that already runs on tight flows. Atlas is the only new North American salt mine with a completed UFS and Early Works underway in twenty-five years.

That is the state of the North American deicing salt market, and that is the price backdrop we are valuing.

With that market in mind, now take the UFS price curve.

SLR escalates the rock salt price at 4% per year for five years from the base date, and 2% per year thereafter. The starting point is defensible, but the fade to 2% is where I completely disagree. Not if we take into account everything above.

My base case is obvious: 4% CAGR. That is the reality of the last several decades. I use 5% CAGR in the bull case and 3.5% CAGR in the bear case. Why not simply use the UFS’s price curve, at least for the bear case? Because that is where the alpha sits. A 2% CAGR strikes me as absurdly low relative to any reasonably plausible scenario. As a shareholder, I am not trying to prove that Great Atlantic Salt can work under a sterile feasibility-study curve. I am trying to estimate the distribution of outcomes as close to reality as possible, because that distribution determines what I am being paid for the risk I take. A 3.5% CAGR is already a hard underwriting line. It assumes a historically low price trajectory, without giving credit to a single one of the factors discussed above.

So, what does this do to NPV?

For each year, I compare my price index to SLR’s price index, apply that ratio to UFS revenue, and then convert the PV-weighted price difference into after-tax NPV using SLR’s own salt-price sensitivity. In the UFS, a 10% increase in the LOM salt price adds C$188.6M of after-tax NPV8. Put simply, I applied the UFS-measured impact of a 10% increase in salt prices to my bull, base and bear cases using linear interpolation.

That gives:

• Bull: +C$753.5M

• Base: +C$392.9M

• Bear: +C$233.5M

The bear case adds C$233.5M to NPV. The base case adds C$392.9M, and the bull case adds C$753.5M.104

This is why the opportunity starts where the UFS stops.

Take a seasoned mining investor. They read the UFS, take SLR’s assumptions, run the same risk math I did, then run the dilution math, and think: “No. Too far-fetched, not enough margin of safety. I’ll pass.” Fair. They have seen hundreds of gold, copper, or uranium projects. They have seen cost overruns, broken studies, optimistic ramp-ups, and shareholder dilution disguised as progress. What they have probably not seen is a salt mine. No comparable new rock salt mine has opened in North America in twenty-five years. No pure-play publicly listed salt mine developer that has gone through this exact path in decades. There is no mental model sitting on the shelf. It is simply the result of hundreds of hours of work spent understanding the Eastern North American deicing salt industry and Atlas Salt’s place within it.

The market gap is therefore partly educational. Atlas needs investors to understand what kind of asset they are looking at.

The CEO deserves credit here. Over the last few months, he has been speaking at many conferences and investor presentations, trying to educate investors on Atlas and on the salt market. That is how I found the company in the first place. In the very short term, this is probably the best way to unlock value. The story is there. It is just buried under the unfamiliar. Undiscovered, in the literal sense.

Shareholders can only hope he keeps doing it over the coming months, before the heavy work really begins.

There is still a limit to what management can say. What are they supposed to do? Stand up and say: “SLR completely underestimated the salt price, our NPV should be 40% higher, believe me”? In a sector where everyone is suspicious, they would lose credibility immediately. Probably with good reason.

My position is different. I am not management. I do not get paid to convince anyone to buy the stock. I invested based on these numbers myself. The assumptions are laid out. The model is visible. The sources are there. All I have done is collect the information, organize it, and show the math. Anyone putting money at risk should verify the work independently.

Once I replace SLR’s long-term 2% fade with a price curve that better matches the salt market, the rest of the model becomes much easier to underwrite. The thesis can absorb an eighteen-month delay, a mine built about 50% above the pre-contingency estimate, and still leave out the 288 Mt of additional Indicated Resources and the 868 Mt of Inferred Resources not included in the mine plan, possible low-cost government funding, higher retail margins, and everything else that could still go right.

There is still one more source of potential upside to address: the sales mix.

Sales Mix

The UFS assumes Atlas sells into four buckets:

• U.S. East Coast: 56.0%, or 2.24 Mtpa

• Canadian Maritime Provinces: 8.5%, or 0.34 Mtpa

• Québec: 15.5%, or 0.62 Mtpa

• Spot sales and commercial: 20.0%, or 0.80 Mtpa

That gets Atlas to the full 4.0 Mtpa production rate. SLR then converts that destination mix into a weighted-average price of C$81.67/t FOB Turf Point. So, regardless of whether the physical sale is CIF, DAP, or FOB, the economic model brings everything back to one FOB Turf Point price (see Section 4). In other words, the UFS collapses geography, channel, contract structure, and product format into one number. Clean? Yes. Lendable? Yes. But too clean for a shareholder.

The UFS treats Atlas mostly as a bulk supplier selling into destination markets at a normalized FOB-equivalent price. It allocates 20% of volume to spot sales and commercial. That line alone is 0.8 Mtpa, and it is explicitly not the same animal as bulk public-sector tonnes. Yet, once the economics hit the model, everything still gets compressed into C$81.67/t FOB. That is where I think the UFS leaves value on the table.

Scotwood had already signed a non-binding MOU with Atlas that contemplated 1.25 to 1.5 Mtpa of volume through a Canadian retail / packaged-salt partnership. Against the 2023 mine plan, that was approximately half to 60% of planned production. Against the new 4.0 Mtpa UFS plan, it is still 31% to 37.5% of annual volume. According to my estimates (see Section 4), if 1.5 Mtpa eventually earns C$60-80/t of incremental economics above the bulk FOB case, that is C$90-120M of annual pre-tax economics before discounting. Over the mine life, discounted at 8%, with no additional price escalation, that rough math gets to around C$630-840M of present value after leakage and taxes.

And Atlas does not see that as the ceiling. Management has said it is discussing a larger retail/distribution allocation with Scotwood and other distributors, which would increase the share of that channel and could push it close to, or even above, half of the sales mix. The company has also mentioned selling directly into Newfoundland and Labrador, where logistics are shorter and Atlas could be able to capture more of the downstream margin itself. So the gap is clear: the UFS still treats most of Atlas’s volume as bulk road salt, while Atlas is actively trying to move a larger share of its salt into higher-margin channels.

Management has both the ability and the incentive to push more salt into higher-margin channels. The fact that the largest packaged retail deicing distributor in the United States approached a junior developer six years before first production tells me the UFS mix is likely too conservative. But it is still only a signal. There is no proof. There is no binding contract. Good. If the risk had already disappeared, the valuation would not look like this.

For me, the bear case is boring bulk economics. And that is essentially what the UFS modeled. So my bear-case starting price is C$81.67/t, the UFS FOB Turf Point price.

From there, I built the base and bull cases. Instead of blindly stacking assumptions around unknown channel mixes and unknown price points, I kept the UFS price as the anchor: C$81.67/t. Then I added a simple mix-driven uplift: +C$5.83/t in the base case, and +C$13.33/t in the bull case:

• Bull: C$95.00/t

• Base: C$87.50/t

• Bear: C$81.67/t

That is not aggressive. At 4.0 Mtpa, the base case uplift is only C$23.3M per year before escalation. Using the C$60-80/t Scotwood economics, that is the equivalent of only 0.3 to 0.4 Mtpa of higher-margin volume, spread across the whole 4.0 Mtpa production base. The base case only assumes a small amount of channel improvement leaks into the FOB-equivalent realized price.

Same idea for the bull. C$53.3M is still below the C$90-120M/year that a full 1.5 Mtpa Scotwood outcome could imply. Even the bull case only captures the equivalent of approximately 0.7 to 0.9 Mtpa of the 1.5 Mtpa (based again on the C$60-80/t channel economics).

Now the mix-only DCF impact.

That gives105:

• Bull: +C$430.8M

• Base: +C$162.7M

• Bear: C$0

The bear case has no sales mix uplift. The base case adds C$162.7M from mix. The bull adds C$430.8M. In all three cases, the full Scotwood math remains mostly outside the model. Any expanded Scotwood allocation, any new retail / distribution partner, any higher spot allocation, and any local direct-sales margin are also outside the model. This is not heroic conservatism; it is just the only way to handle uncertainty when both the probability and the economics are still impossible to pin down.

The final adjusted shareholder NPV before dilution now looks like this:

In the bull case, the cake is enormous. Even in the bear case, there is still significant cake.

But one final question remains: how many slices will the cake have to be cut into?

Share

Dilution

Great Atlantic is too large for the current balance sheet. The UFS pre-production CapEx is C$589.1 million, before the shareholder-level layers added in this model. Atlas can reduce the equity cheque with senior debt, vendor support, subordinate debt, and a bridge facility, but it cannot eliminate it.

I start from the fully diluted share count. Basic shares are approximately 110.7 million. Fully diluted shares are approximately 119.6 million. The current option, RSU, PSU, DSU, and broker-warrant stack is already inside that number, so I use it as the denominator from the start. I then add two small reserves on top: a financing sweetener for warrants or similar instruments needed to clear the construction equity raise, and a future SBC reserve through first production. The SBC reserve is modeled as a percentage of the final diluted share count. I also give no credit for possible cash proceeds from future option or warrant exercises.

In the execution layer, I charge cost overruns on top of the UFS pre-production CapEx number, which already includes the 15% contingency: an additional 15% in the base case and 30% in the bear case. These overruns also have to be financed, at least partly with equity. OpEx and sustaining CapEx overruns are assumed to be funded from operating cash flow. They reduce CFADS and slow debt repayment in the financing model, but they do not create additional construction shares.

I do not assume the entire overrun is funded with equity, that would be too blunt. I also do not assume senior lenders simply increase their exposure on the same terms, that would be too generous. In the base case, 40% of the overrun is funded with junior / cost-overrun debt and 60% with equity. In the bear case, 25% is funded with junior debt and 75% with equity. That adds C$53.0M of equity in the base case and C$132.5M in the bear case. The resulting extra debt and equity then runs through the same constraints as the base financing package: fees, DSCR (debt-service coverage ratio), PIK interest (payment-in-kind), cash sweep, etc. Again, every detail is in the DCF.

Once the base dilution and the dilution caused by a potential CapEx overrun are included, the net construction-equity requirement, before grossing up for issuance fees, is:

• Bull: C$125.0M

• Base: C$240.5M

• Bear: C$382.5M

Great, we have the amount. But at what price?

The main variable is timing: when does Atlas raise, and how much? Before senior financing? After senior financing? After the construction corridor? After a Sandvik conversion? After a binding offtake? Every layer of de-risking should increase the probability of a re-rating, and therefore reduce the number of shares Atlas has to issue for the same amount of cash.

Rather than inventing a full issuance schedule out of thin air, I use a blended issue price across the construction equity for each scenario.

• Bull C$2.5: Atlas re-rates quickly and keeps pushing the equity raise further down the curve.
  Final diluted shares: 177.2M

• Base C$1.75: Atlas finances part of the project relatively early, then moves moderately along the Lassonde curve over the following years.
  Final diluted shares: 279.9M

• Bear C$1: No real re-rating. The share price even weakens, and Atlas issues shares by the hundreds of millions until the relief of the first tonne shipped.
  Final diluted shares: 571.9M

From bull to bear, the share count is multiplied by 3.2. All else equal, that means about 1/3 as much value per share.

This is why the sequence over the next 12 to 24 months is essential. Every binding agreement before the equity raise changes the denominator. Every milestone reached increases the probability of a re-rating and reduces the number of shares issued. Since I started writing this deep dive, the stock has already risen more than 60%, but in my opinion, it is still far from a price that would make equity financing attractive.

I also think management may be making a capital allocation mistake.

In the investor presentation, management suggests that part of unlevered FCF could be returned to shareholders from the first production year. I understand the idea. A dividend can attract a different type of investor, possibly open the door to certain income-oriented funds, and maybe support the share price before a refinancing or a future equity raise. But that is a probability.

What is a certainty, if construction debt is still outstanding, is that every dollar of cash flow paid out as a dividend is a dollar not available for principal repayment. And every dollar of principal not repaid means more interest accumulating. That is arithmetic.

In any case, I doubt the covenants will allow Atlas to distribute that much cash in the first years of production. The only shareholder return that could make sense early would be a buyback, and only if the covenants allow it and if the stock is still obviously undervalued despite production.

Activist moment over.

I have kept you here for 30,000 words. Time to give you what you came for.

How much is Atlas Salt worth?

How Much Is Atlas Salt Worth?

One vocabulary point before moving on: I start from the fully diluted share count. That makes the adjusted NPV per share slightly lower than a basic-share calculation, because the options, RSUs, PSUs, DSUs, and broker-warrant stack is already in the denominator. Run it on basic shares and the per-share value looks higher. But those instruments exist. I prefer to count them from the start.

Here is the DCF summary:

At a share price of C$1.20, the setup looks like this:

• Bull case, close to a 9x.

• Base case, almost a 3x.

• Bear case, down 50%.

These numbers are NPV per share, so they are discounted values, not future share-price targets. Each year, all else equal, the discounting fades and the present value rises.

Of course, all else will not be equal. During the four-year construction period, some milestones will have been reached, some risks will have disappeared, others will have materialized and been quantified. Time reduces uncertainty and risk, but it also reallocates the pie. The longer you wait, the more of the upside, or downside, has already gone to those who were willing to bear the uncertainty.

And this is where the setup becomes more… interesting:

The valuation above excludes several factors that could materially increase intrinsic value. They are the pieces I have deliberately left out throughout this deep dive, framing that choice as “conservatism” or “safety”:

• First, the resource. The UFS mine plan is built around 95 Mt of Probable Reserves, but the deposit is much larger: 288 Mt of additional Indicated Resources at 96.0% NaCl and 868 Mt of Inferred Resources at 95.2% NaCl. In most mining projects, that would be too speculative to count, but salt is different. Bedded salt deposits tend to be homogeneous and laterally continuous, and the conversion from resource to reserve is usually far cleaner than in metallic mining.

• Second, the sales mix. Atlas has the opportunity to sell part of its salt into channels with margins far above the UFS blended FOB pricing framework. It is actively trying to build that opportunity now through retail partnerships. If the sales mix ends up with 50% in retail/distribution and another 10-20% in local and spot sales, which management appears to be targeting, my bull case for this variable could look like a bear case.

• Third, government financing. Any public support would reduce dilution and financing costs, which are by far the largest external cost layer in the DCF. The project has already received a small amount of public funding once, and the province has every reason to want a multi-generational employer in western Newfoundland. It is a big “if,” but also a big “then what?”

• Fourth, early production. Atlas has mentioned the possibility of starting production before the first official UFS production year. If that happens: less debt, less interest leakage and higher NPV. Even the bull case assumes Atlas follows the UFS construction path. But if early production becomes credible, the market will not wait until the first full year of production to reprice the stock.

• Fifth, the tax shield. I have not credited Atlas for the tax shield created by the extra costs I deduct above the UFS. Some of those costs should reduce taxable income, either immediately or over time. The exact value depends on tax treatment and timing, so I leave it out. Directionally, it is upside.

• Sixth, M&A. This is the card I have waited to play since the beginning. My future as an Atlas shareholder may be shorter than the mine life, because Atlas is an obvious potential acquisition target. The likely window is around first production, either shortly before or shortly after, when the project has been de-risked enough for a buyer but before the market has fully capitalized the long-term cash flows. It would also remove several risks from our side of the table. I will come back to this later.

So the model gives us the core of the argument. The omitted upside gives us the shape of the distribution. And the distribution matters more than the point estimate.

My view is that the probability of landing in something equivalent to the base case or better is above 75%, and the probability of landing in the bear case or worse is below 10%. The remaining probability sits somewhere between the bear case and the base case. I am pulling those numbers from my own judgment. There is no formula for this. But for the bear case to become the central outcome, several things need to go wrong at the same time, while many of the things that could go right need to fail to show up.

• The main variable in the bear case is the salt price. For that case to become the central outcome, salt prices would need to stay much closer to SLR’s cautious long-term curve than to the forty-year producer-price history discussed earlier. That is hard to reconcile with the structure of the rock salt market: aging supply, shrinking domestic capacity, import dependence, procurement stress, and a buyer base that cannot choose to stop buying when winter hits. In both 2025 and 2026, parts of the market ran short of salt and spot prices spiked. These episodes are starting to look less like “exceptions” and more like the new shape of the market.

• Even if salt prices compound more slowly than I expect over the next two decades, Atlas would still have to sell its salt in the conservative UFS mix, with limited benefit from retail or higher-margin commercial channels. That is also not what the company appears to be trying to build.

• Even if those two variables land close to the bear case, a lot of other things still have to line up badly: cautious senior lenders, no government funding, no vendor financing, a share price low enough to create heavy dilution, delays and CapEx overruns beyond the UFS assumptions, no early production, and no meaningful offset from any of the omitted upside.

That is a big “then what,” but it needs a lot of “ifs” to get there.

For those who skipped the 30,000 words and think I am being too optimistic: read the full thesis. Every piece above was developed earlier, one by one, number by number.

And even if the bear case materializes and the share price does not give you a clean exit, you can still wait. Once Atlas has repaid the construction debt, the business should be able to return most of its free cash flow through dividends and buybacks. Under the UFS assumptions, once production reaches 4 Mtpa, that means about C$150M to C$200M of after-tax cash flow per year for about two decades. But we will come back to that in a moment.

This is why Atlas may be one of the best risk/reward setups I have come across in my career.

The upside is large, and I think the market is applying the wrong discount. Yes, the risks are real. Yes, there is a real risk of permanent capital loss: a historic earthquake like the 1929 Newfoundland and Labrador event, an armed conflict, name your black swan. But this is no longer the usual junior-mining risk of hoping the ground contains something valuable.

Great Atlantic is shallow. It does not require a vertical shaft. There is no lake sitting above the mine. The product is rock salt, so there is no metallurgical puzzle to solve and no complex processing plant to build. The deposit has been drilled, the salt is there, the end market exists, and the project has already cleared most of the permitting path.

What remains is mostly financing and execution, the two risks no greenfield developer gets to avoid. They can seriously hurt the economics if handled poorly. But they are narrower than the discount suggests. And yet Atlas trades at an NPV multiple one-half to one-third of mining peers that, at a comparable stage, often still carry more geological, metallurgical, permitting, or commodity-price risk.

The second source of the discount is simpler: investors do not know where to put it.

There has been no comparable new salt mine in North America for decades. There are almost no public comps, no futures curve, no liquid benchmark. For mining investors, salt looks too boring, sometimes not even like mining. For other investors, a mine looks too risky. So the asset sits between categories.

You can feel it even at the mining conferences Atlas Salt attends. The questions often start from zero. Sometimes there is almost a smirk in the premise: a company trying to sell rock salt, here, at a mining conference. These are mining people, and even they do not really have a mental bucket for it.

It is undervalued because it is undercovered, and because the risks are clearly overstated, creating potential enormous upside.

What we care about now is how this upside materializes.

Subscribe now

11. Repricing

There are only three paths for value to reach shareholders:

• The market recognizes the setup before the cash flow arrives.

• The mine is built and the cash flow is returned through dividends and buybacks.

• Someone buys the asset.

Recognition, distribution, or acquisition.

Recognition

The first path is simple: reduce the perceived risk and close the knowledge gap around the asset.

Atlas has approximately four years between today and first production, currently targeted around 2030. These are four years during which the company can progressively de-risk itself by climbing the Lassonde curve, as hundreds of companies have done before: permits, binding contracts, financing, early works, construction, and then ramp-up.

Some steps are unavoidable: remaining permits, financing, early works, construction, and commissioning. Without them, there is no mine. Others are not strictly necessary, but would change the speed of the re-rating: a binding offtake, vendor financing, strategic equity, government support. Then there are the catalysts Atlas does not control: tariffs, a supply disruption, or a winter harsh enough to remind buyers that road salt only looks boring when the supply chain works.

Start with the gates Atlas has to clear.

The Critical Gates

These are the gates Atlas has to pass before Great Atlantic becomes a mine. Calling them “catalysts” almost understates the point. If one of them never happens, the thesis breaks. The question is timing, cost, and our sempiternal friend, dilution.

Construction-financing package announcement (expected H2 2026 to H1 2027). Atlas has appointed Endeavour Financial as project finance advisor, and the current financing file already has several visible pieces: senior project debt, possible export credit-agency participation, Sandvik’s non-binding C$132 million equipment and services MOU, and whatever equity remains after the debt stack is set. A term sheet announcement could move Atlas closer to the low end of the funded-developer range, around 0.2x to 0.3x of P/NPV.106 Atlas moving from 0.13x to 0.3x would represent approximately a 2-2.5x re-rate on the market cap alone, before any further catalyst is priced. Closing the financing package is bigger still, and in my comparable-developer framework could be worth another 0.1x turn of P/NPV in comparable cases. As a current shareholder of Atlas Salt, this is the event I am waiting for the most. And every shareholder should be too.

Binding offtake agreement with Scotwood or another major commercial partner (likely around the financing process, if it happens). At some point around the financing process, Atlas will need to turn commercial intent into something lenders can underwrite. That does not necessarily mean Scotwood at any price, nor an unconditional take-or-pay agreement four years before product exists. Buyers rarely accept that kind of delivery risk before financing, FID, construction, and final product specifications are locked. What lenders will want is commercial visibility: enough contracted or highly credible demand to support the first years of production, and enough confidence that Atlas can place its tonnes at the prices assumed in the financing case. Scotwood matters because it opens the packaged and retail channel, where the UFS gives Atlas essentially no specific credit. If the MOU converts on meaningful volume and acceptable pricing, the impact could be far larger than a simple confirmatory catalyst, it could change the NPV itself. So I would not treat Scotwood as being on the critical path. The base case survives without it. But a binding agreement with Scotwood, especially if expanded beyond the original 1.25-1.5 Mtpa framework, would reduce commercial uncertainty, help the financing discussion, and force the market to rethink the project’s economics. And if the market continues to ignore it, the cash flows will have the final say.

Construction decision / FID (expected H2 2027). The formal transition from “development” to “construction.” FID triggers the drawdown of vendor financing, starts the construction clock, and eliminates the “will they actually start construction?” question. Historically, the FID is worth 0.1-0.15x to the P/NPV uplift on its own, and it removes a categorical risk that was weighing on the P/NPV multiple. After FID, execution risk becomes the dominant risk. And more information should not take long to arrive.

Early decline performance (expected H2 2027 to H1 2028). After the main financing package, this is the most important milestone in my view: how fast the declines move, how wet the ground is and how much grouting Atlas needs before it can keep going? Atlas will do the work to mitigate the risk, but only building the mine will answer the questions. A clean first stretch would tell shareholders that the main decline risk is starting to clear. A messy one would not necessarily break the project, but it would bring back the two things every developer and shareholder wants to keep out of the conversation: time and dilution. As long as salt prices keep rising at 4% or more, and Atlas starts converting commercial intent into contracts that lock in volumes with higher-margin channels, I am willing to wait and accept further “bad” dilution. As we have seen, delays and additional financing needs are not the most value-destructive variables in the model.

First salt on surface (expected H2 2029 to early 2030). The first salt reaching surface is the moment the market knows the mine physically works. Production can still ramp slowly after this, but the binary “does it exist or not” risk is gone. The P/NPV multiple could move toward the construction-to-ramp-up range, about 0.6x to 0.8x if execution is clean.107

Commercial production (targeted 2030). This is the first moment where the story can be judged by shipments, customers, margins, and cash conversion. It also changes the shareholder base. Many institutions cannot or will not own pre-production developers, whatever the spreadsheet says. Once Atlas sells salt and begins building an operating record, that restriction starts to fade. The P/NPV could then move closer to 1.0x over the following 12-18 months as operational history accumulates.

These are the mechanical catalysts. Atlas simply has to execute. Not everything is within the company’s control, but that is precisely the point: how management responds to problems will reveal the quality of the team. Each one removes a reason to keep the stock in the “penalty box”.

The next bucket is different. Atlas does not need these variables to hit the base case, but they can compress the timeline if they land.

The Variables

I would treat these as second-order sources of re-rating. They just make the market more willing to believe the thesis once the main boxes are checked.

Additional offtake agreements beyond Scotwood. If Scotwood becomes binding on the currently disclosed terms, Atlas still has approximately 2.5-2.75 Mtpa of nameplate capacity left to place. That remaining volume is where the next layer of value could, and should, appear, ideally through emergency/spot sales or through additional retail/commercial distributors. I would not expect one big “aha” moment here. More likely, each contract removes a little uncertainty and gives the market one less excuse to value the project like an orphaned junior.

Strategic investor participation. The October 2025 LIFE financing included a “strategic investor,” but Atlas has not disclosed the name. If that investor, or someone in the same category, appears again in the construction equity tranche, the signal may matter more than the cheque itself. It tells the market that someone with a strategic reason to care has done some work on the asset and was still willing to put money in. In a junior developer, that can move the discount rate in people’s heads before it ever moves in the model. On its own, that could expand the P/NPV multiple by 0.05x to 0.1x, possibly more depending on the name. The market has seen this before. In exploration and development stories, a recognized backer can sometimes reframe the whole asset almost overnight. When someone like Michael Gentile discloses a position, it is not unusual to see a junior move 3x or 4x in a few weeks simply because the shareholder register suddenly gives the market permission to pay attention.

Analyst coverage. Atlas is no longer completely uncovered. Ventum initiated coverage in March 2026. Still, one analyst is not enough to create institutional ownership, but it is enough to give investors a model, a target price, and a broker who can put the name in front of accounts that were never going to start from zero. Coverage is still thin, but the awareness discount is no longer entirely untouched. Each additional piece of coverage after the financing package is announced increases the likelihood that deep-pocketed investors start paying attention to the setup.

TSX main-board uplisting from TSXV. Many institutions simply cannot own TSXV names, or cannot own them in size. An uplisting widens the buyer base. I would see it less as a catalyst and more as a later-stage “release valve”: once the project is de-risked enough, the exchange itself stops being part of the discount.

Indicated & Inferred Resource conversion. Atlas has 288 Mt of Indicated Resources and 868 Mt of Inferred Resources grading 95.2% NaCl. If even a modest portion is converted over time, the mine life could extend by several more decades. That would lift modeled NPV through duration rather than margin. The market will probably not pay much for it today, which is fair. The interesting point is that conversion should get easier once the mine is underground and Atlas has better geological information than surface drilling can give it. A useful reminder for the patient investor: Great Atlantic is a laterally continuous bedded salt deposit, and salt deposits tend to be relatively homogeneous. Some salt mines have operated for decades on the back of only a handful of initial drill holes. That does not make the Inferred Resource bankable today, but it does make the optionality more credible than it would be in a structurally complex gold or copper deposit, especially in an M&A scenario.

The External Catalysts

Atlas does not control these, but any one of them would compress the re-rate timeline.

1. A severe winter. Spot salt prices crossed C$300/t during the 2025-26 Ontario season. Road salt procurement is backward-looking in the way most municipal procurement is backward-looking: a spot price spike does not fully reset the next contract cycle, but part of it tends to bleed into the following year’s contracted prices. A repeat season would lift the North American road salt pricing environment and make Atlas’s operating leverage harder to ignore.

2. Disruptions to imported salt. A tariff, duty, trade dispute, or ocean-freight shock affecting Chilean, Egyptian, or Moroccan supply could tighten a market that is already structurally short. The current political environment around “buy North American” makes this more plausible than it would have been five years ago. Higher ocean freight would have a similar effect. It would not make imports disappear: even after adding Atlas’s planned 4 Mtpa to North American supply, the market would still need several million tonnes of imports every year. Most likely, the result would be higher delivered prices, mostly passed through to American end customers, whether the buyer is a DOT or a homeowner who still remembers slipping on their icy driveway last winter.

3. Disruption at an existing North American producer. Compass Minerals’ Goderich mine, with approximately 7.5 Mtpa of capacity, has faced labour issues and geological disruptions in recent years. A material shutdown there would leave existing demand with fewer local tonnes and make many procurement managers more willing to pay up, because nobody wants to be the one who ran short of salt when the roads froze.

None of these is required. The Lassonde framework already gives the direction of travel, but external shocks would shorten the distance between recognition and value.

Timing and Sequence

The visual below summarizes the sequence: what has to happen, what can accelerate recognition, and what could force the market to reprice the asset from the outside.

If Atlas re-rates, it will probably happen in steps, both up and down. The important point is that Atlas has more visible paths to positive recognition than to permanent impairment. The negative events, however, are more concentrated in the near term: financing terms, dilution, early-works execution, and the first underground questions.

That is why this setup only makes sense for patient capital. Over a one-year horizon, the risk/reward can be dominated by the timing of one bad event. The thesis can be right and the stock can still trade badly for a while.

But the re-rating is not strictly necessary. A truly patient shareholder can simply wait for the asset to turn into cash flow, then hold it for 24 years, potentially longer if additional resources convert.

Sit-on-your-Ass

At around C$1.20 per share, Atlas does not need to re-rate quickly for the investment to work. It could simply become a cash-flowing asset. A very patient shareholder can sit through the financing, the construction, the ramp-up, and the debt repayment window, then own their share of the mine when the cash starts coming back.

In my model, Atlas has effectively repaid its construction debt by the end of 2036. Add two years for “safety”. By 2039, the project would, under that model, still be generating around C$190 million of after-tax cash flow per year.

Now assume 100% dilution from today. With twice as many shares outstanding, and if that cash flow were fully paid out, that would be approximately C$0.80 per share in annual dividends.

On a C$1.20 purchase price, one year of dividends would return 67% of the initial investment.

Thirteen years later, yes. But at that point, the reserve-backed mine life would still have approximately fifteen years left, before assigning any value to the broader resource base.

The example is deliberately simplistic. It ignores capital allocation choices, tax timing, debt covenants, and whatever else reality decides to add. But it shows the shape of the setup. A truly patient shareholder can buy a share that could, years from now, start behaving almost like a bond with an absurd coupon on original cost.

The UFS implies nearly C$4.6 billion of cumulative after-tax cash flow over the reserve-backed mine life, before shareholder-level financing costs and before assigning any value to the 288 Mt of Indicated Resources beyond the reserve base or the 868 Mt of Inferred Resources sitting outside the mine plan. Against a market cap near C$120 million, I think we can call that an asymmetry.

Again, all the risks discussed throughout this deep dive still apply. But so do the unmodeled upsides.

The market wants recognition. I want it too. But recognition is not the only way to get paid. Waiting can work too.

In fact, there is a way not to wait that long.

Someone Else Pays

Someone else could buy the asset, and I think that outcome is fairly plausible.

Shareholders do not need it for the thesis to work. I am not even sure they should want it, unless the price is exceptional. But ignoring the strategic-buyer angle would be lazy.

Why M&A

Atlas is not worth the same amount to a minority public shareholder and to a potential strategic acquirer.

Atlas owns the scarce piece: a potential new Atlantic supply source in a market short of local tonnes. But in salt, the tonne is only half the asset. The route to the buyer is the other half: terminals, storage, contracts, trucks, retail accounts, procurement relationships, and… trust. That is where a large part of the margin lives.

Atlas has to build or borrow that system, but a strategic buyer may already own it.

That changes the math, literally. The public market discounts Atlas for the missing pieces around the mine. A buyer with those pieces already in place can discount less, probably much less. Its balance sheet and commercial infrastructure are operating assets in themselves.

This is why M&A, if it happens, probably starts quietly. More volume under offtake. A strategic cheque. A distribution agreement. Vendor financing with sharper terms. A partner that wants to get closer before the market understands why. If that sounds familiar, it should. Atlas already brought in an undisclosed strategic investor in the October 2025 financing. Management said that investor’s interest in Atlas and Great Atlantic aligns with its own long-term strategic objectives. I would not turn this into a takeover thesis, but I have read enough headlines like this to know it belongs in the distribution.

In North American salt, that would hardly be unprecedented. The sector has consolidated before. K+S sold its Americas salt business, including Morton Salt and Windsor Salt, to Stone Canyon Industries Holding, Mark Demetree and Affiliates for US$3.2 billion at about 12.5x EBITDA.108 Operating salt assets with customers, logistics, and cash flow have attracted real strategic capital. Great Atlantic is not there yet. But that is exactly the point: the closer Atlas gets, the more it starts looking like something the industry already knows how to buy.

When M&A

The law is the same for shareholders and acquirers: the more Atlas de-risks, the more expensive it “should” become.

Before project financing, the buyer would be taking the whole problem: financing, construction, and execution. That can happen, but only for a buyer strategic enough to want control of the asset before anyone else, and willing to carry the risk to get it.

Once a credible financing package exists, the door opens to more buyers. After FID and the first decline work, the negotiation changes again. A mining operator can then look at the asset and say: we know how to build this. At that point, every meter of decline that behaves as expected should reduce the discount a buyer can demand.

After first salt, the asset becomes much easier to buy, but also much harder to steal.

That is where M&A becomes more interesting for shareholders. Before production, an acquirer is still asking to be paid for the risks shareholders chose to underwrite themselves. A bid at that stage can crystallize value early, but it can also take the trade away before shareholders have been paid for underwriting the hardest risks. The buyer gets to hide the upside behind “remaining uncertainty.” I am not interested in having my alpha buried inside someone else’s risk memo.

After first salt, the discussion changes. The mine exists, the salt reaches surface, and the remaining value is no longer just project risk collapsing. From there, an acquisition is priced more around the buyer: its operating platform, synergies, and the usual industrial logic. The risk discount has shrunk. The buyer can still argue, because buyers always argue, but it becomes much easier to make the buyer pay for part of the synergies.

In short, timing changes valuation. That is why I do not think an acquisition before production would be the best outcome for Atlas shareholders. My job is to take risk and get paid for it. I do not want another company taking the setup too early, then paying me with a discount dressed up as prudence.

But my opinion will barely matter. Management’s will.

Management’s Take

Talking about “management” as one block would be misleading.

Start with Peterson. His line on M&A was clear enough: “We’re focused on building this thing and then discussions [about M&A] will emerge from that. I have no doubt.” That is the right posture: build the mine, make the project work without a buyer, then, if a buyer comes, negotiate from strength.

Nothing revolutionary. He does not close the door, and he does not open it wide enough to invite a lowball. But the M&A angle is clearly in his head. After hearing that answer, I would assign a higher probability to an eventual strategic discussion than I did before.

If that moment comes, Peterson will be the person responsible for extracting value from the buyer. That’ll be his job.

Then there is the elephant in the room: Vulcan Minerals.

Vulcan remains the anchor shareholder, and its CEO, Patrick Laracy, remains the gravity in the room. Through Vulcan, his personal stake, the board history, and the royalty, he is tied to Atlas in more than one way. Most of the time, that is alignment. A higher Atlas share price benefits him directly. A higher takeout price benefits him directly. A financing done later and at a higher price benefits him directly.

Good. But there is always a “but.”

Vulcan also has the 3% Net Production Royalty to protect in a change-of-control negotiation. A buyer may care about buying out the royalty almost as much as buying the equity. Or it may accept the royalty and reduce the premium. Either way, the deal is not only about Atlas shareholders selling their shares.

This is the one place where Vulcan’s alignment can become less clean. Vulcan wants a high Atlas share price, but it also wants to preserve or maximize the value of the royalty. Those two objectives should usually point in the same direction, but they do not have to. A structure that maximizes Vulcan’s total value through the royalty may not be exactly the same structure that maximizes value for minority shareholders through the equity.

I do not think this is the most likely problem. Atlas shareholders would still have a say in most change-of-control structures, whether through a vote or a tender decision. The board is not only Vulcan. Peterson, Howe, and Kelly matter. And the royalty is already inside the UFS economics, so this is not a hidden bomb under the NPV.

Either way, the negotiation is not as simple as “what is Atlas worth?”

Call it the cost of having the people who found the asset still sitting around the table.

Signals to Watch

First, the obvious: rumors are cheap. In a market this obscure, they are usually worse than cheap. They create the illusion of information.

The first signal to watch is the undisclosed strategic investor. That box is already partly checked. If that investor increases its position, or if other actors in the same category appear, the probability of a future strategic transaction goes up.

The second signal is contract architecture.

Every serious contract Atlas signs, and every MOU it converts, changes the M&A map. A binding Scotwood agreement would validate retail demand, but it would also give Scotwood a stronger reason to move closer to the asset, while making the project less clean for a buyer that wants full control of the retail channel. The same logic applies to port access, logistics, vendor financing, prepayments, strategic equity, and offtake.

In M&A, the first bid often starts as something more boring: volume, capital, logistics, or a right buried in the paperwork.

And yes, there are already early signals. The strategic investor. The non-binding MOUs announced years before production. The fact that several counterparties are willing to spend time and credibility around a mine that will not produce before 2030. Atlas is the kind of asset that should interest a lot of people in the sector. But these are only necessary conditions for M&A, not sufficient ones.

I could keep speculating, but the only point that matters is simpler: M&A is neither necessary nor sufficient for Atlas shareholders to make money. It is just another path. Potentially a very good one, depending on timing, price, and who gets to keep the synergies. But the thesis does not need it.

12. My Final Take

Most investors will analyze Atlas Salt through the UFS and the assumptions embedded in it.

They should. That is the official file. That is where the reserve base, the CapEx, the operating cost, the mine plan, and the C$920M after-tax NPV live. But the UFS answers a lender’s question: can this project be built and repay capital under defensible assumptions?

I am an investor, so my question is different: across the most probable range of outcomes, from very good to very bad, what is the present value of my share?

Answering that question forced me to challenge several assumptions inside the UFS, and that is where I found the hidden value. SLR’s conservatism is my opportunity. The salt price used in the UFS is far below recent market evidence. The assumed product mix looks conservative as well, which means the margin assumption is probably conservative too.

Those two variables influence almost every cash flow the mine will generate, from the first year of production to the last. They are far more consequential than a delay, a CapEx overrun, or the final financing structure, yet they are buried inside hundreds of pages of technical work, treated almost like anecdotal assumptions.

There is value to extract, and Atlas has people who know how to extract it, with every incentive to do so.

Construction risk is also limited compared with most peers, and this is starting to show in the financing file. The theory is not yet the practice. But the fact that Atlas is targeting 60% senior debt, while already having announced a non-binding financing MOU with Sandvik, is a large clue.

Put differently, I think the current valuation is pricing in a level of risk that is materially overstated.109

I have seen this kind of alpha before in micro caps, buried under the unknown and inside the assumptions of an official study. Rarely at this scale. It helps that Atlas operates in a sector unknown to 95% of investors and ignored by most of the remaining 5%.

I am a shareholder, and I intend to remain one. I am willing to hold through the full mine life, unless the market offers me an absurd valuation before then. Of course, I will update my thesis with every new piece of information, and I will sell if the assumptions that brought me here prove wrong.

I will keep sharing my analysis of Atlas Salt on Substack for as long as I remain a shareholder.

Subscribe now

13. A Better Way to Play the Setup

Being a shareholder of Atlas Salt is only one way for me to play this opportunity.

There is another way to play the same underlying situation, with objectively more upside, objectively less risk, but a risk profile that is… different.

I know that may sound bold. But I have the math on my side.

This is literally an opportunity inside an opportunity, and I am pretty sure very few investors have found it. Fewer, in any case, than those who have found Atlas Salt.

The full deep dive is here:

If you have any questions or comments, the comment section and my DMs are open.

Take care,
Flo

Do you think this work deserves to be shared?

Share