Uranium: The Quiet Commodity That Became Strategic Infrastructure

For most of the last decade uranium was an afterthought. Cheap natural gas and falling solar costs made nuclear look expensive and politically awkward. Mines closed. Exploration budgets dried up. Producers who survived did so by cutting output and waiting.

That period is over. Uranium now sits at the intersection of three forces that show no sign of reversing: electricity demand from artificial intelligence, national energy security policy, and a supply base that cannot respond quickly even when prices rise.

This is our view on where the market stands and what investors should understand before taking a position.

Where prices are today

The spot price crossed 100 US dollars per pound in January 2026 for the first time in two years, rising roughly 25 percent in a single month. It then cooled. Through the second quarter the price traded in a narrow band between 84 and 87 dollars, and Cameco reported an end of July spot price of 86.36 dollars.

For a market that had just broken a two year high, that consolidation disappointed a lot of holders. Uranium equities weakened alongside it.

The more informative number sits elsewhere. Long term contract prices, negotiated directly between utilities and producers rather than traded on the spot market, reached around 90 dollars per pound by the end of the first quarter. That is the highest level since 2008, and it rose steadily through a year in which the spot price went essentially nowhere.

Enrichment prices tell a similar story. The cost of enriched uranium reached 190 dollars per separative work unit, compared with 56 dollars three years earlier.

When spot prices stall but contract and enrichment prices climb, it usually means the same thing. Traders are cautious. The people who actually need the material are not.

The demand picture

Global installed nuclear capacity stood at 398 gigawatts as of last June. The World Nuclear Association's central scenario has that nearly doubling to 746 gigawatts by 2040. Their optimistic case reaches 966 gigawatts. Even the slow case still gets to 552.

Reactor fuel consumption follows. Roughly 68,900 tonnes of uranium in 2025, rising to more than 150,000 tonnes by 2040 in the reference scenario.

The AI angle is real but often misunderstood. Data centres do not consume uranium. What they consume is enormous quantities of continuous, reliable electricity, and they cannot tolerate the intermittency that comes with wind and solar alone. Nuclear is one of the few sources that delivers carbon free baseload power at the scale hyperscalers now require.

Amazon and Microsoft have both signed agreements to take power from nuclear plants. Meta issued a request for proposals covering 1.4 gigawatts of new nuclear generation in the United States. These are not pilot projects.

Worth noting: several analysts we follow argue the uranium case holds even if you strip data centre demand out entirely. The reactor build programme in China, India, France, Poland, Japan and South Korea would be enough on its own.

The supply problem

In 2024 global mine production covered about 90 percent of demand. The remaining 10 percent came from stockpiles, which are finite.

Bringing new uranium supply online is slow in a way that is difficult to appreciate from outside the industry. Permitting, financing, construction and ramp up routinely take a decade. Mines that were mothballed during the lean years do not simply restart. Skilled labour left the sector. Equipment was sold or degraded. Regulatory approvals lapsed.

This is why price signals alone have not fixed the shortfall. Producers have repeatedly under delivered against their own guidance.

There is also a newer source of pressure. Physical uranium funds, most prominently the Sprott Physical Uranium Trust, buy material on the spot market and hold it. Unlike utilities, they are not price sensitive buyers with delivery schedules to meet. Every pound they accumulate is a pound removed from an already tight market.

One widely followed sector analyst has argued that supplying 250 to 300 million pounds annually within a decade would require sustained prices in the 125 to 150 dollar range. That is a long way above where the market trades today.

Policy is now a price driver

Uranium has been formally designated a critical mineral under the United States Section 232 framework, placing it in the same category as rare earths and lithium for national security purposes. The Department of Energy has committed 2.7 billion dollars over the next decade to expand domestic enrichment capacity.

The strategic logic is straightforward. Enrichment capacity has historically been concentrated in a small number of jurisdictions, including Russia. Western governments have decided that dependency is no longer acceptable.

For investors this matters because it changes the demand profile. Government stockpiling and strategic reserve building is price insensitive by definition.

How to get exposure

There is no single right answer here. Each route carries a different risk profile.

Physical uranium trusts give the cleanest exposure to the commodity itself. No mining risk, no operational risk, no management execution risk. What you get is a claim on stored material. The trade off is that there is no yield and no upside beyond the price of the metal.

Diversified uranium ETFs such as Global X Uranium hold a mix of miners and nuclear component manufacturers, plus some direct physical exposure. Around 64 percent of assets sit in the top ten holdings, with Cameco alone at roughly 23 percent. Nearly 77 percent is invested outside the United States. This is a concentrated fund despite the diversified label.

Pure play miner ETFs such as Sprott Uranium Miners track companies that commit at least half their assets to uranium mining. The link to mining economics is tighter than in broader funds. That cuts both ways. A rising spot price does not repair a flooded mine or a cost overrun at a development project.

Junior miner ETFs target small and micro cap explorers and developers. Drilling results, permitting decisions and takeover interest can move these names violently in either direction. Suitable only for investors who understand they may lose the position entirely.

Individual producers such as Cameco and Kazatomprom offer direct operational exposure with the concentration risk that implies.

What we would flag

The spot price is not the whole picture. If you are trading uranium on spot movements you will be whipsawed. The structural case is expressed in contract prices and reactor build schedules, both of which move on multi year timescales.

Equities and the commodity diverge. Uranium equities weakened through mid 2026 even as fundamentals strengthened. Mining companies carry costs, debt, jurisdictional risk and management risk that physical uranium does not.

Jurisdiction matters enormously. Kazakhstan supplies roughly 40 percent of world output. Concentration of that degree in a single country is a risk that no amount of demand growth eliminates.

This is a long duration position. The thesis rests on reactor construction schedules stretching to 2040. Investors who need liquidity or who cannot tolerate multi year drawdowns should look elsewhere.

Sentiment has run hot before. Uranium rallied hard in 2023 and 2024, then spent most of 2025 range bound between 63 and 83 dollars while holders lost patience. Nothing about the current setup prevents that happening again.

Our view

The uranium case is one of the more coherent structural theses available in commodities right now. Demand is contracted years forward and rising. Supply cannot respond on a comparable timeline. Government policy is actively reinforcing both sides.

What it is not is a fast trade. The gap between where the market prices uranium today and where analysts believe prices need to sit to incentivise adequate supply is substantial, but closing that gap is a process measured in years rather than quarters.

For investors building a commodity allocation with a genuine long horizon, uranium deserves consideration. For anyone looking for momentum, the last eighteen months should serve as a warning.