Tritium, the radioactive isotope of hydrogen, sells for roughly $30,000 per gram, making it one of the most expensive substances produced at industrial scale. That price reflects a convergence of factors: nature produces almost none of it, reactor-based manufacturing is slow and indirect, separating it from other hydrogen isotopes is energy-intensive, and storing it safely requires specialized infrastructure that must account for its radioactive decay and its tendency to escape through solid metal. With fusion energy companies now competing for a supply measured in mere kilograms worldwide, the economics of tritium are only getting tighter.
Nature Makes Almost None
Tritium does exist in the natural world, but in vanishingly small quantities. Cosmic rays striking the upper atmosphere knock neutrons loose from nitrogen and oxygen nuclei, and a fraction of those neutrons interact with other atoms to produce tritium. A study estimating the mean global production rate from cosmic rays arrived at about 0.2 tritium atoms per square centimeter per second across the planet’s surface, strongly suggesting that natural tritium comes almost entirely from this single mechanism.1Geophysical Research Letters. Natural tritium deposition over Antarctica and estimation of the mean global production rate That trickle adds up to only a few kilograms present in the entire global environment at any given time, and it is so dilute in rainfall and ocean water that extracting it would cost far more than producing it artificially. So from the start, any commercial supply has to be manufactured.
How Tritium Is Actually Produced
The world’s primary source of tritium today is the heavy-water nuclear reactor, specifically the CANDU design operated in Canada, South Korea, and a handful of other countries. In these reactors, the moderator and coolant are heavy water, meaning water made with deuterium instead of ordinary hydrogen. When deuterium atoms in that heavy water absorb neutrons during normal reactor operation, a small fraction undergo a reaction that converts them into tritium.2Nuclear Engineering and Design. Tritium: its generation and pathways to the environment at CANDU 6 generating stations The tritium then has to be removed from the heavy water, both because it is a radiological hazard if it leaks and because recovered tritium is the commercial product.
The United States, which does not operate CANDU reactors, takes a different approach for its defense stockpile. Tritium-producing burnable absorber rods, known as TPBARs, are loaded into existing light-water reactors. These rods contain lithium, which captures neutrons and transmutes into tritium. Either way, production is a byproduct of operating a nuclear reactor, and you cannot simply scale it up the way you would a chemical plant. You need neutrons, and neutrons come from fission, which means production is tied to reactor schedules, fuel cycles, and regulatory approvals.
The Separation Problem
Once tritium has been generated inside a reactor’s heavy water or lithium targets, getting it out in pure form is its own engineering challenge. Tritium is chemically identical to ordinary hydrogen and nearly identical to deuterium. All three isotopes form the same bonds, dissolve in the same solvents, and behave the same way in most chemical reactions. The only differences are in mass, and those differences are small enough that separating tritium from deuterium requires extreme methods.
The standard industrial technique is cryogenic distillation, which exploits the tiny differences in boiling point between hydrogen isotopes by cooling them to temperatures near absolute zero and distilling them in tall columns. This process demands enormous energy input and highly specialized equipment. Research into optimizing cryogenic distillation has shown that lowering equilibrator temperatures can improve separation efficiency and reduce the number of distillation stages required, but the fundamental difficulty remains.3Systems and Control Transactions. Simulation and Optimisation of Cryogenic Distillation and Isotopic Equilibrator Cascades for Hydrogen Isotope Separation Processes in the Fusion Fuel Cycle Traditional cryogenic distillation still suffers from low selectivity and high energy consumption, which is why researchers continue to explore alternative separation methods.4PubMed. Comprehensive Design and Experimental Protocol for Scalable and Temperature-Controllable Cryogenic Hydrogen Isotope Separation
Every step of this separation adds cost. The cryogenic equipment itself is expensive, the energy bills are steep, and the entire system has to be designed to handle radioactive tritium safely. There is no cheap shortcut: if you want pure tritium, you pay for the thermodynamics of pulling it away from its nearly identical siblings.
Storing and Handling a Gas That Escapes Through Metal
Even after you have pure tritium, keeping it is expensive. Tritium is radioactive with a half-life of about 12.3 years, meaning any stockpile loses roughly 5.5% of its mass each year to decay. You cannot buy a supply and sit on it indefinitely; the clock is always running. This alone makes tritium fundamentally different from most expensive materials. Gold and platinum sit in vaults for decades without losing an atom. Tritium is constantly disappearing.
Beyond decay, tritium has a remarkable ability to permeate through solid materials, including the stainless steel typically used for containment. Austenitic stainless steels are the standard choice because they resist the mechanical degradation that hydrogen isotopes cause in most metals, but even these suffer over time. Long-term exposure to tritium produces helium inside the steel through radioactive decay, and that helium accumulation can reduce fracture toughness by as much as 95%.5Fusion Engineering and Design. Tritium embrittlement of austenitic stainless-steel tubing at low helium contents That means containment vessels and piping have limited service lives and must be monitored, tested, and eventually replaced.
In environments where tritium is at elevated temperatures and pressures, like the breeding blankets being designed for fusion reactors, permeation through containment walls becomes a serious engineering problem. Research on tritium barriers has found that tritium can permeate through stainless steel and similar alloys at appreciable rates, leading to losses on the order of a couple of grams per day in some reactor blanket designs. Multilayer coatings using materials like alumina and chromium-aluminum alloys show promise for blocking permeation, but their effectiveness degrades under reactor-like radiation and mechanical stress.6PubMed Central. CrAl/Al2O3 multilayer as an effective Tritium/Hydrogen barrier All of this translates into specialized facilities, constant monitoring, robust ventilation and air-cleanup systems, and a regulatory burden that adds still more to the cost of every gram.
Who Buys Tritium at These Prices
Given the cost, tritium is purchased only when no alternative exists. The three main demand sectors are defense, scientific research, and commercial illumination products.
Military programs are the largest single consumer. Thermonuclear weapons require tritium to boost the yield of their fission primaries, and because of the 12.3-year half-life, warheads need periodic tritium replenishment to remain operational. The U.S. Department of Energy manages this through a dedicated production program, and the exact quantities involved are classified, but maintaining a nuclear arsenal is a permanent and substantial draw on the global supply.
In research, tritium serves as a tracer in biological and pharmaceutical studies because it behaves chemically like hydrogen but can be detected through its radioactive emissions. It is also the fuel for experimental fusion devices like ITER, which plan to burn deuterium-tritium plasma. On the commercial side, tritium is used in self-luminous exit signs, watch dials, and gun sights. In these products, tritium gas is sealed inside tiny glass tubes coated with phosphor, producing a glow that lasts for years without any external power source. These consumer applications use only milligrams at a time, but they still pay the going rate per gram.
The Global Supply Is Measured in Kilograms
One of the most striking facts about tritium economics is just how little exists. The entire world’s usable stockpile, accumulated over decades of CANDU reactor operation and defense production, is measured in tens of kilograms. Modeling of global supply scenarios estimates that peak theoretical stockpiles, accounting for contributions from all existing production pathways, would land between about 35 and 48 kilograms.7Fusion Engineering and Design. Tritium supply and demand for fusion: Socioeconomic scenario modelling That is the total, spread across national defense reserves, research stockpiles, and commercial inventories. For context, 48 kilograms is about the weight of a carry-on suitcase’s worth of water.
This tiny supply already faces known demand. ITER alone is expected to need roughly 16 kilograms over its operational life, which supply models suggest can be met under all plausible production scenarios.7Fusion Engineering and Design. Tritium supply and demand for fusion: Socioeconomic scenario modelling The trouble starts when multiple demonstration-class fusion reactors and private-sector machines enter the picture. Under pessimistic supply and high-demand assumptions, shortfalls could exceed 29 kilograms by the mid-2050s, rise to 42 kilograms by 2070, and balloon to 181 kilograms by 2070 if fusion plants need even a modest external top-up because their internal breeding systems are not perfectly efficient.7Fusion Engineering and Design. Tritium supply and demand for fusion: Socioeconomic scenario modelling And if fusion were ever to supply a meaningful fraction of global electricity, the scale changes entirely: providing 10% of global power by 2100 could require annual tritium consumption of around 760 tonnes, a figure so far beyond current production capacity that entirely new supply chains would be necessary.
Fusion’s Chicken-and-Egg Problem
Fusion energy is supposed to eventually produce its own tritium through breeding blankets, lithium-containing structures that surround the reactor and capture the neutrons released by the fusion reaction. When a neutron hits lithium-6, it produces tritium and helium. In theory, a well-designed blanket can produce slightly more tritium than the reactor consumes, a condition described by a tritium breeding ratio greater than one. Researchers have been working to verify this in practice. One detailed assessment of a water-cooled lithium-lead blanket design calculated a total breeding ratio of about 1.14, which is close to but still slightly below the design target of 1.15 needed to guarantee self-sufficiency.8Fusion Engineering and Design. Nuclear analyses in support of the water-cooled lithium lead breeding blanket design development Other blanket concepts using helium cooling and lithium-lead as both the breeding material and neutron multiplier are being studied with the explicit goal of achieving a breeding ratio above one to ensure steady-state tritium supply.9Journal of Physics: Conference Series. Studied of Tritium Breeding Ratio on HCLL (Helium Cooled Lithium-Lead) Fusion Blanket by Using MCNP Program
But even if breeding blankets work perfectly from day one, every fusion reactor still needs a startup inventory of tritium before it can begin breeding. An analysis of a 500-megawatt thermal fusion plant estimated a baseline startup inventory of roughly 327 grams, with an operating reserve of about 642 grams needed to sustain 24 hours of continuous operation through internal recycling alone.10Fusion Engineering and Design. Approach to startup inventory for viable commercial fusion power plant At $30,000 per gram, 642 grams represents about $19 million in fuel cost before the reactor generates a single watt of power. That is manageable for a single plant. But if dozens of fusion startups are all trying to commission reactors in the same decade, they are all drawing from the same global stockpile of a few tens of kilograms. The result is a textbook scarcity problem: limited supply, growing demand, and no rapid way to increase production.
This is why the breeding blanket is not just a technical feature but an economic necessity. If fusion reactors cannot reliably breed their own fuel, the industry faces permanent dependence on reactor-produced tritium at prices that would make fusion electricity uncompetitive. The margins here are thin. A breeding ratio of 1.14 instead of 1.15 may sound like a rounding error, but over years of operation it determines whether a plant is tritium-positive or slowly running out of fuel.
Why the Price Is Unlikely to Drop Soon
Several structural factors work against any significant price decrease in the near term. First, CANDU reactors are aging. Canada’s fleet, historically the world’s largest source of commercial tritium, is not being replaced at the same rate as plants retire. Ontario Power Generation’s Darlington tritium removal facility is the single most important extraction site globally, and any disruption there would tighten supply immediately.
Second, there is no factory that produces tritium on demand. Every gram is a byproduct of running a nuclear reactor for other purposes, which means production responds to electricity demand and reactor scheduling, not to the tritium market. You cannot simply commission a new tritium plant the way you might build a new semiconductor fab. The infrastructure is nuclear, the lead times are measured in decades, and the regulatory approvals are among the most rigorous on Earth.
Third, that 5.5% annual decay rate means the stockpile is constantly shrinking even when no one buys any. If production merely matches consumption, the inventory still declines. Maintaining a stable stockpile requires production to outpace both sales and decay, and right now the margin is thin.
Finally, the safety and regulatory overhead is irreducible. Tritium is a radioactive material that becomes part of water when it oxidizes, making it biologically mobile and environmentally persistent in ways that other isotopes are not. Handling facilities need tritium-rated ventilation, double-walled containment, real-time air monitoring, and extensive waste-management infrastructure. None of that gets cheaper with time. Experimental validation of blanket breeding performance, for instance, requires dedicated neutron irradiation campaigns that are themselves expensive and time-consuming.11Nuclear Fusion. Neutronics experiment of tritium breeding in supercritical CO2 cooled lithium-lead blanket mock-up by D–T neutron irradiation
Tritium Versus Other Expensive Substances
At $30,000 per gram, tritium is expensive but not record-setting in the world of rare isotopes and specialty materials. Californium-252, a neutron source used in cancer treatment and industrial applications, has been priced at millions per gram. Antimatter, if you could buy it, would dwarf everything else. What makes tritium unusual is that it is expensive and needed in bulk. Most exotic isotopes are used in microgram or nanogram quantities. Tritium is needed in hundreds of grams for a single fusion reactor, kilograms for a national defense program, and potentially tonnes for a future energy grid. The intersection of high unit cost and large required quantities is what makes its economics uniquely challenging.
For self-luminous products like exit signs and watch dials, the amount of tritium involved is small enough that the per-unit cost is manageable. A typical tritium exit sign contains about 25 curies of activity, which corresponds to a fraction of a gram. The retail price of the sign absorbs the tritium cost alongside the phosphor tubes, housing, and regulatory compliance. But even in these small-scale applications, alternatives are being explored. Promethium-147 and nickel-63 have been investigated for use in betavoltaic batteries and self-powered devices, though each comes with its own set of tradeoffs in half-life, energy density, and shielding requirements.
What Happens If Fusion Succeeds
The real wild card in tritium economics is the success or failure of fusion energy. If fusion remains a research endeavor limited to a handful of experimental machines, the current supply chain can probably cope. ITER’s needs are large but finite, and the existing stockpile appears sufficient to cover them. The economic pressure on tritium pricing would remain roughly where it is today.
But if fusion works and begins to scale commercially, the picture changes radically. Private fusion companies are already investing heavily in deuterium-tritium reactor designs, and each of those companies will need startup tritium that can only come from the existing global inventory.10Fusion Engineering and Design. Approach to startup inventory for viable commercial fusion power plant The modeling suggests that without breakthroughs in supplementary production, like optimized TPBAR systems in light-water reactors or entirely new production pathways, tritium availability could become the binding constraint on how fast fusion energy can be deployed.7Fusion Engineering and Design. Tritium supply and demand for fusion: Socioeconomic scenario modelling In that scenario, $30,000 per gram starts to look like a bargain compared to what a competitive market for a genuinely scarce strategic resource could produce.
Some fusion concepts sidestep the problem entirely by using fuels that do not require tritium, such as deuterium-helium-3 or proton-boron-11 reactions. These alternatives are far harder to achieve in practice because they require higher temperatures and produce less energetic reactions, but their appeal grows in direct proportion to how intractable the tritium supply problem turns out to be. The cost of tritium, in other words, is not just a line item in a budget. It is shaping the strategic direction of an entire energy technology.