A price record is supposed to feel like a party. Uranium got one this autumn, and the guest list looks oddly thin. The long-term price, the one that reflects what utilities agree to pay for fuel years down the road, climbed past its 2007 peak of about $95 a pound. Cameco published a month-end long-term reading of $96.50 for September, the highest in its data going back to 1996. Meanwhile the shares of the very companies that dig the stuff out of the ground have been sliding. So what is going on?
A record that was not a rally
Start with the numbers, because they are stranger than the headlines suggest. Cameco’s September data put the long-term price at $96.50 per pound and the spot price at $89.63. A separate price indicator cited by Uranium Unleashed put the long-term figure at $97 and spot at roughly $90 as of October 9. Different indicators, slightly different readings, same story. Spot sits well above where it spent the spring and summer, when the same source says it traded in a range of roughly $86 to $87, but it is still below its January peak of $100.25.
Now look at the equities. Cameco closed at $85.18 on October 2, according to a Yahoo Finance piece by Rich Duprey. That is down 6.9% for 2026 and about $50 below its 52-week high of $135.24. Among the newer reactor names, Oklo has lost about half its value and NuScale Power is down more than 45%. The VanEck Uranium and Nuclear Energy ETF, ticker NLR, is off about 16.4%. A commodity at a multi-decade high and a sector in the red. That is not how the script usually goes.
Why a higher price is not the same as higher profits
The first piece of the puzzle is that a producer does not sell everything at today’s price. Cameco’s realized price last quarter was about $65 a pound, per reporting that draws on the company’s results, while spot was near $88. Why the gap? Much of what Cameco sells goes out under older contracts and market-related contracts that carry floors and ceilings. According to the Startup Fortune write-up, market-related pricing made up 69% of revenue last quarter, up from 51% a year earlier. Those structures protect a producer when prices fall. They also cap what it earns when prices climb.
The company’s own guidance tells the same tale. Per the Uranium Unleashed summary, Cameco expects to deliver 29 million to 32 million pounds this year at a realized price of $85 to $89. That is a very respectable number. It is also lower than the headline long-term price would imply for a business that has spent years waiting for this moment. Quarterly revenue, the Startup Fortune piece reports, missed analyst consensus by more than 25%. Investors had priced in a faster handoff from record prices to record earnings, and the handoff is slower than they hoped.
There is a second, less comfortable issue. A price only means something if buyers transact at it. Term contracting volumes fell about 15% from a year earlier, to just over 38 million pounds through the end of August, based on figures in the same reporting. Duprey’s article makes the point bluntly: a record long-term price means little until utilities actually sign. TD Cowen, quoted in the Startup Fortune piece, describes utilities showing sticker shock on pricing and holding back from locking in long-term deals. When buyers stall, the quoted price can be a number on a screen more than a reflection of heavy trading.
The AI power story has a timing problem
For the last two years the bullish case for nuclear has leaned on data centers. Technology companies need enormous amounts of reliable power, and reactors deliver it around the clock. That part of the story has not gone away. Reporting this month describes a 20-year power purchase agreement between Google and Constellation Energy covering about 890 megawatts of uprated capacity across 11 Constellation units, and Constellation announced a 20-year deal with Amazon on October 1 for 690 megawatts at Calvert Cliffs in Maryland, as summarized by Uranium Unleashed. Those are real contracts with real money behind them.
But look at the timing. TD Cowen’s view, again as quoted by Startup Fortune, is that hyperscalers need megawatts in 2027 rather than gigawatts in 2037. That distinction hurts the small modular reactor names hardest, which helps explain why Oklo and NuScale have fallen so far. Uprating an existing plant or extending its life can add power in a few years. A new reactor design that has not yet been built is a longer wait. The market, it seems, is willing to pay for the power that shows up soon and discount the power that shows up later.
Policy is pushing in the other direction, slowly. On September 30 the United States and South Korea announced a non-binding framework that anticipates about $120 billion of Korean investment in deploying Westinghouse reactor technology in the United States, covering up to six AP1000 units and up to two APR1400 units. Cameco holds a 49% interest in Westinghouse and acknowledged the framework. Note the words non-binding and anticipates. Frameworks are not steel in the ground.
Supply is the quiet half of the equation
If demand has a timing problem, supply has a reliability problem. Kazatomprom, the world’s largest producer, cut its 2026 production target by 10%, citing sulfuric acid shortages and equipment delays, according to the Startup Fortune article. Cameco has had its own troubles, including slower-than-expected ground freezing at McArthur River and an extended mill shutdown at Key Lake in the third quarter. Supply that keeps missing its targets is a big part of why long-term prices keep rising even when buyers drag their feet.
There is also a geopolitical thread worth watching. On October 7, Kazatomprom shareholders voted on a supply contract with Uranium One Group, a Russian entity, for delivery to the Siberian Chemical Plant. According to Uranium Unleashed, the resolution failed to win a majority, with 88.34% of voting shares simply withholding their votes. At the same meeting, shareholders approved, with 88.63% in favor, a spot-term contract with a subsidiary of China’s State Power Investment Corporation. The same source notes that the sovereign wealth fund Samruk-Kazyna holds a majority of voting shares and suggests the outcome was settled before the ballots were counted. That is the author’s reading, not an established fact, but the direction of travel is clear enough. Kazakh uranium is being steered, whether by design or by default, away from one customer and toward another.
What a clear-eyed investor takes from this
Three lessons are worth keeping, and none of them is a trade.
First, the commodity and the company are two different bets. Owning exposure to the price of uranium is not the same as owning a producer whose contracts, costs, and operating record determine how much of that price it keeps. Cameco’s realized price sitting $20 or more below spot is the clearest illustration you will find.
Second, watch contracting, not just quotes. A long-term price can drift higher on thin volume. The more useful signal is whether utilities start signing multi-year deals in size. If they do, the record price starts to look earned. If they do not, it may prove to be a high-water mark on a quiet market. Cameco is due to report third-quarter results on October 30, and Uranium Unleashed notes that a Kazatomprom update is expected in late October or November. Those are the next checkpoints, and they will say more than any single day of trading.
Third, be honest about time horizons. Analysts remain constructive in the aggregate. Stifel, for example, raised its long-term price forecast to $120 a pound from $105, according to Startup Fortune, and the same article cites an average analyst price target for Cameco of $126.75. Forecasts are opinions, though, not outcomes, and the same analysts were presumably less cautious when the shares were near $135. A stock can be right about the long run and still punish you in the short run, which is exactly what this year has looked like.
The uranium story is not broken. A shortage of mined supply, a wave of new demand from power-hungry technology companies, and government support for nuclear build-out are all still on the table. But a record price and a record stock are two separate things, and 2026 has been a lesson in how long the gap between them can last. The question for the months ahead is not whether uranium is expensive. It is whether the people who need it are finally ready to commit.
This article is written for educational and informational purposes only and does not constitute financial or legal advice. The views and analytical frameworks presented draw on publicly available information and reported commentary from industry participants. Readers are encouraged to consult primary sources and form their own informed views on these complex topics.
