Uranium at 18-Year Highs: Real Nuclear Renaissance or a Hyperscaler-Fueled Bubble?

The number most people quote for uranium is the wrong one. The spot price — roughly $90 a pound this month, drifting back near a six-month high — is the figure that makes headlines. But utilities don't buy most of their reactor fuel on the spot market. They sign long-term contracts years ahead for the same reason they build reactors in the first place: certainty. And that number, the long-term contract price, hit $97 a pound on June 30 — an 18-year high that now sits above spot.

That gap is the overlooked signal. Earlier this year the pattern looked like textbook froth: spot spiked past $100 in late January, running well ahead of the slower-moving term market. Since then spot has cooled back to the low $90s while the term price kept climbing to $97. The surge was the speculative part; what remains is utilities paying a premium to lock in future supply. The market isn't chasing a trade — it's buying insurance against not having fuel in 2030.

The durable case is real, and it's written in contracts

The demand story behind all this is not manufactured. For 13 straight years, utilities contracted less uranium than their reactors consumed, running down coverage that now has to be rebuilt. The arithmetic is forcing them: U.S. utility coverage falls from roughly 98% of requirements in 2026 to 60% by 2030 and about 9% by 2033, and Europe's drops hard after 2030. Supply won't come easily. The world's largest producer, Kazatomprom, supplies about 39% of global output and is again struggling with sulfuric acid shortages; flooding cut Cameco's June output and it suspended mining at one of its flagship deposits. Analysts figure the price needed to coax enough new supply into the market is on the order of $125–150 a pound — well above where either market trades.

Then there is the hyperscaler chapter, which is what turned nuclear from an industrial story into an AI one. Microsoft committed roughly $16 billion to restart Three Mile Island; Google signed the first U.S. corporate SMR-fleet deal; Meta lined up around 7.7 gigawatts of new nuclear across partners. These are decade-out power contracts, not fuel purchases — but they underwrite the reactor buildout that needs the fuel. The term market at $97 is pricing exactly that.

Where the froth lives: the equities, not the fuel

From a valuation perspective, though, the balance sheet of common sense gets strained quickly. Cameco, the largest Western producer, trades at a market value around $42 billion. Measured against trailing results, the stock goes for roughly 90 times EBITDA and about 168 times trailing earnings, and even the forward multiple sits in the mid-70s — partly because a large share of its realized revenue still comes from older contracts signed at far lower prices. The market is paying now for a recovery that has not fully shown up in reported cash flow yet.

The more telling comparison is with the companies that have no earnings at all. Oklo, which has not delivered a reactor, carries a market cap of about $7.4 billion. NuScale goes for roughly $4.4 billion and something like 400 times sales. Set either against Centrus, an actual fuel company with real earnings, worth about $3.3 billion — the market values two pre-revenue reactor developers at more than a cash-generating fuel supplier. That is where froth is genuinely embedded: in optionality priced as if the buildout is certain, not in the term market that is doing the real backing.

The survival test, and the one number that decides it

For the cash-flow discipline, the key miner to check is Cameco — and it passes the survival question cleanly. Its debt-to-equity runs near 0.14x with an essentially clean balance sheet, and trailing operating cash flow of about $680 million covers its roughly $276 million of capital spending. Because much of its revenue is priced off long-term contracts rather than spot, a sharp uranium pullback would dent the model but not break it; miners selling into term contracts are insulated from the daily spot tape. The problem is the opposite one: survival isn't the scarce thing here, cheapness is. A clean balance sheet does not make 90 times trailing EBITDA a bargain.

So drop the binary. The demand is durable and the term curve — $97 sitting on top of $90 — is confirmation, not a lagging artifact. The speculation is concentrated in the equities that promise to capitalize on it, several priced with little margin of safety. That is why the clean falsifier runs through the commodity rather than the stocks: if the long-term price converges back down to spot — if utilities stop paying the premium and the $97 term drifts toward $90 and below — the structural story dies and leaves a thin, speculative spot market behind. Spot collapsing while term held would, by contrast, confirm the thesis, because the durability lives in the term market, which is exactly what the current gap says. Watch the term price. It is ahead of every headline.