November 15, 2026.
Introduction: The Uranium Investment Paradox
Nuclear power is having a strong comeback, and in my reading it is winning the argument. Reactors are being built in China and India, life extensions are routine, Japan is restarting plants, and the United States has put government money behind a large new Westinghouse program. The World Nuclear Association counts 443 operable reactors with 405,667 megawatts of net capacity, 82 under construction and 119 planned, on a table it updated on November 11, 2026. A strong industry, however, does not make every uranium investment a good one.
Uranium equities have not rewarded it lately. The US-listed Global X uranium ETF (URA) closed at USD 36.55 on October 28, 2026 and USD 32.34 on November 6, 2026, a fall of 11.5%. Cameco closed at USD 83.49 on October 29, fell 7.9% the next day after its third-quarter results, and closed at USD 70.73 on November 13. Meanwhile Cameco’s month-end indicators for October 30, 2026 were USD 91.55 per pound for spot uranium and USD 98.00 for long-term contracts.
That gap is the paradox, and it is why I keep four questions apart. Will nuclear power grow? Will that raise uranium demand? Will certain businesses benefit? And will buying them at today’s prices earn a satisfactory return? A yes to the first three does not carry over automatically to the fourth. My research supports the first two strongly, the third selectively and the fourth only weakly.
To test the fourth question I use a hurdle of 9.5% a year in nominal Canadian dollars over ten years. That is my own discipline at Sovereign Canadian, not a market consensus, and you may set a different bar. I applied the same bar to every security and scenario.
The short answer, at the research cutoff: using closing prices dated November 6 or 13, 2026, no security I valued passes the full test for a Buy, which needs an expected return above 9.5%, a margin of safety below fair value, and acceptable downside and confidence. One, Ur-Energy, clears the bare 9.5% return test on its November 13 close, narrowly once cash is assumed to earn less than 9.5%, but its bear case is a total loss and the price is about 14% above my Buy price. Kazatomprom comes within about 5% of the bare return test on my most generous cash assumption, and further away on more cautious ones (Part 11). For most of the sector, the price would have to fall by a third to a half before the numbers work. Waiting, with a written list of prices at which I would act, is a legitimate answer.
Part 1: Why Nuclear Energy Matters Again
Four forces explain the revival. The first is life extension: extending an existing reactor’s operating life avoids most of the cost and time of building a new one, and most of that demand is already in reported fuel requirements. The second is construction, led by China. Its State Council approved eight more reactors on July 31, 2026, at a reported cost of about CNY 160 billion, and the country has roughly 62.5 gigawatts operating against a 2030 target of 110, according to Caixin on August 1, 2026. The Association’s table shows 82 reactors under construction worldwide.
The third force is Western policy. In October 2025 the US government and Westinghouse’s owners announced a term sheet for at least USD 80 billion of new reactors. Cameco’s third-quarter report says the Department of Commerce will receive a participation interest in Westinghouse that vests when the government invests at least USD 80 billion in US reactor projects before January 2029. That investment is a condition, not a given. The fourth force is the hope that small modular reactors and data-centre demand will create a new class of buyer. I treat that as real but late. Uranium demand from small reactors before about 2032 is negligible, and listed small-reactor companies have fallen sharply this year, so my model counts them only in the bull case.
Cost overruns, delayed financing, political reversals and competition from solar and batteries can slow the revival, but they matter less for uranium than for reactor builders, because the existing fleet carries most of the fuel demand.
Part 2: Understanding the Uranium Market
The mine sells uranium oxide (U3O8), measured in pounds, which must be converted, enriched and fabricated into fuel. Conversion and enrichment have been tighter than mining since 2022, and the United States still imports most of its enriched uranium.
Two prices matter. Spot covers small volumes for near delivery. The long-term price covers multi-year contracts, where most volume trades. Cameco’s series shows spot at USD 91.55 and long-term at USD 98.00 on October 30, 2026. The long-term price has climbed from USD 86.50 at the end of 2025 and is above its December 2007 peak of USD 95.00 in nominal terms. Spot peaked at USD 136.00 in June 2007 and has not closed a 2026 month above USD 100.
Producers do not receive spot. They receive their contract book. Cameco’s realized price was USD 66.64 for the nine months, while month-end spot averaged about USD 87. Kazatomprom’s own realized price was USD 79.45 in the third quarter against USD 64.51 in the first half, so its contracts are repricing faster. US utilities, per the Energy Information Administration, bought 87% of their 2025 uranium under long-term contracts at an average of USD 55.91 per pound and paid USD 76.01 for the 13% bought on spot contracts. Their blended cost was USD 58.46, far below today’s indicators, because old contracts are still running off.
Financial buyers matter in a small, opaque market. The Sprott Physical Uranium Trust cannot issue units below net asset value, and its 15.18% discount on November 13, 2026 means it cannot raise money to buy more uranium, a sign financial buying has stalled.
Part 3: Global Uranium Demand Through 2036
The Association puts 2025 reactor requirements at 68,920 tonnes of uranium, about 179 million pounds of U3O8. The United States needs 19,011 tonnes, China 13,872, France 8,389, Russia 6,251 and South Korea 4,703.
I built my own forecast by region. It is a set of judgments, not a reactor-by-reactor schedule, and it sits below the Association’s reference path. My base case reaches 99,900 tonnes in 2036, up 45% from 2025 or about 3.4% a year. My bear case is 84,800 and my bull case 118,000. The Association’s 2025 Fuel Report has a reference scenario of about 150,000 tonnes a year by 2040.
| Region | 2025 (Association) | 2036 base (mine) | What drives the base case |
|---|---|---|---|
| China | 13,872 | 30,000 | About 40 reactors under construction, new approvals |
| United States | 19,011 | 20,000 | Uprates and life extensions, little new construction |
| India | 1,884 | 4,300 | Units under construction plus a long pipeline |
| Japan | 2,135 | 4,500 | Continued restarts |
| Russia | 6,251 | 9,500 | Domestic and export builds |
| All regions | 68,920 | 99,900 | 3.4% a year |
Tonnes of uranium per year. The 2036 figures are my estimates, not the Association’s.
The most important US data point is the stockpile. Energy Information Administration data show 118 million pounds held by reactor owners at the end of 2025 and 170 million pounds held by all US commercial holders, the most since 2003. Utilities report 186 million pounds of requirements for 2025 to 2035 not yet covered by contracts. That is the contracting opportunity, and the inventory lets utilities wait.
Part 4: Where Uranium Supply Comes From
World mine production was 60,213 tonnes in 2024, about 90% of requirements, up from 74% in 2020. Inventories, recycled material and enrichment tails supplied the rest. Kazakhstan produced 23,270 tonnes, Canada 14,309, Namibia 7,333 and Australia 4,598.
Kazakhstan sets the pace. Kazatomprom reported on November 3, 2026 that nine-month production was 20,508 tonnes on a 100% basis, up 10%, and reiterated full-year guidance of 27,500 to 29,000. Its own sales fell 7% over the nine months while group production rose, so restraint cuts both ways. Kazakhstan has also raised its mineral extraction tax from 9% to 12.4%, acid supply remains a cost risk, and a new law sets licence-transfer thresholds.
Canada is the highest-grade supplier and has been unreliable this year. A flood and bridge collapse in May 2026 disrupted McArthur River and Key Lake. Cameco produced 3.5 million pounds in the third quarter against 4.4 million a year earlier, narrowed 2026 production guidance to 19.5 to 20.5 million pounds from 19.5 to 21.5, and raised unit-cost guidance to CAD 67 to 70 per pound from CAD 63 to 67.50.
Smaller suppliers are growing but not decisively. The United States produced about 2.1 million pounds in 2025, the most since 2017. Namibia’s Langer Heinrich produced 4.82 million pounds in the year to June 2026 at a cost of USD 43.30 per pound. Denison’s Phoenix targets first production in the second half of 2028, while NexGen’s Rook I is economic at low prices but has a funding gap and mostly uncontracted reserves. Secondary supply adds about 13 million pounds a year in my model, falling to about 10 million by 2036.
Part 5: Is There Really a Uranium Shortage?
Not yet, and not in my base case before the mid-2030s. That is narrower than “no shortage ever,” and the precision matters because the investment case often leans on the word.
I define a deficit as reactor demand minus mine and secondary supply. In my base case the cumulative deficit from 2026 to 2036 is 81 million pounds against cumulative demand of about 2,454 million, or 3.3%. That figure is my model’s output, not an observed shortage, and it is a small residual of two very large numbers, so modest changes in either move it a lot. The shape matters more than the sum. The annual gap is 2.2 million pounds in 2026 and 9.2 million in 2030, the market is roughly balanced around 2033 as Rook I, Phoenix and a Mongolian project arrive, and the gap reopens to 23.5 million in 2036 because I hold existing mines flat while demand grows. The bear case has a surplus every year, 157 million pounds in total. The bull case reaches a deficit of about 80 million a year by 2036.
The base case is fragile in both directions. The whole 81 million disappears if supply runs 3.6% higher every year or demand runs 3.3% lower. It grows if mines deplete or projects slip: a one-year project delay lifts it to 127 million pounds, a two-year delay to 173 million, 2% annual depletion to 177 million, a 10% cut in Kazakh output from 2028 to 150 million, and 2% depletion plus a one-year delay to 224 million. Using the Association’s own demand paths, the cumulative gap is 33 million pounds on the lower path, 525 million on the reference path and 1,142 million on the upper path, so whether there is a shortage depends mainly on whose demand forecast you accept. Commentary that cites annual deficits of 20 to 85 million pounds is above my base case in every year and is matched only by my bull case; I do not endorse the high end.
Two buffers explain the calm near term: US commercial inventories of about 170 million pounds, which let utilities wait (spot below the long-term price suggests many are), and price-driven supply from Kazakh output beyond guidance, restarts and deferred projects. Inventories and secondary supply are why a modelled deficit does not mean utilities face scarcity today.
Part 6: What Could Uranium Cost in 2030 and 2036?
I use three outcomes rather than one forecast, with probabilities of 30% for the bear case, 50% for the base case and 20% for the bull case. These weights are my judgment, and Part 11 shows what changing them does. Spot today is about USD 90 to 92 per pound depending on the measure: Cameco’s month-end figure was USD 91.55 on October 30, 2026, and the price implied by the Sprott trust’s holdings was USD 90.11 on November 13, 2026.
| Scenario | Probability | 2030 spot (nominal) | 2036 spot (nominal) | 2036 spot in 2026 dollars |
|---|---|---|---|---|
| Bear | 30% | USD 60 | USD 62 | USD 48 |
| Base | 50% | USD 104 | USD 122 | USD 95 |
| Bull | 20% | USD 160 | USD 180 | USD 141 |
| Probability-weighted | 100% | USD 102 | USD 116 | USD 90 |
These are my scenarios, not market forecasts.
The weighted deck rises about 2.5% a year in nominal terms and is roughly flat after inflation. I am not forecasting a boom; I am forecasting that uranium roughly keeps its value, with a wide range around that.
Why not higher? Because the cheap tonnes are profitable far below today’s price. Denison’s Phoenix has a reported cash operating cost of USD 6.28 per pound and an all-in capital cost near CAD 700 million. NexGen’s 2021 feasibility study for Rook I showed operating costs of CAD 7.58 per pound, although its 2024 update raised life-of-mine cash operating costs to about CAD 13.86 and pre-production capital to about CAD 2.2 billion from CAD 1.3 billion. What holds projects back is financing, contracting, permitting and time. Deep Yellow has deferred a final investment decision on Tumas in Namibia with the long-term price in the mid-USD 90s. The company points to market conditions, and I read that as a sign that price alone is not enough without long-term contracts. A deficit does not automatically produce a USD 170 price.
Part 7: Ways to Invest
| Route | What you own | Main risk |
|---|---|---|
| Physical trust | Stored uranium, through units | No redemption, so the discount can persist |
| Producer | A mining and marketing company | Operating problems, contract ceilings, sovereign risk |
| Developer | A project not yet in production | Dilution, delays, a binary outcome |
| Royalty company | A claim on other people’s output | Valuation, leverage, a changed business |
| Fuel-cycle company | Enrichment, conversion or fabrication | Not a uranium price instrument |
| ETF | A basket of the above | The same names, plus fees |
No uranium investment route is a direct claim on nuclear growth. Each is a claim on a business or an asset with its own economics, and those often diverge from the theme.
Part 8: The Companies Worth Understanding
For each company I estimate a fair value per share, a probability-weighted ten-year annualized return from the dated price (the “PW return”), a return-test price at which that return equals 9.5%, and a Buy price that also demands a margin of safety (the lower of the return-test price and 85% of fair value). The return bands describe the modelled return only and are not recommendations: Well above hurdle at 11.5% or more, Above hurdle at 9.5% to 11.5%, Near hurdle at 8% to 9.5%, Short of hurdle at 6% to 8%, Well short of hurdle at 3% to 6%, Far short of hurdle at 0% to 3%, and Negative return below zero. A formal Buy is a separate test, set out under the Part 11 table.
Cameco (TSX: CCO, NYSE: CCJ)
The market is paying for Westinghouse more than for uranium. Cameco’s third-quarter report on October 30, 2026 was soft: revenue of CAD 619 million, an adjusted net loss of CAD 2 million and adjusted EBITDA of CAD 267 million against CAD 310 million a year earlier. Production fell to 3.5 million pounds from 4.4 million, and the cost of produced uranium rose to CAD 57.40 per pound from CAD 36.36. Cameco’s 49% share of Westinghouse earned CAD 143 million of adjusted EBITDA in the quarter, but CAD 427 million for nine months against CAD 569 million a year earlier.
I value Cameco in parts: uranium and fuel services, the Westinghouse stake, a possible government interest and a Canada Revenue Agency dispute over 2007 to 2013 (CAD 3.3 billion of adjustments remain). At USD 70.73 on November 13, 2026 my fair value is USD 40.64 and the PW return is 2.2%, far short of the hurdle. That was 0.7% at the USD 82.41 I used before the results, so the fall has helped, but only partway.
The decisive question is what Westinghouse is worth. Cameco announced a confidential draft registration statement for a Westinghouse IPO on July 30, 2026. Bloomberg News reported on September 18, 2026, citing unnamed people, that the company was said to be targeting a valuation above USD 50 billion. Neither Westinghouse nor Cameco has confirmed that, and I found no public filing as of November 15, 2026. On my assumptions, Cameco’s price equals my fair value if Westinghouse is worth about USD 50 billion (USD 47 billion to USD 53 billion, depending on whether the government interest vests). My own bottom-up estimate is about USD 19 billion (USD 9 billion bear, USD 39 billion bull). The stock therefore discounts something close to the reported target; if an IPO prices there, Cameco is roughly fairly valued, not cheap. Westinghouse’s net debt is undisclosed; I assume USD 3 billion, and each USD 3 billion either way moves my fair value by about USD 1.40 a share.
Kazatomprom (London: KAP, depositary receipts)
Kazatomprom is the closest to the hurdle among the companies that fall short of it. Its November 3, 2026 report reiterated 2026 production guidance of 27,500 to 29,000 tonnes and showed a realized price of USD 79.45 per pound on its own sales, up from USD 64.51 in the first half. At USD 60.30 on November 13, 2026 (the London close, quoted in US dollars), my PW return is 8.9%, fair value is USD 54.23, the return-test price is about USD 57 and the Buy price about USD 46. The bear case loses 3.9% a year, the shallowest of the producers I valued.
I am cautious for three reasons. The state of Kazakhstan owns about 75% indirectly, and tax, licence and marketing decisions can change. My model assumes realized prices keep moving toward the long-term price; the third quarter supports that, but no 2027 guidance was published. And dividend withholding for depositary-receipt holders is unverified: Kazakhstan’s general rate is 15%, with an exemption for securities listed on a Kazakh exchange, and I could not confirm how it applies to London receipts. At 0%, 5% or 15%, my PW return would be 8.9%, 8.6% or 8.0%. The return-test price is about 5% below the current price, so a modest fall would carry Kazatomprom over the bare 9.5% test, though not the Buy test, which needs a price near USD 46.
Boss Energy (ASX: BOE)
Boss’s September quarter, released on October 28, 2026, was mixed. Drummed production was 359,000 pounds, sales were 301,000 pounds at AUD 117 and the all-in sustaining cost was AUD 79. Fiscal 2027 guidance was held at 1.25 to 1.30 million pounds, with all-in sustaining costs of AUD 83 to 92. The company said record flow offset an expected decline in solution grade in maturing wellfields and expects costs to rise as it falls. The Alta Mesa plant produced about 10,000 pounds on a 100% basis, and plant production has stopped pending new wellfields because of state permitting delays. Cash was AUD 39.2 million, with 1.641 million pounds in inventory.
The shares closed at AUD 1.425 on November 13, 2026, after a volatile fortnight that included AUD 1.65 on November 11. My PW return is 7.7% at AUD 1.425, 8.2% at AUD 1.355 and 8.8% at AUD 1.29. Fair value is about AUD 1.24, the return-test price about AUD 1.21 and the Buy price about AUD 1.06, so the price would have to fall about 15% to reach the return-test price. Confidence is low and the bear case loses 7.0% a year.
Ur-Energy (NYSE American: URG, TSX: URE)
Ur-Energy is a small US producer incorporated in Canada. Its November 9, 2026 results showed 190,000 pounds sold at USD 64.95, a product loss of USD 0.97 million and drummed production down 25.8% from the second quarter. The release’s table classes only 10,000 pounds of the quarter’s sales as produced and 180,000 as non-produced. It deferred 600,000 pounds of deliveries to 2027 and 2029 and guided to about 700,000 pounds for 2026 at about USD 67. The shares closed at USD 0.61 on November 13, 2026, down from USD 0.81 on the day of the results (quotes are to the cent). At that price my PW return is 10.7%, above the hurdle, because the return-test price of about USD 0.68 is already above the market. That is a pass on the bare test and nothing more. The Buy price is about USD 0.54, 12% lower, the bear case is a total loss and confidence is low to medium, so it is not a formal Buy. In my model the deferrals alone do not hurt value, since the pounds are still sold later, but they push cash needs forward.
NexGen and Denison
NexGen (NXE) closed at USD 7.76 on November 13, 2026. My PW return is 5.1% and fair value USD 5.05. Rook I is a world-class deposit. My replication of its 2021 feasibility study came within 1.3% of the published net present value, and my valuation uses the higher capital cost of the 2024 update. On November 9, 2026 NexGen reported CAD 821.6 million of cash and short-term investments and said it still needs additional financing to finish construction. My valuation uses the earlier balance sheet; updating it would lower fair value by about 2%. Reserves are mostly uncontracted. Denison (DNN) closed at USD 2.19 on November 6, 2026. Phoenix was about 24% complete on September 30, with CAD 161.1 million spent so far this year, a 2026 forecast of CAD 278.6 million and first production targeted for the second half of 2028. My PW return is 4.2% and fair value USD 1.20. Both are well short of the hurdle.
Uranium Energy, Centrus and Uranium Royalty
Uranium Energy (UEC) closed at USD 6.66 on November 13, 2026 and produced 229,294 pounds in its 2026 fiscal year. My PW return is 2.0%, with a 17% a year loss in the bear case. Centrus (LEU) is a fuel-cycle company, not a uranium producer. It closed at USD 104.60 on November 6, 2026. Its third-quarter profit of USD 31.6 million included a USD 46.4 million non-cash warrant gain, and about USD 5.4 billion of its USD 7.0 billion backlog is contingent. The Department of Energy does not intend to exercise further high-assay fuel options (about USD 0.8 billion of backlog), and the 2027 budget leaves out cascade funding. My PW return is 5.0%, well short of the hurdle. The valuation uses the balance sheet from before the third quarter, because I could not confirm how much of the higher cash is customer prepayments.
Uranium Royalty, now a US corporation on Nasdaq as UROY, is mostly a trona royalty business after buying about 92% of the Sweetwater royalty entities in July 2026. It sold all 593,255 pounds of its physical uranium after year-end at an average of USD 85.91. At USD 2.78 on November 13, 2026 my PW return is 0.6%, the weakest I valued. I did not rate Paladin, whose last verified price was from October 16, or Energy Fuels, whose value depends on rare earths I did not value.
Part 9: Physical Uranium and the Sprott Discount
The Sprott Physical Uranium Trust (TSX: U.UN and U.U) is the cleanest way to own uranium itself. On November 13, 2026 it held 81,697,348 pounds, with a net asset value of USD 21.62 per unit (CAD 30.85) against a market price of USD 18.34 (CAD 26.20), a discount of 15.18%. The uranium was worth USD 7.36 billion, or USD 90.11 per pound. The annual fee is 0.71%.
A 15% discount looks like a bargain, and in one sense it is: if it closed tomorrow, units would rise about 18%. But that is a trade, not a ten-year return, and the structure makes it hard to close. Units cannot be redeemed, the prospectus has no buyback program, and the trust may not issue units below net asset value, so it cannot use the discount to raise money and buy uranium.
For a ten-year holder, the question is what uranium does. On my scenarios the probability-weighted return is about 2.0% a year at USD 18.34, far short of the hurdle. To reach 9.5%, spot would have to be about USD 205 in 2036 if the discount fully closes, USD 241 if it stays and USD 256 if it widens to 20%. My bull case is USD 180. At a flat inflation-adjusted USD 100 for the decade, units return 4.5% a year if the discount closes and 2.8% if it does not. SPUT is a tactical way to own uranium without company risk, best when the discount is wide and spot is stable. It does not clear a 9.5% hurdle at today’s prices.
I did not rate Yellow Cake plc. The last price I could verify was from August 13, 2026, too old to rely on.
Part 10: Uranium ETFs
ETFs offer a basket instead of company picking, but the baskets overlap heavily with the companies above.
| Fund | Structure and listing | Fee | Largest exposures (dated snapshots) | Share I could not model |
|---|---|---|---|---|
| URA (Global X) | US-domiciled, NYSE Arca | 0.69% | Cameco 23.2%, Sprott trust 7.7%, NexGen 6.1%, Kazatomprom 5.0% (October 16, 2026) | About 44% |
| URNM (Sprott) | US-domiciled, NYSE Arca | 0.75% | Cameco about 21%, physical trusts about 22%, CGN Mining 5.1% (October 23 to 28, 2026) | About 24% |
| URNJ (Sprott juniors) | US-domiciled, NYSE Arca | 0.80% | Deep Yellow 5.4%, Bannerman 4.8%, IsoEnergy 4.8% (September 3, 2026) | About 45% |
| HURA (Global X Canada) | Canadian ETF, TSX | 0.75% (0.69% proposed) | Not re-checked for this article | Not modelled |
I do not publish a ten-year return forecast for any ETF. Between about a quarter and 45% of each fund’s weight sits in companies I did not model, such as small-reactor developers, Japanese industrials and Chinese miners. From structure alone: the funds are cap-weighted, so their biggest positions are the biggest and most expensive companies, mostly ones I rate well short of the hurdle, and a fund that owns a company at a price I would not pay will not beat that company’s return. Physical trusts are 8% of URA and about 22% of URNM, so each fund is partly an investment in discounted trusts.
The Canadian fund has a pending event. On October 21, 2026, Global X Canada announced a special meeting of HURA unitholders for November 26, 2026 at 10:00 a.m. in Toronto, with an adjourned meeting on November 27 if there is no quorum. Unitholders will vote on replacing the Solactive Global Uranium Pure-Play Index with the Solactive Global Uranium & Nuclear Components Index, which caps physical uranium trusts at 10%. If approved, the management fee falls from 0.75% to 0.69%, and the ticker changes from HURA to URA. Approval needs a majority of votes cast, and the manager may delay, modify or end the change even after a yes vote. The circular says the manager expects the fund to realize capital gains in carrying it out, with no special distribution expected. I have not assumed the vote passes. If you hold HURA, read the circular, because the fund you hold on November 27 may not be the one you bought.
Part 11: What the Valuation Models Actually Say
For each company I forecast cash flows under each price scenario, add what sits outside them (net cash, inventory, the Westinghouse stake), allow for dilution, and compute the ten-year return from today’s price to the end of 2036. Surplus cash is assumed paid out and reinvested at 9.5%, a convention I test at the end of this part. The headline is the return on probability-weighted terminal wealth. The mean of the three scenario returns is lower for companies with wide outcomes (Ur-Energy’s bear case is a total loss), so I use it only as a warning light.
| Security | Reference price and date | PW return | Return band (not a recommendation) | Return-test price | Fall needed | Buy price | Bear case | Formal Buy now? |
|---|---|---|---|---|---|---|---|---|
| Ur-Energy | USD 0.61, November 13 | 10.7% | Above hurdle | USD 0.67 to 0.69 | none (price is 10% below) | USD 0.53 to 0.55 | total loss | No |
| Kazatomprom | USD 60.30, November 13 | 8.9% | Near hurdle | USD 56 to 58 | 5% | USD 45 to 47 | loss of 3.9% | No |
| Boss Energy | AUD 1.425, November 13 | 7.7% | Short of hurdle | AUD 1.20 to 1.22 | 15% | AUD 1.05 to 1.10 | loss of 7.0% | No |
| NexGen | USD 7.76, November 13 | 5.1% | Well short of hurdle | USD 5.10 to 5.20 | 34% | USD 4.25 to 4.35 | loss of 6.7% | No |
| Centrus | USD 104.60, November 6 | 5.0% | Well short of hurdle | USD 68 to 70 | 34% | USD 55 to 56 | loss of 1.1% | No |
| Denison | USD 2.19, November 6 | 4.2% | Well short of hurdle | USD 1.30 to 1.35 | 39% | USD 1.00 to 1.05 | loss of 8.2% | No |
| Cameco | USD 70.73, November 13 | 2.2% | Far short of hurdle | USD 35 to 36 | 50% | USD 34 to 35 | loss of 8.4% | No |
| Sprott trust (SPUT) | USD 18.34, November 13 | 2.0% | Far short of hurdle | about USD 9 | 51% | about USD 9 | not applicable | No |
| Uranium Energy | USD 6.66, November 13 | 2.0% | Far short of hurdle | USD 3.20 to 3.30 | 51% | USD 3.20 to 3.30 | loss of 17.1% | No |
| Uranium Royalty | USD 2.78, November 13 | 0.6% | Far short of hurdle | USD 1.15 to 1.25 | 57% | USD 0.70 to 0.75 | loss of 7.2% | No |
The return-test price is where the PW return equals 9.5%. Every price is a closing price on the date shown: Centrus and Denison are from November 6 and the rest from November 13. A formal Buy needs a price at or below the Buy price, a bear case no worse than a 5% annual loss and confidence above low-to-medium. Only Kazatomprom meets the last two, so it alone could earn a formal Buy, at its Buy price. Ur-Energy clears the bare return test but meets none of the three. No security qualifies today.
Three observations stand out. First, the ranking is robust. Kazatomprom, Boss and Ur-Energy are nearest the hurdle in nearly every version of the model, including the cash tests below, and Cameco, Uranium Energy and the Sprott trust are furthest. Second, price matters a great deal. A 20% lower price lifts Kazatomprom’s PW return to 11.3%, Boss’s to 10.1%, Ur-Energy’s to 13.2% and Cameco’s to 4.5%. Third, the answer depends on how bullish you are.
| Scenario weights (bear, base, bull) | Kazatomprom | Boss | Ur-Energy | Centrus | NexGen | Denison | Cameco | Uranium Energy |
|---|---|---|---|---|---|---|---|---|
| 30%, 50%, 20% (mine) | 8.9% | 7.7% | 10.7% | 5.0% | 5.1% | 4.2% | 2.2% | 2.0% |
| 25%, 50%, 25% | 9.7% | 8.9% | 12.2% | 5.7% | 5.9% | 5.0% | 3.1% | 3.2% |
| 20%, 50%, 30% | 10.6% | 10.1% | 13.6% | 6.3% | 6.7% | 5.9% | 3.8% | 4.3% |
| 10%, 40%, 50% | 12.7% | 13.3% | 17.3% | 8.3% | 8.8% | 8.1% | 6.0% | 7.5% |
Even if you put half your probability on the bull case, NexGen, Denison, Centrus, Cameco and Uranium Energy do not reach 9.5%. A reader who weights the bull case that heavily is making a bet on Kazatomprom, Boss and Ur-Energy, not on the sector’s best-known names. A different lens gives a similar ranking. The flat inflation-adjusted uranium price each stock needs to earn 9.5% is USD 94 for Kazatomprom, USD 96 for Ur-Energy, USD 113 for Boss, USD 122 to 124 for NexGen, Denison and Uranium Energy and USD 138 for Cameco, against about USD 90 today. Cameco needs the most because its contracts cap what it realizes: even at USD 160 spot, its own sensitivity table shows USD 104 in 2030.
The Cameco result deserves one more number. At today’s price, Cameco earns 9.5% only if Westinghouse is worth about USD 52 billion today and compounds at 9.5% a year until exit. If you assume it compounds at 12%, the required value today is about USD 41 billion, and at 6% it is about USD 71 billion. Whatever the assumption, that is the bet you are making.
For Kazatomprom, an extra 10% or 20% haircut to terminal value for sovereign risk would lower the return to 7.7% or 6.5%. For every company the weakest assumption is how far realized prices follow spot, which is why I show ranges.
One convention deserves its own test. My model pays out surplus cash and reinvests it at 9.5%, my hurdle. That convention is internally consistent with my return-test calculation, but it assumes distributed cash can earn my 9.5% hurdle. A required return is not a guaranteed reinvestment return, and for mines that run down it matters: for Kazatomprom, Boss, Ur-Energy and Denison, roughly half to nine-tenths of terminal wealth comes from distributions and what they earn. At a 6% reinvestment rate the PW returns fall to 8.0% for Kazatomprom, 6.1% for Boss, 9.7% for Ur-Energy, 3.0% for Denison and 1.8% for Cameco. At 4% they are 7.5%, 5.3%, 9.1%, 2.4% and 1.6%. Kazatomprom stays below 9.5% in every case, Ur-Energy only just misses at 4%, and no other security gets close. Kazatomprom’s return-test price falls from about USD 57 to about USD 53 at 6%, while Buy prices move much less because the margin-of-safety limit often binds. I keep 9.5% as the headline because the return-test and Buy prices are built on it, and the conclusion that nothing passes the full Buy test survives the generous end of the range. Read the headline returns as the top of a range, not a promise.
Part 12: What Could Go Wrong?
The most direct risk is price. In my bear case spot slides toward USD 60 by 2030 and stays there. At a flat inflation-adjusted USD 50, every uranium equity in my sensitivity tests loses money, and one loses everything. That is not exotic: spot was USD 47.68 at the end of 2022.
Second, good news may not reach shareholders. Producers are paid their contract book, and developers raise equity, so dilution can consume much of a higher price. Third, mines miss plans: a flood in May, a longer shutdown at Key Lake, an expected decline in solution grade at Honeymoon and deferred deliveries at Ur-Energy.
Fourth, supply is concentrated. About three-quarters of 2024 mine supply came from Kazakhstan, Canada and Namibia, and Kazakhstan raised its extraction tax and changed licence rules this year. Enrichment depends on Russia and on government programs, whose support is not uniform. Fifth is the Westinghouse bet: Cameco’s price now discounts a valuation near the reported IPO target, and no public filing exists.
Sixth is correlation. Owning several uranium stocks diversifies company risk, not uranium risk. After the end of October, Cameco fell 15% and the URA fund 11% while the uranium price slipped less than 2%. Seventh, discounts to net asset value can persist.
The most useful discipline is to separate being right about the theme from being right about the price you paid. You can be right about nuclear and lose money anyway if you paid for ten years of good news in advance.
What would make me more constructive? A priced Westinghouse IPO well above USD 50 billion with disclosed debt, utilities contracting well beyond replacement, a Sprott trust at a premium, or simply lower prices. Less so? Spot slipping further below the long-term price, a Kazakh output surge or a failed financing at Rook I.
Part 13: Canadian Investor Considerations
Volatile, mostly capital-gain investments suit a TFSA or RRSP, where growth is sheltered. Our TFSA guide and RRSP versus TFSA comparison cover the mechanics.
Domicile matters more than listing. Cameco, NexGen and Denison are Canadian corporations. Ur-Energy is a Canadian corporation, continued under the Canada Business Corporations Act, with its headquarters in Colorado. Uranium Energy and Centrus are US corporations, and Uranium Royalty became one when it redomiciled in July 2026, with a Nasdaq listing only. Kazatomprom’s depositary receipts are issued by a Kazakh company and trade in London. Our guides to TSX, NYSE and Nasdaq listings, global stock markets and OTC stocks and depositary receipts explain access.
The Canada Revenue Agency says Form T1135 is required when the total cost, not the market value, of specified foreign property exceeds CAD 100,000. Shares of non-resident corporations count, and so do shares of Canadian corporations if held outside Canada. Property held in an RRSP or TFSA is excluded, and simplified reporting applies below CAD 250,000 of total cost. So a Canadian holding Uranium Energy, Centrus, Uranium Royalty or the Kazatomprom receipts in a non-registered account, with total foreign property costing more than CAD 100,000, files the form.
US dividend withholding depends on the account. The treaty rate is 15%, and the treaty exempts retirement arrangements, an exemption brokers apply to RRSPs when US shares are held directly. The CRA warns that foreign dividends paid to a TFSA may be subject to withholding, which generally cannot be recovered, while in a non-registered account the tax can usually be claimed as a foreign tax credit. A Canadian-listed fund can lose withholding tax inside the fund. Most of these companies pay no dividend and Cameco’s is small (CAD 0.24 a share a year), so this matters mainly for fund distributions.
US estate tax is separate. Shares of US-incorporated corporations are US property for estate-tax purposes, and the Internal Revenue Service requires a nonresident’s estate to file when US assets exceed USD 60,000. That is a filing threshold, not the amount exempt from tax. The Canada-US treaty can eliminate or reduce the tax, but relief must be claimed and depends on your worldwide estate. Advisers generally treat Canadian-listed funds as not US property even when they hold US stocks, which is commentary, not an IRS rule. A US-listed ETF is less clear, and US shares in an RRSP or TFSA are still US property. With large US holdings, speak to a cross-border professional.
Three narrower points. The Sprott trust’s annual information form says its units are a qualified investment for registered plans while they are listed on a designated stock exchange such as the TSX, provided the holder deals at arm’s length and has no significant interest; confirm with your broker. Ur-Energy’s 10-K says it may be a passive foreign investment company for US tax purposes, a warning and not a determination, and relevant only to US persons, for whom the IRS requires Form 8621. Kazakh withholding on depositary receipts is unverified. Capital gains are only partly taxable in Canada, and our capital gains guide has the current inclusion rate.
Currency adds noise. I assume the Canadian dollar is as likely to strengthen as weaken around 1.40, testing end rates of 1.25, 1.40 and 1.55, so expected Canadian-dollar and US-dollar returns are equal. The Sprott trust’s figures imply about 1.43 on November 13, 2026, so this simplifies.
Part 14: What I Would Actually Do
- Decide whether a uranium investment earns a place in your portfolio at all. At today’s dated prices nothing passes my full test, so the default is no allocation, with cash or short-term bonds as a legitimate home while you wait.
- Write the prices down. Set alerts at the Buy ranges in Part 11: Kazatomprom near USD 45 to 47, Boss near AUD 1.05 to 1.10, Ur-Energy near USD 0.53 to 0.55, Centrus near USD 55 to 56 and Cameco near USD 34 to 35, and re-run the numbers when a price arrives.
- If you want exposure now regardless, cap it at 1% of your portfolio and treat 2.5% as a ceiling, not a target. In my best-ranked mix (half Kazatomprom, 15% each Ur-Energy and Boss, 10% each NexGen and the Sprott trust), a 1% sleeve loses about 0.5% of the portfolio in the bear case, gains about 3.8% in the bull case, and its probability-weighted ten-year terminal value falls short of a 9.5% benchmark by about 0.3% of the total portfolio. That is a ten-year terminal-value shortfall expressed as a percentage of the total portfolio’s starting value, not an annual return shortfall. At 2.5% the figures are 1.2%, 9.5% and 0.7%, with the same meaning; at 10%, the bear case costs almost 5% of the portfolio.
- Start with Kazatomprom, the company closest to the hurdle and the one with the shallowest bear case among the three nearest the hurdle. Treat Boss and Ur-Energy as small, low-confidence satellites, and the Sprott trust as a tactical discount trade only.
- Scale in on price and events, not the calendar. Watch the HURA vote on November 26, 2026, a public Westinghouse registration statement and its pricing, Kazatomprom’s 2027 guidance and Cameco’s fourth-quarter results on February 12, 2027.
- Choose the account before the stock. Keep volatile, non-dividend names in a TFSA or RRSP where you can, and check Form T1135 and US estate-tax exposure before holding US-incorporated shares outside a registered plan.
- Review quarterly, update entry prices after each set of results and write down in advance what would make you sell: a sustained gap of spot below the long-term price, a Kazakh supply surge or a failed project financing.
Conclusion: A Good Industry Is Not Always a Good Investment
Nuclear power is growing, and uranium demand with it. My base case has demand up 45% by 2036. The supply side is more contested: my model shows a modest cumulative deficit, not a structural one, and it is a model output, not an observed shortage. It depends on project timing and on how much the old mines can still produce. Uranium probably roughly holds its value against inflation, with a wide range around it.
The step from there to a good investment is where the story weakens. Producers are paid their contract book, not spot. Developers dilute. The largest company is priced for a Westinghouse valuation not yet tested in a public market. Funds hold the same names and are partly invested in physical trusts. On the prices I could date, none of the securities I valued passes my full test, even on my most generous cash assumption. Ur-Energy clears the bare return test with a bear case that is a total loss, and Kazatomprom, the nearest of the rest, carries sovereign risk and an unverified withholding question.
That is a verdict about price, not against uranium investment as an idea. If the price comes to you, or the facts change, the answer can change quickly. Until then, the honest position for a Canadian investor with a 9.5% hurdle is to know the companies, write down the prices and wait.
Research Methodology and Cutoff
This article reflects research completed on November 15, 2026, with November 13, 2026 as the last trading day. It builds on an earlier full-length study from October 2026, updated for later events and rechecked against primary documents: company results and filings, tax-agency guidance, the Sprott trust’s annual information form, World Nuclear Association and Energy Information Administration data, Cameco’s price page and Global X Canada’s circular. I re-ran the valuation model on the updated prices and reconciled its output to the figures printed here.
Limits matter. Centrus and Denison prices are November 6 closes and every other price is a November 13 close, each dated where used. The Kazatomprom, NexGen, Uranium Energy, Ur-Energy and Uranium Royalty closes come from price aggregators, not the exchanges, and Ur-Energy’s is quoted to the cent. Headline returns assume surplus cash is reinvested at 9.5%, which Part 11 tests. I treated Bloomberg’s Westinghouse valuation report as unconfirmed because it relies on unnamed sources. Some US estate-tax and withholding points rest on adviser and broker commentary, not agency rules, and are flagged. The scenario weights, the return hurdle and the Westinghouse and Kazatomprom adjustments are my judgments.
Sources
| Source | Link |
|---|---|
| World Nuclear Association, reactors and uranium requirements | https://world-nuclear.org/information-library/facts-and-figures/world-nuclear-power-reactors-and-uranium-requireme |
| U.S. Energy Information Administration, Uranium Marketing Annual Report | https://www.eia.gov/uranium/marketing/ |
| Cameco, uranium price page | https://www.cameco.com/invest/markets/uranium-price |
| Cameco, third-quarter 2026 results | https://www.cameco.com/media/news/cameco-reports-2026-third-quarter-results |
| Kazatomprom, 3Q26 operations and trading update | https://www.londonstockexchange.com/en/news/article/KAP/kazatomprom-3q26-operations-and-trading-update/58e3-a1-a0-b9-3f-41118e |
| Sprott Physical Uranium Trust | https://sprott.com/investment-strategies/physical-commodity-funds/uranium/ |
| Ur-Energy, third-quarter 2026 results | https://www.sec.gov/Archives/edgar/data/0001375205/000110465926124599/urg-ex99d1.htm |
| Ur-Energy, 2025 annual report on Form 10-K | https://www.sec.gov/Archives/edgar/data/1375205/000110465926025923/urg-20251231x10k.htm |
| Boss Energy, September 2026 quarterly presentation | https://www.bossenergy.com/images/media/BOE_September_2026_Quarterly_Results_Presentation.pdf |
| Centrus Energy, third-quarter 2026 results | https://www.sec.gov/Archives/edgar/data/0001065059/000162828026070229/ex991-10q2026_11x04.htm |
| Denison Mines, Q3 2026 results (Form 6-K exhibit) | https://www.sec.gov/Archives/edgar/data/0001063259/000106325926000032/dmc-pr20261104q3pressrele.htm |
| Uranium Royalty Corp, annual report filing | https://www.sec.gov/Archives/edgar/data/2143673/000119312526321135/0001193125-26-321135-index.htm |
| Global X Canada, management information circular | https://www.globalx.ca/wp-content/uploads/2026/10/GX_Management-Information-Circular-Oct-2026.pdf |
| Bloomberg Government, Westinghouse IPO report (paywalled) | https://news.bgov.com/capital-markets/westinghouse-said-to-target-over-50-billion-valuation-in-us-ipo |
| Canada Revenue Agency, questions and answers about Form T1135 | https://www.canada.ca/en/revenue-agency/services/tax/international-non-residents/information-been-moved/foreign-reporting/questions-answers-about-form-t1135.html |
| Internal Revenue Service, about Form 8621 | https://www.irs.gov/forms-pubs/about-form-8621 |
| Internal Revenue Service, Form 706-NA instructions | https://www.irs.gov/instructions/i706na |
| Sprott Physical Uranium Trust, annual information form | https://sprott.com/media/z0zjfn40/sput-annual-information-form.pdf |
