Cameco

IThe business

Cameco digs uranium out of the ground in northern Saskatchewan, buys more of it, processes some of it, and sells it to the people who run nuclear reactors. It also owns 49% of Westinghouse, which designs reactors, makes their fuel and services them. So there are two businesses here, and they are valued in different ways. Most of what follows is about keeping them apart.

All company figures on this page are in Canadian dollars, marked C$, because that is how Cameco reports, under IFRS. Uranium prices are in US dollars a pound, because that is how the world trades it. Per-share values are in US dollars against the New York close. The rate I use is 1.396, the Bank of Canada’s average on the price date, and the Toronto close of C$107.60 agrees with it to within 0.3%.¹⁴¹⁵ TTM means the twelve months to 30 September 2026, and this page includes the third-quarter report.⁵

Over those twelve months revenue was C$3.48B, adjusted EBITDA C$1.76B and net earnings C$398M C (calculated).²³⁵ Adjusted EBITDA is Cameco’s own measure. It splits into C$1.25B for uranium, C$181M for fuel services, C$638M for its share of Westinghouse, and minus C$310M for head office C (calculated).⁵⁶ Those numbers flatter the middle of the business, as I show under Financials, because only a small part of Westinghouse’s EBITDA ever reaches Cameco as cash.

Uranium

Three mines do nearly all the work. McArthur River is the world’s largest high-grade uranium mine, with reserves averaging 6.48% U3O8, and its ore goes to the Key Lake mill, the world’s largest. Cameco owns 69.805% and 83.33% of them. Cigar Lake is the highest-grade uranium mine in the world, at 16.33%, and after buying part of TEPCO’s stake in July Cameco owns 57.418%. Its ore is milled at McClean Lake, which Orano owns and runs. Inkai, in Kazakhstan, is a 40% joint venture with Kazatomprom that dissolves uranium out of sand underground.⁶⁷ To put the grades in context, Inkai’s are 0.03%: Cigar Lake’s rock holds more than 500 times as much uranium.⁶ Grade is why Saskatchewan can be cheap even though it is deep, wet, cold and frozen on purpose, since the ground at both mines has to be frozen before it can be mined.

In 2025 those Saskatchewan assets gave Cameco 21.0M lb and it delivered 33.0M lb. The gap matters. Cameco commits to more than it mines and buys the rest. In 2025 it purchased 9.6M lb, including its Inkai share, which equity accounting shows as a purchase.⁶ In 2026 it plans 19.5M to 20.5M lb of production and 29.5M to 30.5M lb of deliveries, and it now expects to pay about C$110 a pound for what it still has to buy.⁷ That is about what it receives.

Why the market is short

A reactor needs fuel years ahead. Natural uranium has to be converted, enriched and made into fuel before it can be loaded, so utilities buy under long-term contracts, and the spot market is a thin side-show. It traded about 9M lb in the third quarter, against about 15M lb a year earlier.⁷ Annual use is much larger: UxC reports about 815M lb burned over five years, about 163M lb a year, and only 589M lb contracted C (calculated).⁶ UxC also puts uncovered needs at about 3.1B lb through 2045.⁶ That is the heart of the bull argument. Utilities have under-bought for years, and at some point they have to come back.

On the supply side, more than two thirds of the fuel supply typically comes from state-owned firms.⁶ New projects keep slipping, and Cameco’s own list for the quarter reads like a catalogue of delay: Kazatomprom’s acid plant pushed back to between late 2027 and early 2028, Deep Yellow deferring its decision on Tumas, Global Atomic’s Dasa now US$777M and second-half 2028, Boss cutting Honeymoon’s ramp, Peninsula withdrawing guidance.⁷ These are the facts of the company’s market update, so a reader is right to note who chose them. The demand side in that update is just as one-sided: 39 countries pledging to triple capacity by 2050, 78 reactors under construction, and Google paying for uprates in Georgia, Finland and at Constellation.⁷

Prices have responded, slowly. The long-term price stood at US$96.50 at 30 September, up US$13.50 in a year, and the spot price at US$89.63.⁷ Cameco says the long-term price is approaching its highest since 2011 in constant dollars.⁷ But long-term contracting in the first nine months was about 42M lb, down slightly from 45M lb a year earlier, and all of 2025 saw 116M lb.⁷⁶ Against 163M lb burned, that is still below replacement. Management’s line is that contracting is rarely linear. It has been saying some version of that for several years, and it may be right, but the buyers have not yet arrived in size.

Fuel services

Cameco refines uranium at Blind River and converts it at Port Hope, Ontario. Conversion turns the concentrate into UF6, the gas that goes to the enricher. Cameco has about 18% of the world’s primary conversion capacity, and few others do this at all.⁸ The segment earned C$181M of adjusted EBITDA over twelve months C (calculated).⁵⁶ It is a small, good business: Orano’s Malvési plant expects about 9M kgU in 2026 against a target of at least 13M, and the long-term conversion price of US$54 a kgU is far above where it sat before the war in Ukraine split the market.⁷ The Port Hope licence expires in February 2027, with hearings this quarter.⁷

Westinghouse

Westinghouse has two parts. Operating Plants serves the reactors already running: it services about 63% of the roughly 417 in operation and fuels about 40%.⁷ That business had a backlog of US$13.2B at the end of 2025 and took US$5.2B of new orders in the year.⁷ New Plants sells the AP1000, the large reactor design, plus the smaller AP300 and the eVinci microreactor, both still on paper or in test. Its backlog was US$0.8B.⁷ Its pipeline lists up to 91 reactors: up to 20 under the DOE and Commerce programmes, 5 at the long-lead stage (2 in South Carolina, 3 in Poland), 4 in early services, 12 in front-end design and 50 in origination.⁷ At 100%, Westinghouse’s revenue was C$5.36B over nine months, against C$2.28B for Cameco’s own consolidated business.²

Cameco and Brookfield bought it in November 2023. Cameco paid US$2.1B for 49%.¹³ Since then three things have happened, and they account for most of what the shares have done. In October 2025 the US government signed a binding term sheet with Cameco and Brookfield to arrange at least US$80B of new Westinghouse reactors; the shares rose 23% that day, from US$86.62 to US$106.91.²⁰¹ In June 2026 the Department of Energy gave a conditional commitment of up to US$17.5B for long-lead parts for up to ten AP1000s.²¹ In July Westinghouse confidentially filed for an initial public offering.²² Bloomberg then reported, citing people familiar, that it wants a valuation above US$50B, with a public filing possible as soon as October.⁹ I searched EDGAR and found no public filing.²³

The rest

Cameco owns 49% of Global Laser Enrichment, with an option to reach up to 75%. It holds the exclusive licence to a laser enrichment process, which has reached technology readiness level 6. It now has to show level 7, and it targets first operations by 2030, subject to a final decision. Cameco spends C$50M to C$55M a year on it as research.⁶⁷ It also has idled mines in Rabbit Lake and the United States that cost C$62M to C$67M a year to keep on care and maintenance.⁷

IIThe moat

A durable business keeps others from taking its profit. Cameco’s defence is in the rock and the permits. Over the last decade it was also tested hard, and the record is worth reading before the argument.

The long-term uranium price fell from US$39 in 2016 to about US$30 in 2018, and spot averaged US$21.78 in 2017.¹⁹²⁴¹⁸ From 2016 to 2025 Cameco’s adjusted net earnings totalled C$1.72B, about C$172M a year, and it lost money on an adjusted basis in 2020 and 2021 C (calculated).^4101617 A business with pricing power does not behave like that. In the slump its only real defence was to stop mining. It shut McArthur River and Key Lake from 2018 to 2022 and bought pounds in the market more cheaply than it could mine them.¹⁸¹⁶

That choice is the strongest thing I can say about the company and also the weakest. It is strong because few miners have the discipline to close their best asset for four years. It is weak because it shows what the asset was worth in a bad market: less than the market price of what it produced. So the value here depends on the uranium price more than on any one thing Cameco controls.

What does protect it

• Grade. The Athabasca Basin holds the richest uranium in the world, and Cameco controls the largest high-grade reserves. Its share of reserves was 433M lb at the end of 2025, and of measured and indicated resources a further 404M lb.⁶ Ore this rich can pay for mining methods nobody else would use.
• Permits and time. The McArthur River and Key Lake licences run through 2043 and Cigar Lake’s through June 2031.⁶ New mines take many years to license and fund. Deep Yellow, with a 3.6M lb a year project, could not finish its funding in October.⁷
• Where it is. A utility in Ohio, Korea or India that can’t buy from Russia and worries about Niger has few other large places to go. Niger took the SOMAÏR mine from Orano and has awarded the permit to a state company.⁷ India signed in March for nearly 22M lb over nine years, about C$2.6B, on market-related terms.²⁵
• Conversion. With 18% of primary capacity and few rivals, this part has real scarcity.⁸
• Westinghouse’s installed base. A reactor is licensed around its fuel and systems, and changing suppliers is slow. Fuel and outage work repeat for decades. This is the most dependable income Cameco has a share in, and I would call it the only true annuity in the group.

Where it is thin

The reserves don’t last forever. Cameco’s own estimates end Cigar Lake in 2036 and McArthur River in 2044; Inkai’s licence runs through July 2045.⁸ In 2025, production removed 25M lb of reserves and they fell to 433M lb from 457M lb.⁶ That is about 17 years at the 25M lb that 2025 took out, and an honest valuation has to pay something to replace them, which mine does at the end.

And costs are not fixed. Life-of-mine operating cost estimates in the latest annual information form are C$21.72 a pound at McArthur River and Key Lake and C$23.94 at Cigar Lake, and Cameco expects the McArthur figure to change with its new technical report.⁷ The reported cash cost for produced pounds was C$29.70 over nine months and C$40.73 in the third quarter, inflated by the Key Lake outage, and the company has now raised 2026 unit cost of sales guidance from C$61.50 to C$65.00 at the start of the year to C$67.00 to C$70.00.⁵⁷⁶ Saskatchewan also takes a growing cut. Basic royalties and the resource surcharge come to 7.25% of sales, and a profit royalty rises to 15%.⁶ Kazakhstan moved to a tax on extraction of up to 18% this year, up from 9%, with up to 2.5% more depending on the price.⁶ Governments are partners in this business, and they get richer first.

IIIThe competitive advantage

The question is what Cameco can do that rivals can’t. On cost it isn’t clear that the answer is much. On the numbers each side publishes, the two biggest listed producers are close.

Kazatomprom’s scale is roughly double. Its uranium comes out of the ground by pumping acid through sand, which needs no shafts, and its costs have been rising fast too: first-half cash cost was up 37%.¹¹ It depends on sulfuric acid, on long transport routes, and on a state that can change the tax, as it just did. For a western utility, that is the opening Cameco lives in. Over the years, buyers have been willing to pay for safe supply in conversion and enrichment, and management says that focus is now moving up to uranium.⁷

The second part of the edge is the book. Cameco holds commitments of about 230M lb with 39 customers, the five largest being 56%.⁸⁶ It has spent years telling customers it would not sell at any price, and I think that discipline is real. The 2026 volume cut, from 33.0M lb to about 30M lb, is a choice: it mines less and buys less at high prices rather than chase deliveries.⁵⁷ New contracts layer in market-linked prices with floors and higher ceilings.⁷

The third part is that nobody else owns both ends. A miner with a stake in the world’s main reactor servicer gets paid at two points on the same chain. Whether that integration is worth anything beyond the sum of the two is doubtful. Utilities buy uranium, conversion, enrichment and fabrication separately. Cameco does sell to Westinghouse, on normal terms, but that is a line in the related-party note and not a channel.² So I treat the two as separate holdings.

New supply

Rivals in waiting are mostly in the same basin. Orano and Denison have begun producing at McClean North with a new method, and other projects in the basin are in permitting.⁶ If they get built they add low-cost western pounds, which is the exact product Cameco claims is scarce. I have not modelled them one by one, but they are why my base price rises only with inflation. The Athabasca has more to give at the right price.

The AP1000’s real rivals

Westinghouse’s competition is different. Its large reactors face Korea’s APR1400 family, France’s EPR, and in some markets China’s and Russia’s state builders, plus every small reactor start-up. Korea’s KHNP won the Dukovany project in the Czech Republic, and Westinghouse took a part in it, which added about US$170M to Cameco’s share of revenue in 2025.⁷ The Korean framework of September is an odd case. It would finance up to eight US reactors, six AP1000s and two APR1400s, with Westinghouse taking about US$2B a reactor on the latter.⁷²⁷ The best thing Westinghouse has is that its design has run: it is the only generation III+ pressurized water reactor in service with fully passive safety, and the NRC renewed its certification this month.⁷ Its worst thing is what the first ones in Georgia cost, which the company itself lists as the base for its near-term costs of US$20B to US$26B for two units.⁷

IVManagement

Tim Gitzel, 63, has been chief executive since July 2011, so he has run Cameco through the long slump and the recovery. He joined from Orano’s mining business and sits on the boards of the World Nuclear Association and the Nuclear Energy Institute.¹² In September 2025 Grant Isaac, long the chief financial officer and the person investors know best, became president and chief operating officer.¹² That has the look of a succession, though no one has said so. Heidi Shockey became chief financial officer the same day and plans to retire: Leah Hipperson takes over on 1 January 2027.⁵ That makes her the third chief financial officer in under two years, which is worth watching.

How they are paid, and what that pays them to do

Gitzel’s total direct pay for 2025 was C$9.89M: C$1.275M of salary, a C$2.62M bonus and C$5.99M of long-term awards, 60% performance share units and 40% restricted.¹² For 2026 the board raised his salary 14%, his target bonus to 140% of salary, and his long-term target from 470% to 600% of salary, after raising it from 370% a year earlier.¹² Directors explain this as a move above the median of the peer group. Peer groups rise when peers raise pay, so I read that as a pay rise.

The annual bonus is 50% financial, split equally between adjusted net earnings and funds from operations, with the rest on safety, environment and community goals.¹² In 2025 adjusted net earnings of C$627M beat the C$555M maximum and funds from operations of C$1.20B came in near the C$1.215B maximum.¹² Those targets were set before the Dukovany payment, which added about US$170M of revenue to Cameco’s share of Westinghouse in a single quarter.⁷

The performance units for 2025 tested three things over three years: 35% realized uranium price against two industry benchmarks, 35% production against budget, and 30% adjusted EBITDA growth against budget.¹² For 2026 the production test is gone. The weights are now 50% relative realized price, 40% EBITDA growth against budget, and 10% milestones for growing Westinghouse.¹² The stated reason is that production affects the EBITDA measure anyway. That is true. It is also true that production has missed the plan at McArthur River two years running, and that the 2023 award’s operations score was set at threshold only after the committee adjusted the cost measure for strategic decisions.¹²⁶

So what does this pay them to do? Get a better price than the industry, which is what the contracting discipline is, and that seems right. Grow EBITDA, which includes a share of Westinghouse’s EBITDA that doesn’t arrive as cash. Advance Westinghouse. I found no measure of return on capital and none of free cash flow per share. A team paid this way is well aligned with a higher uranium price and with an IPO. It is less well aligned with the question an owner asks, which is how much cash each dollar spent will bring back.

How much of their own money is in

Quite a lot. At the end of 2025 Gitzel owned 532,211 shares, worth C$66.9M at the year-end price, more than 52 times salary against a requirement of five, with no change on the year.¹² Add restricted and performance units and it was C$88.0M. He also filed two Forms 144 in January 2026, giving notice of a proposed sale of 75,000 shares, worth about US$7.5M at the time.²⁸ A Form 144 is a notice, not a record of a trade, and Canadian insiders report completed trades on SEDI, which I did not check. Two other officers filed notices for 7,000 shares between them.²⁸ Those are modest against his holding, and the January timing was close to the top, which says little about intent.

The company’s own five-year table is more telling. For the 2021 grant, C$100 of chief executive pay became C$195 by the end of 2025, while C$100 of shares became C$747.¹² Shareholders did better than he did, which is not always so in a proxy. The say-on-pay vote at the 2025 meeting was 97.19% in favour.¹²

What they did with capital over a full cycle

The period from 2016 to 2025 covers a slump and a recovery. Cameco produced C$5.92B of operating cash flow, spent C$1.48B on plant, paid C$760M of dividends and bought back no shares C (calculated).^41016171819 It issued C$953M of shares in 2022, and weighted shares rose 10% from 395.8M to 435.4M.¹⁶⁴ Total debt fell from C$1.49B to C$1.0B.¹⁹⁴ The cut in the dividend from C$0.40 to C$0.08 in 2018 was the right thing to do, and doubling it to C$0.24 since 2023 is proportionate.¹⁸⁶

On the whole this is a good record of a patient team. Management did the hard things at the bottom and bought the best asset of the decade when nobody wanted it. Two caveats. The company’s statements about what it will do next, such as positioning McArthur River for its licensed 25M lb, depend on delivery the last two years have not shown.⁶ And Cameco has run ten years without a share buyback, so we have no record of how it treats a large cash surplus. The Westinghouse IPO may be the first test.

VFinancials

Reading Cameco’s statements takes some care, because three things in them don’t mean what they seem to.

• Revenue contains purchased pounds. A pound bought at market and sold at contract adds revenue and almost no profit. So revenue growth says little on its own.
• Westinghouse is one line. Cameco books its share of Westinghouse’s net result, a loss of C$80M over nine months, because purchase accounting amortizes the intangibles. The company prefers adjusted EBITDA, C$427M for the same period.⁷ Neither number is cash.
• Inkai shows up as a purchase and a dividend. Cameco buys its share at a 5% discount to spot, and gets the profit back later as a dividend, net of tax, usually in the following year.⁷

The third quarter

The third-quarter report explains the fall in the shares. Revenue was C$619M, about flat. Gross profit fell 28% to C$122M, and adjusted net earnings were a C$2M loss.⁵ Cash from operations was C$3M, and for nine months it was C$112M against C$731M a year earlier.⁵ Unit cost of sales in uranium was C$72.17 a pound, up 21%, with the Key Lake outage running through it.⁷ Management narrowed 2026 production to 19.5M to 20.5M lb, the bottom half of its old range, and raised unit cost guidance by about C$3 a pound.⁷ It says this was planned. Most of it was. But the outage carries over into the fourth quarter, and the company warns that restart problems could cut production further.⁷ That leaves about 6.4M lb to make in the last quarter C (calculated).

What an owner can really take out

I start from operating cash flow, which already includes what Inkai and Westinghouse paid out, and already bears the cost of share units, because Cameco settles them in cash or buys shares in the market rather than issuing new ones.¹² Cash-settled stock pay was C$71M in 2025 and equity-settled C$15M.⁴ So I don’t subtract it twice. I take off capital spending, and interest paid, which Cameco puts under financing.

That gives C$1.02B for 2025 and C$308M for the twelve months to September C (calculated). Against net earnings of C$398M and adjusted EBITDA of C$1.76B, it means less than a fifth of the headline measure reached an owner C (calculated). The market value is C$46.8B, so the owner yield is about 0.66% C (calculated).

Neither number is normal. 2025 was too high: it had a US$171.5M Dukovany distribution from Westinghouse, and low cash tax of C$90M. The last twelve months are too low: cash tax was C$295M because the CRA dispute has drained Cameco’s tax pools and pulled cash payments forward, and working capital took C$274M in nine months.⁷² If I normalize, taking off the one-time Dukovany cash from 2025 and using a cash tax near what the current profit would bear, I get about C$750M a year, or about 1.6% of market value E (estimate). For comparison, the ten-year sum of owner cash is C$3.83B, about C$383M a year.

The gap at Westinghouse is the wider one. Cameco’s share of adjusted EBITDA was C$780M in 2025. Its cash receipts from Westinghouse in 2025 were US$220.5M, about C$305M, and US$171.5M of that was Dukovany C (calculated).⁶³⁰ The ordinary distribution has been US$49M a year.³⁰ That is about 9% of a C$780M number C (calculated). Westinghouse carries about US$3.25B of net debt at 100%, capital spending of US$330M to US$410M, interest of about US$270M, and expenses left out of adjusted EBITDA C (calculated).²⁷ So a stake that the IPO talk prices at more than US$20B has paid its owner the equivalent of US$49M a year in normal times.

Balance sheet

Cash was C$860M and debt C$997M at 30 September, so net debt was about C$137M.⁷ There is C$400M due in October 2027 at 2.95%, a C$1.0B undrawn revolver and C$1.6B of financial assurances outstanding.⁷⁶ Ratings are BBB, Baa2 and BBB.⁶ Westinghouse’s debt doesn’t have recourse to Cameco.⁶ Reclamation and waste provisions are C$979M on the balance sheet and about C$1.37B undiscounted.²⁶

The CRA dispute is the odd item. Since 2008 Canada’s tax agency has challenged Cameco’s marketing structure. Courts ruled for Cameco on 2003, 2005 and 2006, and the Supreme Court refused to hear an appeal in 2021.⁷ CRA still holds C$559M of security, C$209M in cash and C$350M in letters of credit, for later years, and has shifted its argument for 2014 to 2019. Cameco has also paid C$346M of cash tax for 2024 to 2026 because of that reassessment.⁷ I leave all of this out of my values. If Cameco wins everything and gets back the C$209M held in cash and the C$346M it has paid, that is about US$0.90 a share C (calculated), and it is not going to be resolved soon.

How the shares have traded

The stock rose from US$5.50 in March 2020 to US$134.09 in January 2026, 24 times C (calculated).¹ It has lost 43% since then. The 7.9% fall on the day of the third-quarter report took it to its lowest close of the past year.¹ The rest of the sector has fallen further, so this is not only about Cameco. Uranium Energy is 62% below its 52-week high, Centrus 74%, Oklo 82% and BWX Technologies 46% C (calculated).³² NexGen is down 41%. The uranium price had not fallen as of 30 September, the latest figure I have. What fell was the premium the market was willing to pay for the idea of nuclear growth, and any Westinghouse IPO has to price into that market.

VIValuation

First, how this goes badly wrong

Before the numbers, I asked what would most likely ruin this investment, and worked back. Three things stand out.

• The reactors don’t get ordered. Everything said about new build rests on documents that are conditional, non-binding or both. The US partnership is a term sheet, and the government’s participation interest ends unless it makes a final investment decision on at least US$80B before January 2029. The Korean framework is non-binding. The DOE loan is a conditional commitment.⁶²⁷²¹ I can find no AP1000 in the company’s western pipeline that has reached a final investment decision. The first Polish unit is due in 2036, and a near-term pair costs US$20B to US$26B.⁷ Elections, interest rates or one bad overrun can stop all of it. If they do, a stock priced for a nuclear revival loses that value quickly, and it has already lost part of it.
• The buyers stay away while the book runs off. Cameco has about 28M lb a year committed through 2030, but less in 2029 and 2030. The lower years are where the old contracts stop protecting it.⁷ If utilities keep waiting, and if Kazatomprom, the financial holders of 78M lb in the Sprott trust, or new mines fill the gap, then the term price rolls over just as Cameco’s floors roll off.⁶ It has happened before: from 2015 to 2018 the term price fell by a third, from about US$46 to about US$30.¹⁹¹⁸
• A single point fails in Saskatchewan. Most of the company’s value passes through one mill at Key Lake, another at McClean Lake run by someone else, and one road. In May a flood took out a bridge on the supply route and stopped Key Lake. In July an acid plant failure at McClean Lake suspended Cigar Lake for two weeks. This quarter Key Lake is in an outage planned to run longer than in past years, and it has carried into the fourth quarter.^332634357 Each was small. The mines also carry the old risk of water, and Cigar Lake’s frozen ground is there because the rock above it is full of it. Cameco has no large second site.

None of these needs anything new to happen. They set my bear case. They also explain why I start from the cash that actually comes out of the two businesses, and don’t just apply a multiple to EBITDA.

How I built it

One discounted cash flow doesn’t work here, since Westinghouse is equity-accounted and pays out rarely. I value the parts separately. Uranium and fuel services come from a pounds model: sales, Cameco’s own table of realized prices, production at a cash cost, Inkai bought at 95% of spot, the rest bought at spot, Saskatchewan royalties by formula, Inkai dividends after the new extraction tax, fuel services, head office, capital spending at the company’s plan, and 25% tax. After 2036 I charge 25% more capital to pay for replacing Cigar Lake’s pounds. I then deduct all C$979M of reclamation provisions, and I leave the annual reclamation cash out of the flows so it isn’t counted twice.

Westinghouse is built at 100% in US dollars. Operating Plants starts near US$790M of EBITDA in my base and grows 7% a year, then 4%. New Plants is pairs of reactors started each year. Westinghouse’s own framework says a near-term pair is worth US$8B to US$11B of work to it at a target margin of about 20%, and an nth-of-a-kind pair US$6B to US$8B.⁷ So I use US$1.9B of EBITDA a pair for the first five, and US$1.4B after that, spread over ten years. I take off US$100M a year for the costs it leaves out of adjusted EBITDA, plus capital spending of about US$430M, tax at 23%, and working capital, and subtract about US$3.25B of net debt at the end. If the US warrant vests, the government takes 20% of value above US$17.5B, and Cameco gets 49% of what’s left.⁶

What the price implies

At $76.91 Cameco has a market value of US$33.5B. That is 94 times trailing adjusted earnings per share, and about 27 times trailing adjusted EBITDA, which includes a share of Westinghouse’s C (calculated). Multiples of a year like this one mean little, so I ask the sum-of-the-parts question in reverse. If I take my base for uranium and fuel, what must Westinghouse be worth?

This is the most useful table on the page. The price almost exactly equals my uranium value plus US$50B of Westinghouse with no deduction for the US warrant. The price is consistent with something fairly simple: the reported IPO number taken nearly at face value, with uranium at about my base. It is equally consistent with a discount on Westinghouse and a richer value for uranium, and I can’t tell which. Either way, the market isn’t paying for the current cash flows of either business.

The US$50B itself deserves a hard look. Westinghouse’s 2026 adjusted EBITDA guide, at 100%, is about US$816M. At US$50B of equity value plus net debt that is 65 times.⁷ I can get there at my 10.0% rate only with about 10 reactors a year for ever, starting from a programme with no reactor at final decision. Or with a much lower rate. A listed market in a good mood will pay for such a story. It did for Oklo, now 82% below its 52-week high.³² What a market pays and what an owner gets out are two different numbers, and I care more about the second. But I can’t rule out the first, and the IPO may be close. If Westinghouse does list at US$50B and stays there, that would add about $17 a share to my bull, and $31 to my base.

The figure also rests on one Bloomberg report that quotes unnamed people, and I could read only its opening paragraphs.⁹ It says “valuation” and does not say whether that means equity or equity plus debt. I treat it as equity, which is the generous reading. If it includes the US$3.25B of net debt, equity is nearer US$46.75B, and that takes about $3 a share off the price this table reproduces E (estimate).

A listing is not a given either. Holtec Nuclear filed for a US listing on 10 July and asked to withdraw on 25 September, saying only that it no longer wished to offer securities at this time.⁴⁶ One case says little, but it shows a nuclear listing can be pulled, and Westinghouse has yet to file in public.

One more item is in none of my cases. The Korean framework provides for Korea to take a cornerstone stake of 5% to 10% in Westinghouse.⁴⁷ Neither the price nor whether the shares are new is stated. New shares sold below the IPO price would dilute Cameco’s 49%, so I count it as a risk and not a value.

There is also a holding-company point. Even if Westinghouse lists at US$50B, Cameco’s stake won’t turn into cash on the day. Cameco would own a large listed minority stake, with lock-ups, and tax on any sale, inside a company that also runs mines. On this site Prosus shows how far that can go: it costs about US$0.81 for each US$1 of Tencent. I don’t apply a discount like that here, but a reader might.

What I can’t judge

Two inputs decide this page, and I can’t know either. I don’t know what uranium will cost in 2032. I don’t know how many AP1000s get built. Nobody does. So here is the grid. Read my $39 as one cell in it.

A US$20 rise in the uranium price adds about US$7 a share at the middle of the table. A US$10B rise in Westinghouse adds about US$9. Those two are about equal bets, and they are not independent. A real nuclear build-out helps both. A failure hurts both. So the two parts are one thesis held twice, which is the main reason for care.

The discount rate

I use 10.0%. The ten-year Treasury yields 4.94%, Canada’s 3.84%.³⁶³⁷ With Damodaran’s 4.20% implied equity risk premium and a five-year beta of 1.18 on month-end returns against SPY, the capital asset pricing model gives 9.9% C (calculated).³⁸¹³⁹ I use US rates since the product is priced in US dollars. That is above the 9.0% on my Mastercard page and below the 10.5% on Broadcom, which seems right for a company with contracted volumes but an unknown price. I use the same rate for Westinghouse, though Operating Plants deserves less and New Plants more.

Three cases

The most important modelling fact is that higher prices arrive late. Here is the company’s own table, which it updates each quarter. It shows what it would realize on its committed book at a given flat spot price.

At US$160 spot, 2027 gets US$77. That’s ceilings at work. At US$40 it gets US$47: floors. So the book cuts both ways, and it makes this a bond-like business in the near term and a commodity business only from about 2029. My base realized price reaches the market in 2031, when old contracts have largely run off. In 2027 my base EBITDA barely moves, because Cameco still has to buy pounds at about US$100 to deliver into contracts near US$73. That squeeze is the reason current earnings look weak, and it is also why they can improve without any change in the market.

A rough check on uranium. In July Cameco paid about C$23 a reserve pound for TEPCO’s share of Cigar Lake.²³¹ At that rate its 433M lb of reserves come to about US$7B, or US$16 a share, before any resources, conversion or the Inkai cash C (calculated). My base uranium value of $20 a share sits a little above it. It isn’t a clean comparison, because a partner’s small stake sold to other partners is a thin market. But an informed seller did take that price this summer.

Other uses of the same dollar

The ten-year Treasury pays 4.94% with no risk to the money R (reported). My base here returns 2.2% a year at this price, below the Treasury. The bull returns 9.9%, about what I require. The bear loses 8.3% a year E (estimate).³⁶ The closest structure on this site is Prosus, a listed business with one enormous stake inside: it trades 22% below my base value, and the stake is already public. If the appeal here is an asset with a growing list of nuclear customers, the fair comparison is a company with a similar lock on its market and a known record of cash, such as ASML at 7.4% a year on my base, Broadcom at 7.5%, or NVIDIA at 9.7%. Those pages use their own rates and dates. On my numbers the same dollar earns more elsewhere unless the bull case happens.

The margin of safety

• Price$76.91Close, Q4 2026 [1]
• Base value$39My estimate E (estimate)
• Margin of safety below$17set from my bear case C (calculated)

I set the margin of safety from the bear case, because with a price this uncertain a discount to base is no protection. Below $17, 57% under my base value, my $13 bear value is at least three quarters of the price paid. That price describes a market that has given up on nuclear altogether. The stock last closed there in August 2021, at US$16.82 or below, when the term price was under US$40.¹²⁹ It isn’t a level I expect to see offered, and there is no margin of safety at $76.91. Between that and my base is a wide middle where a reader has to take a view on one of the two things I can’t judge.

Two different questions

How good is the business, and how good is the price? On the first, the assets are the richest uranium deposits on earth in a safe country, 18% of conversion, 49% of the company that services most of the world’s reactors, and little debt, under a team that has been patient for fifteen years. It is still a business that earned C$172M a year in adjusted profit over the last decade, and that depends on a price it doesn’t set. On the second, the price has fallen 43% and at 97% above my base it still pays for a reactor programme that doesn’t exist yet. This looks like a fine company at a price that needs a very large amount to go right. A fall in the share price doesn’t change that by itself, though another one like it would start to.

Where this could be in 5 and 10 years

For my base to be right at five years, Westinghouse has to have started about five AP1000 pairs by 2031, and its EBITDA at 100% has to reach US$1.8B. Uranium stays flat after inflation, near US$108 a pound in 2031. Owner cash from uranium and fuel is then about C$1.3B a year C (calculated). At ten years it needs 12 pairs, Westinghouse EBITDA of US$3.6B and owner cash of C$1.4B E (estimate). Westinghouse is then 56% of the value, against 46% today. The number rises because the cash comes nearer, at 10% a year, and not because I add anything.

The bear and the bull differ on the uranium price and the pairs. In the bear the term price is US$68 in 2030 and US$77 in 2036, with five pairs in ten years. In the bull it is US$150 and US$174, with 24 pairs. The discount rate is 10% in all three. At 11% my base is $33 today, and at 9% it is $46.

The most likely way my base is wrong is the build-out. I can find no AP1000 in the western pipeline that has reached a final investment decision. If orders slip by years, the five-year value moves toward the bear’s $15. The price is above my base value at both dates.

These are estimates of value, not forecasts of the share price. The market price can sit above or below value for years.

VIIThe case for the price

What most people miss

Three things cut in Cameco’s favour and get less attention than the headlines. First, today’s earnings carry contracts signed years ago, and those are rolling off on a schedule. At a flat US$100 spot, Cameco’s own table goes from US$73 in 2027 to US$88 in 2030.⁷ That is US$15 a pound on about 28M lb, about C$590M a year, with no change in the market C (calculated). It is a third of trailing adjusted EBITDA. Second, every pound McArthur River mines replaces one bought at about C$110.⁷ Getting from 14.5M lb to 18M lb, Cameco’s share of about 2.4M lb, saves about C$190M a year at a C$80 gap C (calculated). Third, Westinghouse is working on 28 of about 30 US uprate applications, a business that needs no new reactor.⁷

The bull case

In plain words: the world burned 815M lb in five years and contracted 589M lb. Utilities have to refill, new mines are late, and the safe supply is in Saskatchewan. On top of that, the western world has picked one large reactor, and Cameco owns nearly half of the company that designs it and services most of the fleet.⁶⁷

Management makes the case calmly. Tim Gitzel said on 30 October that “customer requirements do not disappear”, and that fuel-cycle capacity remains constrained, and added that AI demand may be extra upside but the thesis does not depend on it.⁵ Christopher Souther of Truist cut his target to US$125 from US$130 on 22 October, as reported by TIKR; I have not seen the note.⁴⁰ TIKR puts the average analyst target at $128, from the 21 analysts it tracks.⁴⁰ Among holders, ARK held 361,570 shares and Sprott Inc. 214,954 at 30 June, and BlackRock reported 5.02% in July.⁴¹⁴²⁴³ Those are holdings, and they aren’t stated reasons. None of seven well-known funds I checked listed Cameco at 30 June: Altimeter, Atreides, Coatue, Duquesne, Appaloosa, Situational Awareness and Berkshire Hathaway.⁴⁸ The one owner I found making the case in his own words is a partner in Westinghouse. On 30 September Connor Teskey, chief executive of Brookfield Asset Management, called the US “one of the most attractive markets in the world to build new nuclear generation”.⁴⁷ He has money in it, so I treat that as an owner’s view and not a neutral one.

Against reality

Today’s numbers support the first half. The term price is up US$13.50 in a year, Saskatchewan has reserves, and the book gives a floor: at US$40 spot, 2030 still realizes US$54.⁷ What’s missing is the buyer. Long-term contracting was 42M lb in nine months, on pace for well under the 163M lb burned each year C (calculated).⁷⁶ Cameco says several utilities have requests for proposals out with responses due before year-end, which is the first place I would look for a change.⁷ Production is short of plan, and unit costs are rising. And no western reactor order is final.

If Cameco delivers 18M lb at McArthur River, if the term price stays above US$100 in today’s money, if the US orders reach US$80B before January 2029, and if Westinghouse lists near US$50B and stays there, then $76.91 is possible. My grid says so. Today, against that list, the term price is close at US$96.50 and the other three have not happened.

Multiply it out. Westinghouse lists at or above US$50B: 35%, my guess, in a market where Oklo is 82% below its high. It stays there after the lock-ups: 50%. Uranium stays at about US$100 or better in real terms: 60%. McArthur River gets back to plan: 60%. Together that is about 6% E (estimate). They overlap, because a strong nuclear market helps all four, so it is perhaps 15%. Each one looks reasonable alone. All four together do not. The 25% I give the bull in my weighted value of $42 is, if anything, generous.

What a 10x would take. Ten times $76.91 is $769, about US$335B of market value. At 20 times earnings that needs about US$17B of net income. At a 25% net margin, above Cameco’s best year of the ten (17% in 2025), that is about US$68B of revenue, 27 times the US$2.5B Cameco books today C (calculated). The whole world’s uranium bill, 163M lb at US$96.50, is about US$16B C (calculated).⁶⁷ So Cameco would need to earn more than every reactor on earth now pays for uranium. Even in my bull, with uranium above US$170 and Westinghouse at six reactors a year, 2036 cash for the two together is about US$4B E (estimate). By 2036 it would mean 26% a year for ten years. Tenfold from this size has happened. NVIDIA was worth about US$50B in November 2016, 539.5M shares at US$93.36 before later splits, and it has since risen about 88 times C (calculated).⁴⁴⁴⁵ But chips were a market that could grow many times over. Cameco’s limit is the size of the market it sells into, about 15% of a US$16B bill, plus half of a reactor maker. Tenfold from here is close to impossible, and I would not build anything on it.

Where I come down

The book’s roll-off moved me: without it my 2030 figures would be far lower, and it is why near-term earnings understate the business. The reported IPO value did not, because it hasn’t been tested in a market that has just fallen hard. Real orders would raise my Westinghouse value most: one more pair a year adds about US$10 a share to my base E (estimate).

VIIIRisks, and what would change my view

The three ways this goes badly wrong are at the start of Valuation. These are the rest.

• A reactor accident. Anywhere, any design. Public support for nuclear can change overnight, and Cameco’s demand with it.
• Kazakhstan and transport. Inkai pays Cameco a share of a lower-cost business, in a country that raised its extraction tax this year, through routes the company says carry risk. Dividends can be delayed or kept back.⁶⁷
• Trade policy. US tariffs on uranium remain uncertain, and a Section 232 proclamation of January 2026 reserves the right to tariffs or price floors on processed critical minerals, uranium included.⁶⁷ Most of Cameco’s pounds cross that border.
• Purchases. Cameco is short pounds each year. Spot rising US$5 would cut this year’s cash flow by about C$17M and adjusted earnings by C$7M, because ceilings cap sales while purchases are at market.⁷ It is odd for a uranium miner to lose from a higher uranium price in the short run. It has 4.0M lb of product loans drawn too.⁷
• Customers. Five customers are 56% of commitments.⁶
• Currency. Costs are in Canadian dollars and sales in US dollars. Cameco has hedged US$1.45B for 2026 at 1.37, and a weaker US dollar would cut margins.⁷
• The US warrant. Its terms are binding in a term sheet, and definitive agreements are still to come. They could end up different from what I model.⁶
• Westinghouse’s debt and spending. AP300 design could take US$0.8B to US$1.2B more, eVinci up to US$1.2B, and each DOE-backed project needs up to US$500M of Westinghouse equity before a final decision.⁷ Those are cash calls on a business that pays its owners little.
• Tax. The CRA case can still go the wrong way for later years, and the agency has changed its argument twice.⁷
• Ownership limits. Federal law restricts how much of Cameco any one holder may own: non-residents to 15% each and 25% of votes.⁸ That makes a takeover at a premium unlikely.
• Port Hope. The conversion plant’s licence expires in February 2027, and hearings are this quarter.⁷