The Future Uses of Uranium: AI Data Centres, SMRs & Space

Published: July 20, 2026

The Future Uses of Uranium

A single ceramic pellet of uranium fuel, about the size of a fingertip, holds as much energy as a tonne of coal. That fact is the whole reason uranium matters, and it is why a metal most people file under “bombs and power plants” is suddenly at the centre of the future economy. The reactors that run on those pellets supply about a tenth of the world’s electricity today, and after two decades of stagnation they are being built, restarted, and fought over again — because the data centres training artificial intelligence need enormous, around-the-clock power, and nuclear is one of the few sources that can deliver it.

Uranium’s story runs far past the power station. It drives every nuclear submarine on a core of weapons-grade fuel that lasts the life of the ship, it is heading back to the Moon inside compact fission reactors, and its dense, spent byproduct armours tanks and shields radiation. It is about to heat steel mills and make hydrogen. And uranium comes wrapped in an honest set of limits worth stating up front: it does not power the Mars rovers, it is not what runs a quantum computer, and nothing about a robot runs on it. This is what uranium actually does now, what it is about to do, and why a market this strategic keeps getting caught short of supply.

Why Uranium? The Most Concentrated Energy Source We Use

Uranium earns its place on one overwhelming property: energy density. The nucleus of a uranium-235 atom, split in a controlled chain reaction, releases millions of times more energy than the chemical bond that burns in coal or gas. That is not a marginal advantage over other fuels — it is a difference of scale so large that it changes what is physically possible, letting a single power station run a city for eighteen months on a few truckloads of fuel instead of a coal train arriving every day.

Because the advantage is so lopsided, every substitute is a downgrade by orders of magnitude. A source that matches uranium’s carbon-free, weather-independent, around-the-clock output at that density does not exist: fossil fuels emit carbon and burn by the trainload, while wind and solar are clean but intermittent and land-hungry. The question a grid planner faces is rarely “uranium or something denser” — there is nothing denser on offer — but “how much land, fuel, and emissions are we willing to trade to avoid it.”

These figures come from the Nuclear Energy Institute, and they explain why uranium is worth mining, enriching, and guarding despite the difficulty of doing all three. What follows is where that energy actually goes.

What Makes Uranium Fissile

Natural uranium is not ready to use straight from the ground. It is a mix of two isotopes, and only one of them — uranium-235 — splits easily enough to sustain a chain reaction. The catch is that uranium-235 makes up just 0.7% of natural uranium; the other 99.3% is uranium-238, which does not readily fission. To fuel an ordinary reactor, that 0.7% has to be concentrated, or “enriched,” up to between 3.5% and 5% uranium-235 — the single most important step in the whole fuel cycle, and the one where the world’s supply chain chokes.

Uranium-238 is not waste, though. It is “fertile”: bombarded with neutrons inside a reactor, it transmutes into plutonium-239, which is itself a fissile fuel. That quiet fact underpins two later chapters of uranium’s story — the fast reactors that could run on today’s leftovers, and the thorium reactors that still need a uranium spark to start. For now, the thing to hold onto is that a reactor lives on that thin 0.7% sliver, and everything strategic about uranium follows from the effort to concentrate it.

Uranium in Nuclear Power Reactors

Generating electricity is uranium’s overwhelming purpose, consuming the vast majority of the roughly 68,900 tonnes of uranium the world uses each year. Inside a reactor, fissioning uranium heats water into steam that spins a turbine — the same basic idea as a coal plant, but with a fuel millions of times denser and no carbon coming out of a stack. The World Nuclear Association counts about 440 operable power reactors supplying close to 9% of global electricity, with another 79 under construction, most of them in Asia.

The appetite of a single reactor is modest by fuel standards but demanding by supply standards. A typical 1,000-megawatt reactor needs on the order of 160 to 210 tonnes of natural uranium a year, which is milled, converted, enriched, and fabricated into fuel assemblies before loading. About a third of the core is swapped out for fresh fuel every 12 to 18 months, so a reactor is a steady, decades-long customer rather than a one-time buyer — which is exactly why utilities sign uranium contracts years in advance and why a supply squeeze takes so long to unwind.

How Much of the World’s Electricity Is Nuclear

Nuclear power provides roughly 9% of the world’s electricity and about a quarter of its low-carbon electricity, from those 440-odd reactors totalling around 400 gigawatts of capacity. In several countries the share is far higher — France draws about two-thirds of its power from uranium — and the fleet generated some 2,667 terawatt-hours in 2024. After years of flat or falling output in the West, the direction has reversed: reactors are being licensed for longer lives, shuttered plants are being switched back on, and new build is accelerating in China. The demand that turnaround creates for uranium is the thread running through the rest of this article.

Why Data Centers Are Reviving Nuclear Power

The most consequential new pull on uranium is not a reactor technology at all — it is artificial intelligence. Training and running large AI models takes enormous, uninterrupted electricity, and the data centres that do it are straining grids. The International Energy Agency projects that global data-centre electricity use will more than double, from about 415 terawatt-hours in 2024 to roughly 945 terawatt-hours by 2030 — an increase larger than the entire electricity consumption of many countries — and nuclear is one of the few sources that can supply clean power around the clock, which is what a data centre needs.

That logic has produced a wave of deals no one predicted five years ago. In September 2024 Microsoft agreed to buy the entire output of a restarted Three Mile Island reactor from Constellation, a plant shut in 2019 and now being brought back specifically to feed data centres, per Utility Dive. Amazon has backed the small-reactor developer X-energy and taken nuclear capacity from Talen Energy; Google has contracted for reactors from Kairos Power; Meta and Oracle have launched their own nuclear procurements. The buyers are technology companies, and the product they want is baseload uranium power — the same demand explored in depth in our look at why AI is reviving nuclear power.

Small Modular Reactors & the HALEU Question

Much of the new nuclear enthusiasm rides on small modular reactors, or SMRs — factory-built reactors a fraction the size of a conventional plant, meant to be cheaper, faster, and sited next to the customer that needs the power. The field crossed a real threshold in 2026: TerraPower’s Natrium plant in Wyoming received its US construction permit in March and began nuclear-island construction in April, the first American approval for a commercial non-light-water reactor in more than half a century. Others — NuScale, X-energy, Kairos, Oklo, Rolls-Royce SMR — are at various stages behind it.

What Is HALEU & Why Advanced Reactors Need It

Many of these advanced designs share a fuel problem. They run not on ordinary 5%-enriched uranium but on HALEU — high-assay low-enriched uranium, concentrated to just under 20% uranium-235. Higher enrichment lets a reactor be smaller, run longer between refuellings, and waste less, which is why more than half of the advanced designs under development need it. The problem is that, until recently, Russia was the only country producing HALEU commercially. After the United States banned Russian uranium imports in 2024, Western commercial supply of this specific fuel effectively vanished, and the sole US producer, Centrus Energy, had delivered only about 920 kilograms to the Department of Energy through mid-2025 — a trickle against the tonnes the new reactors will need. Washington has committed $2.7 billion to closing that gap, but HALEU is the bottleneck that could stall the whole SMR wave.

Nuclear Heat Beyond Electricity: Steel, Hydrogen & Desalination

Electricity is only half of what a reactor can sell. The other half is heat — and heat is where uranium’s next industrial market is opening. Heavy industry runs on high-temperature process heat that today comes almost entirely from burning coal and gas, and the International Energy Agency reckons industrial heat accounts for roughly a fifth of all global energy use and a large share of industrial carbon emissions. A reactor that delivers heat, not just power, aims squarely at that gap.

How Hot a Nuclear Reactor Gets

Ordinary reactors are temperature-limited: a water-cooled plant produces steam at around 300°C, fine for electricity and low-grade heat but far short of what a furnace needs. The new high-temperature gas-cooled reactors change that. X-energy’s Xe-100 runs helium out of the core at about 750°C and can deliver industrial steam near 565°C, and China’s HTR-PM, the world’s first commercial pebble-bed reactor of this kind, entered operation in December 2023. That jump in temperature is what turns a reactor from a power source into an industrial heat source.

What Is Pink Hydrogen

Point that heat and power at water and you get “pink hydrogen” — hydrogen split from water using nuclear energy. It matters because high temperature makes the splitting far more efficient: conventional electrolysis needs around 50 kilowatt-hours of electricity per kilogram of hydrogen, but high-temperature steam electrolysis uses roughly a third less energy, because the reactor’s heat does part of the work the electricity would otherwise have to. Idaho National Laboratory and Bloom Energy demonstrated the efficiency in 2022, and four US nuclear plants — starting with Constellation’s Nine Mile Point, the first to make clean hydrogen in 2023 — have been fitted with electrolysers to prove it at scale. Pink hydrogen turns a reactor into a chemical factory.

Can Nuclear Reactors Power Heavy Industry

The flagship test of that idea is a chemical plant, not a power grid. Dow and X-energy are building four Xe-100 reactors at Dow’s Seadrift site in Texas to supply the plant’s power and steam, a project the developers say will cut roughly 440,000 tonnes of carbon dioxide a year and which cleared its US environmental review in May 2026 — the first advanced-reactor construction permit sought for an American industrial site. It is the clearest sign yet that uranium’s future customers include the chemical and manufacturing industries, not only utilities.

Steelmaking, Cement & Chemicals

Honesty matters on the limits here. The very hottest industrial processes — cement kilns and blast-furnace ironmaking run well above 1,100°C — are still beyond even a high-temperature reactor’s reach, so nuclear’s near-term industrial role is chemicals, refining, process steam, and hydrogen rather than directly firing a steel furnace. Steelmaker Nucor has explored SMRs to power electric-arc furnaces with reactor electricity, which is a different play from direct heat. The realistic picture is uranium decarbonising the middle band of industrial heat first, and chipping at the highest temperatures only as the technology matures.

Nuclear Desalination

Reactor heat can also make fresh water. Thermal desalination needs only low-grade heat — well under 130°C — so even an ordinary reactor’s waste heat can drive it, making this one of the easiest non-electric uses of nuclear energy rather than the hardest. It is not new: Kazakhstan’s BN-350 reactor at Aktau produced both electricity and up to 80,000–120,000 cubic metres of drinking water a day for nearly three decades until 1999. With water scarcity worsening across the arid Gulf and North Africa, and SMRs offering right-sized heat, nuclear desalination is drawing fresh interest in exactly the regions that need it.

Do Nuclear Submarines Use Uranium?

Yes — and this is one of uranium’s most important and least-understood jobs. A nuclear-powered submarine or aircraft carrier carries a compact reactor that lets it run for decades without refuelling, and the United States alone operates on the order of 70 nuclear submarines and 11 nuclear aircraft carriers, with Russia, China, the United Kingdom, France, and India fielding their own. The reactor is what gives a submarine effectively unlimited range and lets it stay submerged for months, limited only by food and crew.

Why Naval Reactors Use Weapons-Grade Uranium

The crucial difference from a power station is the fuel. Where a civilian reactor runs on uranium enriched to about 5%, a US naval reactor runs on highly enriched uranium at around 93% uranium-235 — effectively weapons-grade. The reason is packaging: that extreme concentration puts an enormous amount of energy in a tiny, shielded core that fits inside a hull and, critically, does not need refuelling for the life of the vessel. France is the notable exception, running its naval reactors on low-enriched fuel, but the major submarine powers accept weapons-grade fuel for the compactness it buys. That choice sits at the heart of the AUKUS agreement, under which the US and UK will transfer nuclear-powered submarines to Australia — the first time weapons-grade naval reactors will go to a country without nuclear weapons, in sealed cores it cannot open.

How Long a Submarine’s Nuclear Core Lasts

The payoff of all that enrichment is longevity. A modern submarine reactor core is designed to last around 33 years — the entire service life of the boat — and a carrier’s core can run for roughly 50 years, so many of these vessels are built to be refuelled once, or never. A conventional submarine must surface or snorkel to run its diesels; a nuclear one carries its whole energy supply sealed inside from launch to retirement. That single sealed core, fuelled once, is uranium’s quiet strategic masterpiece.

Uranium in Space: Reactors, Rockets & a Plutonium Myth

Uranium is heading into space — but not everywhere people assume, and the difference is worth getting right. Where uranium genuinely belongs is in fission reactors built to run off-world. NASA and the Department of Energy proved the concept with the KRUSTY experiment in 2018, a small reactor fuelled by highly enriched uranium that ran for 28 hours, and they are now developing surface reactors to power a lunar or Martian base where sunlight is unreliable.

Is NASA Building a Nuclear Reactor on the Moon

It is trying to. NASA’s Fission Surface Power project aimed at a 40-kilowatt reactor for the Moon, and in August 2025 the agency directed an accelerated push toward a larger 100-kilowatt lunar reactor targeting launch around 2029 — explicitly framed as a race with China and Russia, which have announced their own lunar-reactor plans. Newer designs lean toward HALEU rather than weapons-grade fuel. A separate effort, the DARPA and NASA DRACO nuclear-thermal rocket, was cancelled in 2025 as falling commercial launch costs undercut its rationale — a reminder that not every nuclear-space idea survives contact with economics.

Do Space Probes Like Voyager Run on Uranium

Here is the myth to retire: the famous deep-space probes do not run on uranium. Voyager, Cassini, New Horizons, and the Perseverance and Curiosity Mars rovers are powered by radioisotope generators fuelled by plutonium-238, not uranium. Plutonium-238 is chosen because it self-heats intensely from radioactive decay for decades without any chain reaction, and no uranium isotope substitutes for that steady warmth. So while uranium may soon run a reactor on the lunar surface, the nuclear “batteries” that have carried humanity’s probes to the edge of the solar system are a different element entirely — a distinction most write-ups miss.

Was There Ever a Nuclear-Powered Airplane?

Uranium’s history in the air is a cautionary tale, and it splits into two very different things that are easy to confuse. The first is uranium as fuel.

The Nuclear Bomber That Never Flew on Its Reactor

In the 1950s the United States spent about a billion dollars trying to build a nuclear-powered bomber that could stay aloft for days. A test aircraft, the Convair NB-36H, flew 47 times carrying an operating reactor — but the reactor only ran shielding experiments and never actually powered the plane. President Kennedy cancelled the program in 1961, noting that after fifteen years and a billion dollars a useful aircraft was still remote, and intercontinental missiles had made the whole idea pointless. Uranium as an aviation fuel is a road the world tried and abandoned.

Why Old Boeing 747s Carried Depleted Uranium

The second use is uranium as dead weight, and it did fly. Early Boeing 747s, built before 1981, used depleted uranium as counterweights in their control surfaces — several hundred kilograms of it — chosen purely because uranium is extraordinarily dense and could balance a control surface in a small space. This had nothing to do with power; it was ballast. Boeing switched to tungsten from 1981, and the practice is now a historical footnote. Keep the two straight: one was an attempt to fly on uranium and failed, the other was uranium riding along as weight and has since been retired.

What Depleted Uranium Is Used For

Every kilogram of enriched reactor fuel leaves behind several kilograms of depleted uranium — the leftover uranium-238 after the fissile uranium-235 has been skimmed off. The world has accumulated a vast stockpile of it: around 1.6 million tonnes globally, growing by more than 50,000 tonnes a year, with the US Department of Energy alone holding over 700,000 tonnes. For most of the nuclear age this has been a waste-storage problem in search of uses, and the uses it has found all trade on one property.

Why Depleted Uranium Is Denser Than Lead

That property is density. Depleted uranium is about 19 grams per cubic centimetre — roughly 1.7 times as dense as lead — so it packs enormous mass into a small volume. Everything depleted uranium does industrially and militarily follows from that single fact: where you need the most weight or the most stopping power in the least space, few materials compete, and the ones that do (like tungsten) are far more expensive.

Depleted Uranium in Radiation Shielding

Its highest-value civilian use is shielding. Because it is so dense, depleted uranium stops gamma radiation in far less thickness than lead — the World Nuclear Association rates it about five times more effective than lead by volume — so it is used to shield radiotherapy equipment, industrial radiography sources, and the casks that transport radioactive material. A small amount of depleted uranium does the shielding work of a much bulkier lead casting, which matters when the shield itself has to be moved or fitted into a machine.

Depleted Uranium in Armor-Piercing Rounds & Tank Armor

Its most controversial use is military. Depleted-uranium penetrators are prized because they are “self-sharpening” — the metal shears to a fresh point as it punches through armour, rather than blunting like tungsten — and pyrophoric, igniting on impact. The same density makes depleted uranium effective as armour, and post-1998 M1 Abrams tanks carry it in their turret plating. These rounds remain in service and in the news: the United States sent 120-millimetre depleted-uranium tank ammunition to Ukraine in 2023. The health debate around them is real, and worth stating plainly — the World Nuclear Association and the World Health Organization treat depleted uranium’s chemical toxicity to the kidneys, not its weak radioactivity, as the primary hazard.

Is Uranium Used in Quantum Computing?

Not in any real machine — and it is worth saying so directly, because the honest answer is more useful than a hopeful one. The metals that actually run quantum computers are niobium and aluminium, which form the superconducting qubits, and gold, which plates the cryogenic wiring, as covered in our piece on the future uses of gold. Uranium is nowhere in that stack.

Where uranium genuinely appears is one step back, in the laboratory chemistry that might one day feed quantum technology. Uranium’s unusual electronic structure makes it a subject of research into “single-molecule magnets” and molecular qubits — the first actinide single-molecule magnet, reported in 2009, was a uranium compound — because its electrons can store quantum information in ways researchers want to understand. That work is real and cited, but it is basic science years from any device, and uranium’s radioactivity makes it an unlikely material for a shipped product. The accurate framing is that uranium is a curiosity in quantum research, not a component of quantum computers.

Do Robots Use Uranium?

No. A humanoid or industrial robot runs on lithium-ion batteries, rare-earth magnets, copper, and silicon — the same materials detailed in our teardown of the metals inside a humanoid robot — and there is no uranium anywhere in it. If you came looking for uranium powering robots, the straight answer is that it does not, and no serious design proposes it.

The only honest connections are indirect, and worth naming rather than inflating. Uranium may power the grid and the microreactors behind the machines — the electricity that charges a robot fleet or runs an automated factory can come from a reactor — and robots are increasingly the tool used to decommission reactors, working in radioactive spaces too dangerous for people. In other words, uranium does not go inside a robot; it may supply the power behind one, and robots help clean up after it. That is the whole of the relationship, and pretending otherwise would be exactly the kind of stretch this site avoids.

How Uranium Makes Medical Isotopes

Uranium’s most surprising everyday role is in the hospital. The single most-used tool in nuclear medicine is technetium-99m, the radioactive tracer behind roughly 80% of all nuclear-medicine scans — the bone scans, cardiac scans, and cancer-staging images that add up to more than 50 million procedures worldwide every year. Almost no patient realises where it comes from.

It comes from splitting uranium. Technetium-99m is generated from a parent isotope, molybdenum-99, which is produced mostly by fissioning uranium-235 targets in a handful of research reactors, then shipped rapidly to hospitals because it decays within days. So the same nuclear reaction that lights a city also produces the tracer coursing through tens of millions of medical scans a year — a genuinely large, genuinely hidden use of uranium. For non-proliferation reasons the industry has been shifting these targets from highly enriched to low-enriched uranium, but the underlying fact holds: uranium is quietly essential to modern diagnostic medicine.

Do We Have Enough Uranium?

The world currently mines less uranium than its reactors consume. Mine production was about 60,200 tonnes in 2024 against reactor requirements near 68,900 tonnes, so mining meets only around 90% of demand; the rest has long been filled from secondary sources — civilian stockpiles, reprocessed material, and, for two decades, Russian warheads dismantled and down-blended into reactor fuel under the “Megatons to Megawatts” program. That program ended in 2013, and the secondary cushion that once covered the gap is thinning, which is the real supply story.

How Much Uranium Is Left in the Ground

Geologically, there is a great deal. Identified resources recoverable at reasonable cost stand at roughly 5.9 million tonnes — about 90 years of supply at current consumption, per the World Nuclear Association, and more becomes economic as prices rise. The distribution holds a quiet irony: Australia sits on about 28% of known resources but mines relatively little, while Kazakhstan, with a fraction of the reserves, is the world’s dominant producer. Running low on uranium in the ground is not the near-term problem.

Can You Extract Uranium From Seawater

And there is a near-limitless backstop: the ocean. Seawater holds an estimated 4.5 billion tonnes of dissolved uranium — on the order of 500 times all identified land resources — at a concentration of about 3.3 parts per billion. Researchers at Oak Ridge and, more recently, in China have shown it can be pulled out with special adsorbent fibres that grab uranium ions from passing water, with newer methods steadily cutting the cost. It is strictly laboratory and pilot work today, with no commercial operation and China’s own roadmap pointing to mid-century, so it is not a source now. But its importance is as a price ceiling: because the ocean’s uranium is effectively inexhaustible at a known, bounded cost, the world can run short of cheap uranium but can never truly run out — and since fuel is only about 5% of the cost of nuclear electricity, even expensive uranium barely moves the price of power.

How Uranium Is Mined

Most uranium no longer comes from a pit or a shaft at all. The largest single method is now in-situ recovery, which accounts for about 52% of world production, followed by conventional open-pit and underground mining at around 44%, with a few percent recovered as a by-product of copper and gold mining. How uranium is mined depends almost entirely on how the deposit sits in the ground.

What Is In-Situ Recovery & Why Kazakhstan Uses It

In-situ recovery barely looks like mining. Instead of digging, operators pump a mild solution down wells into a uranium-bearing sandstone layer, dissolve the uranium underground, and pump the liquid back up to recover it — no pit, no shaft, no tailings pile. It works only where the geology cooperates, and Kazakhstan’s permeable, uranium-rich sandstones are close to ideal, which is why the country runs almost entirely on in-situ recovery and does it more cheaply than anyone. That low cost is a big part of why Kazakhstan dominates world supply.

How Rich Is High-Grade Uranium Ore

The opposite extreme is Canada’s Athabasca Basin, home to the richest uranium ore on Earth. Where a typical deposit runs around 0.1% uranium, the Athabasca’s unconformity deposits grade up to about 20% — a hundred times richer — which is why two Canadian mines, McArthur River and Cigar Lake, are among the most productive in the world despite being conventional underground operations. Grade is destiny in uranium mining: a percent or two of uranium in the rock changes the entire economics.

Which Countries Produce the Most Uranium

Production is heavily concentrated. Kazakhstan alone supplies about 39% of the world’s mined uranium, with Canada around 24% and Namibia about 12%; the top six countries together account for roughly 90%. That concentration carries geopolitical risk, sharply illustrated by Niger, once a major supplier to Europe, whose output collapsed about 68% between 2019 and 2024 across a 2023 coup and the loss of Western operator access. A market that leans this hard on a few producers is exposed to any one of them stumbling.

How Uranium Becomes Reactor Fuel

Uranium out of the ground is nowhere near ready to power a reactor. Between the mine and the fuel rod sit four industrial steps, and the value — and the geopolitical leverage — concentrates in the middle of them.

From Yellowcake to Fuel Rods

Milling: Making Yellowcake (U₃O₈)

At or near the mine, ore is crushed and chemically leached to extract the uranium, which is dried into a coarse powder of uranium oxide — “yellowcake,” or U₃O₈. This is the form uranium is actually bought and sold in, and the “uranium price” quoted in markets is the price of a pound of yellowcake.

Conversion: Turning Yellowcake Into UF₆ Gas

Next, yellowcake is converted into uranium hexafluoride, or UF₆, the only common uranium compound that becomes a gas at low temperature — a necessary form because the enrichment step works on a gas. Conversion is itself a chokepoint: only a handful of plants worldwide do it, and prices for the service spiked as the West moved away from Russian supply. The sole US conversion plant, Metropolis Works in Illinois, restarted in 2023 after years idle and was spun into a new owner, Solstice, in 2025 — part of a broader push to rebuild the Western front end.

Enrichment: Raising the U-235 (LEU & HALEU)

Enrichment is the heart of the fuel cycle and the step that matters most. Thousands of fast-spinning centrifuges separate the slightly lighter uranium-235 from uranium-238, raising the fissile share from 0.7% to the 3–5% a reactor needs (low-enriched uranium), or to just under 20% for HALEU. The work is measured in “separative work units,” and it is the most capital-intensive, technically guarded, and geographically concentrated part of the whole chain — which is exactly why enrichment, not mining, is where the supply chain truly bottlenecks.

Fabrication: Pellets, Rods & Assemblies

Finally the enriched uranium is turned into ceramic pellets, stacked into zirconium-alloy fuel rods, and bundled into fuel assemblies engineered to each reactor’s exact design. Only then is uranium ready to generate power — the fingertip pellet from the opening of this article is the end product of all four steps.

Why Russia Controls Uranium Enrichment

The uncomfortable fact is that Russia dominates the most critical step. Rosatom controls roughly 44% of the world’s uranium-enrichment capacity, per the World Nuclear Association — more than any other country by far. That dependence is why the United States banned Russian enriched-uranium imports in 2024 and paired the ban with billions in funding to rebuild domestic enrichment, and why Europe’s Urenco and France’s Orano are expanding. Enrichment is where uranium’s strategic value really lives.

Why the West Is Reshoring Uranium Supply

The result is a scramble to rebuild an allied supply chain that bypasses both Russian enrichment and unstable suppliers like Niger. Western governments and companies are restarting conversion at Metropolis in the US and running Cameco’s Port Hope in Canada and Orano’s plants in France, expanding enrichment across the US and Europe, and leaning on a “friendshoring” group of the US, Canada, Australia, and France. After decades of treating the fuel cycle as a low-priority commodity business, the West is reshoring the front end of nuclear power as a matter of energy security.

Can Nuclear Waste Be Recycled Into Fuel?

Here is the fact that reframes the entire nuclear-waste debate: when fuel comes out of a reactor as “spent,” more than 90% of its energy is still in it, according to the US Department of Energy. Used fuel is about 96% uranium and roughly 1% plutonium, with only a few percent genuine fission-product waste. Most of what is called “waste” is really unused fuel, and whether to recycle it is one of the central choices in uranium’s future.

What Is MOX Fuel & Who Makes It

Recycling that fuel is already done commercially, chiefly in France. Reprocessing separates the reusable uranium and plutonium from the true waste and blends them into mixed-oxide, or MOX, fuel that goes back into reactors. Orano’s La Hague plant, licensed to reprocess about 1,700 tonnes of used fuel a year, is the world’s largest, and nearly all commercial MOX is made in France. The United States, by contrast, chose in 1977 not to reprocess and instead sends its used fuel toward direct disposal — a policy choice, not a technical necessity, and one increasingly questioned as the recycling case strengthens.

Can Reactors Run on Depleted Uranium

The bigger prize is a reactor that runs on the leftovers. Fast neutron reactors can extract up to 60 times more energy from uranium than today’s reactors by converting fertile uranium-238 into fissile plutonium as they run — which means the world’s 1.6-million-tonne mountain of depleted uranium, plus its spent fuel, could in principle power civilisation for centuries. This is not only theory: Russia’s BN-600 and BN-800 have operated as commercial-scale fast reactors for decades, China started its CFR-600 in 2023, and India brought a prototype fast breeder to criticality in early 2026. One honest caveat matters here — the much-hyped TerraPower Natrium being built in Wyoming is a fast reactor, but the demonstration unit is a HALEU burner, not a waste-eater; the ability to run on depleted uranium is design heritage and future ambition, not what the first plant will do.

Why Thorium Reactors Still Need a Uranium Seed

Even the great alternative to uranium depends on it. Thorium is often floated as a cleaner nuclear fuel, but thorium is “fertile,” not fissile — it cannot sustain a chain reaction on its own and must be bombarded with neutrons from a fissile “seed” of uranium-235 or plutonium to breed into usable fuel. Uranium is the spark that lights any thorium cycle. China’s experimental thorium molten-salt reactor in the Gansu desert, which achieved thorium-to-uranium fuel conversion in 2025, and India’s long-running three-stage program both start from uranium and plutonium. Thorium may be nuclear’s future, but uranium is the match that has to strike first.

What Is a Fusion-Fission Hybrid Reactor

At the speculative frontier sits a stranger idea: using a fusion reaction purely as a neutron source to drive a surrounding “blanket” of uranium or thorium, multiplying the energy output many times over without the fission core going critical on its own. China announced plans for Xinghuo, billed as the world’s first fusion-fission hybrid power plant, targeting operation around 2030 — but this is an announced megaproject, not a demonstrated technology, and hybrids have lived mostly on paper since the 1950s. It belongs in the story as a possibility to watch, honestly labelled as unproven, alongside the superconducting magnets that make fusion conceivable at all in our piece on REBCO superconductors and fusion.

The Biggest Uranium Mining Companies

Uranium production concentrates into a short list of companies, most of them state-linked, and their decisions move the market as much as any geological fact. The field, as it stands in mid-2026, with 2024 output as the yardstick:

Two currents run beneath the table. One is geopolitical fragility: Orano’s expulsion from Niger shows how fast a national producer can lose a mine, and Kazatomprom’s dominance means its production choices set the tone for the whole market. The other is a wave of Western developers racing to add supply — Canada’s NexGen Energy won a construction licence for its high-grade Rook I project in March 2026, Denison Mines is advancing an in-situ operation, and US producers like Uranium Energy Corp and Energy Fuels are restarting domestic mines — though a new mine can take a decade to reach full output, which is the crux of the supply problem the market now faces.

Why Uranium Supply Can’t Keep Up With Demand

High prices are supposed to summon new supply, but uranium responds slowly. A new mine takes 10 to 15 years — often longer — to move from discovery through permitting to production, so today’s price does little to change this decade’s output. And miners will not commit that capital until they trust the price: even with uranium’s long-term price near $80 a pound in 2025, Kazatomprom judged that demand did not yet justify returning to full capacity, and analysts peg the “incentive price” needed to build new greenfield mines at roughly $75 to $90 a pound.

Layered on top are the market’s own scars. A decade of low prices after 2011 idled mines that are slow and costly to restart, in-situ operations depend on chemicals like sulphuric acid that have run short, and the secondary supply that once padded the gap is thinning. The result is a supply curve that barely bends to demand on any timescale investors care about — the same structural inelasticity that has kept the market in a slow-burning deficit even as reactor demand climbs.

How Much Uranium the World Will Need by 2050

For anyone weighing uranium as an investment, the demand trajectory is the whole argument — and every credible forecaster has been revising it upward. The World Nuclear Association’s 2025 fuel report sees reactor uranium requirements roughly doubling by 2040 in its reference case, while the International Atomic Energy Agency, raising its projection for a fifth straight year in September 2025, sees installed nuclear capacity reaching as much as 992 gigawatts by 2050 — about 2.6 times today’s level. Set against a mine industry that can barely cover 90% of current demand, those curves are the reason uranium draws investor attention at all.

Two honesty notes on that table. The capacity figures are the forecasters’ own; the uranium-requirement figures are hard-sourced only to 2040, so the 2050 tonnage cells are an estimate, scaled from installed capacity at roughly 170 tonnes of uranium per gigawatt per year — a reasonable rule of thumb, not a published number. And these are projections, not promises: they assume reactors get built roughly on schedule, which the history of nuclear construction says is optimistic. What is not in serious dispute is the direction. More than twenty countries pledged at COP28 to triple nuclear capacity by 2050, China is building roughly half the world’s new reactors, and even the low case has demand rising. The supply side, meanwhile, cannot move nearly as fast — which is the entire investment thesis in one sentence.

How to Invest in Uranium

Uranium is unlike gold or silver in one fundamental way, and it shapes everything about investing in it: you cannot legally hold the physical metal. What follows is market analysis for information only, not investment advice.

Can You Buy Physical Uranium

Not as an individual, no. Uranium oxide is a regulated nuclear material — owning, storing, and moving it requires licensing, and it must sit in approved facilities — so there is no equivalent of stacking gold coins in a safe. The way investors get “physical” exposure is through listed trusts that hold the metal in licensed storage on their behalf. The largest is the Sprott Physical Uranium Trust, which holds tens of millions of pounds of uranium oxide and, by buying pounds off a thin spot market and sitting on them, has itself become a force tightening supply; Britain’s Yellow Cake plc plays a similar role. Buying these is buying a claim on vaulted pounds, not metal you can ever take home.

Uranium ETFs Compared: URA, URNM, URNJ & NLR

Most equity investors reach for a fund, and they differ mainly in what kind of uranium exposure they hold. Moving along the spectrum trades pure metal exposure for company and sector exposure:

Can You Hold Uranium in an IRA

You can hold uranium exposure in a retirement account, but not the metal. Uranium ETFs and mining shares are ordinary securities that sit in any standard brokerage IRA like any other stock. What you cannot do is put physical uranium in a precious-metals IRA — those are limited by tax rules to specific gold, silver, platinum, and palladium bullion, and uranium is neither a precious metal nor something an individual can take licensed possession of. To hold “uranium” for retirement is to hold the trusts or the equities, never the element.

The exposures carry real and specific risks. Uranium is famously boom-and-bust — it spiked near $140 a pound in 2007 and then languished for a decade, and ran to roughly $106 in early 2024 before pulling back toward the mid-$80s by mid-2026 — so the timing risk is severe. The physical trusts can trade well above or below the value of the pounds they hold; SMR and new-build timelines slip, so demand narratives can run ahead of actual reactors; and the supply base is concentrated in Kazakhstan, Russia, and other single points of failure. There is also a distinct fuel-cycle play in enrichment names like Centrus, a bet on the HALEU bottleneck rather than on the metal itself.

The Outlook for Uranium Demand

Uranium’s future is a tug-of-war, and it is worth seeing honestly from both sides. Pulling demand up: a reactor fleet being extended and restarted, an SMR wave reaching construction, the AI data-centre power crunch, and entirely new uses in process heat and hydrogen — a broader set of customers than uranium has ever had. Pulling the other way: the chronic tendency of nuclear projects to slip and overrun, cheap and fast-building renewables competing for the same clean-power demand, and the possibility that fuel recycling, fast reactors, and one day seawater extraction stretch each tonne of uranium much further.

The base case does not need the moonshots. Uranium’s core advantage — dense, carbon-free, around-the-clock energy that nothing else matches — is enough on its own to keep reactors being built, and the supply side simply cannot expand as fast as even conservative demand forecasts. Layer a structural mine deficit and a concentrated, choke-pointed fuel cycle on top of steadily rising demand, and the metal that spent decades out of fashion looks set to stay strategically tight well into the century it is helping to power.

A Short History of Uranium

Future Uses of Uranium FAQ

Related Articles