Approved Article With Interlinks

The fusion energy industry is pulling in billions of dollars in private and public capital, and yet almost nobody talking about it is pricing in a brutal physical reality: the fuel these machines need to switch on barely exists anywhere on Earth.

That fuel is tritium, a rare radioactive form of hydrogen. Fusion power depends on it, but supply, production, and self-generation form a tangle of limits that most deployment forecasts quietly ignore.

As of 2026, the entire global civilian stockpile of tritium sits below 30 kg, and it decays away at roughly 5.5% every single year. That combination makes tritium supply the quiet, deciding bottleneck for any credible commercial fusion timeline.

Here is the framework you need for understanding this hard supply limit, and for identifying which obscure corners of the nuclear industry actually hold the keys to fusion becoming a real business rather than a permanent science project.

The microscopic math of global tritium inventory

Start with the numbers, because they are almost absurd. According to the Federation of American Scientists (FAS), summarising Fusion Industry Association (FIA) data in an August 2026 analysis, global civilian tritium stocks amount to just 25-30 kg. That is the total, worldwide, for the entire fusion ambitions of the human race.

Now set that against demand. The same FAS brief notes that commissioning and starting up a single commercial fusion reactor may require up to 10 kg of tritium. One plant. Roughly a third of everything that exists.

The FAS tritium capacity analysis puts the global civilian stockpile at 25-30 kg, with a single commercial reactor requiring up to 10 kg for startup and more than 55 kg per year for operation, numbers that make the supply constraint concrete rather than theoretical.

The operating figures are worse. A 2025 Kleinman Center analysis estimates that a 1 GW fusion reactor could consume nearly 55-56 kg of tritium per year if it is not breeding its own fuel. Read that again: a single plant’s annual appetite exceeds the entire global civilian inventory, twice over.

Then there is the clock. Tritium has a half-life of about 12.3 years, which means the stockpile shrinks by roughly 5.5% annually whether anyone touches it or not. There are no meaningful natural reservoirs to top it back up.

This is the part that should reset your thinking. The supply is not just small; it is actively rotting away while you read this.

DOE Office of Science warning (2022) Presenter Pearson noted there is “not enough external tritium to support any fusion program beyond the start-up of the first (or perhaps the first few) commercial reactors.”

That single sentence sets the ceiling on the entire industry’s near-term ambitions. Eurofusion analyses echo it, warning that Canadian production after ITER “may not be sufficient to start DEMO.”

What this tells you is blunt. The fusion sector is racing against a biological half-life, and external supply can realistically fuel only the first one to three commercial plants on the planet.

So when you see a fusion company promising rapid, multi-plant fleet rollouts through the 2030s, you now have the numbers to be sceptical. Fuel security, not plasma physics, is quietly deciding which companies actually cross the commercial finish line.

Why self-sustaining reactors cannot solve the startup problem

The standard rebuttal to all of this is that fusion reactors make their own fuel. The technology is called a breeding blanket, and understanding why it does not rescue the near-term picture is the single most useful piece of technical fluency you can carry into any fusion investment conversation.

Here is the mechanism in plain terms. A breeding blanket is a layer of lithium-containing material wrapped around the fusion chamber. High-energy neutrons flying out of the plasma strike lithium-6 atoms, triggering the reaction ⁶Li + n → ⁴He + T, which produces fresh tritium inside the blanket.

For a reactor to become fuel self-sufficient, it needs what physicists call a Tritium Breeding Ratio (TBR) above 1. That means each fusion neutron must ultimately generate more than one tritium atom, once you account for losses and ongoing decay.

Sounds elegant. The problem is the order of operations.

The sequencing paradox

The breeding chain runs in a strict sequence that cannot be shuffled:

1. External tritium is loaded to provide the starting fuel.
2. The plasma ignites and begins fusing that fuel.
3. Fusion generates the high-energy neutrons.
4. Those neutrons hit the blanket and breed new tritium.
5. The reactor eventually reaches self-sufficiency.

Look at step one. You cannot breed tritium until neutrons are flying, and neutrons only fly once the plasma is already burning, and the plasma only burns because you fed it external tritium first. The blanket produces nothing on day one.

This creates an unavoidable gap period. Every new facility must first draw down the shrinking global stockpile before it can contribute a single gram back to it.

That gap is not a rounding error. It is the entire near-term supply crisis.

And this assumes the technology even works at scale, which nobody has yet shown. There is no completed demonstrator anywhere on Earth proving closed fuel-cycle operation.

Prof. Jens Linow put it starkly in a 2026 Euronews feature, noting that tritium regeneration inside a power station has “never been tried before” and that “there is no demonstrator; there is no laboratory experiment in which this has been successfully tested.” ITER’s test blanket modules are only planned for the 2030s.

What this means for you as an investor is direct. You cannot lean on internal fuel generation to solve a supply problem that bites in the 2020s and 2030s, because breeding is a post-2030 hope at best.

That forces your attention onto the external supply chain, which is where the real near-term story lives. It also gives you a sharp question for any fusion company pitch: show me your tritium procurement strategy, not just your plasma physics.

For readers wanting to pressure-test the breeding blanket economics before accepting any fusion company’s fuel cycle projections, our dedicated guide to the lithium-6 supply chain covers sourcing concentration risks, geopolitical export controls, and the enrichment capacity constraints that sit upstream of tritium breeding.

The CANDU choke point and the fragility of current supply

If external supply is the only game for the next decade, where does it actually come from? The answer is unsettling in its narrowness.

Almost all commercially available tritium is an incidental byproduct of heavy-water moderation in CANDU fission reactors, the Canadian-designed plants that use heavy water to sustain their reactions. Tritium accumulates in that water over years of operation, and specialised facilities extract it.

That is it. According to DOE briefings, CANDU reactors are effectively the only commercial source on the planet. No large-scale dedicated tritium production infrastructure exists outside CANDU operations and the US military weapons programme, and that military stock is off-limits for civilian fusion.

The volumes are tiny. The Kleinman Center’s October 2025 report, citing the Canadian Nuclear Association, states that roughly 30 CANDU-style reactors globally generate less than 4 kg of tritium per year in total. Canadian output specifically sits around 2 kg per year.

Set that against startup demand of up to 10 kg per plant, and you see the trap. Global annual production would take years to fuel a single reactor, and decay eats into the stockpile the whole time.

Worse, this supply is aging out. Production depends entirely on fleet-management choices: refurbishment schedules, life-extension decisions, and whether these decades-old reactors keep running at all. No major expansion of civilian tritium production has been announced since 2024.

Key extraction facilities

Two facilities do most of the world’s heavy lifting.

The Darlington Tritium Removal Facility in Ontario, operated by Ontario Power Generation, is the principal commercial extraction site globally. It pulls tritium from the moderator water across OPG’s CANDU fleet, not just Darlington’s own units. Its future output hinges on Ontario’s refurbishment and life-extension decisions running into the 2030s and beyond.

The Wolsong Tritium Removal Facility in South Korea is the other major civilian extraction site. Together, Darlington and Wolsong hold the bulk of the world’s civilian tritium inventory.

What this tells you is that global fusion ambitions are tethered to a handful of aging fission plants in two countries. That means you should watch Canadian and South Korean nuclear policy, and specifically CANDU life-extension rulings, as closely as you watch any fusion breakthrough. A single refurbishment delay in Ontario ripples across the entire global fusion timeline.

Canada’s nuclear strategy in 2026 encompasses not just SMR exports and uranium supply but also the life-extension decisions for its CANDU fleet, and those refurbishment timelines at plants like Darlington and Pickering will directly govern how much tritium remains available to the global fusion sector through the 2030s.

Where the immediate commercial leverage sits

Here is where the risk story flips into an opportunity, and it is the part the market is largely ignoring while it chases fusion startups.

Every viable D-T fusion plant, regardless of which company or reactor design wins, will need tritium. That makes the investment case for the fuel supply chain agnostic to the outcome of the fusion race itself. You do not need to pick the winning reactor; you need to own the fuel choke point.

CANDU operators and tritium recovery specialists sit in a structurally advantaged position as the dominant source of recoverable tritium worldwide. With supply this tight and demand this desperate, the entities controlling production and extraction could hold serious pricing and allocation leverage as commercialisation approaches.

The recovery expertise itself is the real moat. Extracting and purifying tritium from heavy water is a highly specialised technical field with very few qualified operators globally. That scarcity of skill, not just scarcity of fuel, is what becomes increasingly valuable.

The high-value commercial segments cluster around a handful of technical capabilities:

• Recovery: extracting tritium from heavy water at CANDU facilities.
• Purification: refining raw tritium to fusion-grade specifications.
• Storage: safely holding a decaying, radioactive gas.
• Transport: moving it in small, tightly regulated containers under export controls.

There is also a potential expansion pathway worth watching. The World Nuclear Association notes that if breeding falls short, supplementary tritium could be produced by deliberately irradiating heavy water or lithium in existing fission reactors. That would create entirely new revenue streams bolted onto legacy nuclear infrastructure, and it would need specialised firms to design and run it.

What this means for your capital is a genuine pick-and-shovel thesis. Rather than gambling on a single fusion reactor design surviving to market, you can position around the tritium recovery infrastructure that every design will depend on.

These are the supply chain participants likely to hold real pricing and allocation power over the next two decades. In a gold rush, the reliable money has often been in the shovels, and tritium recovery is the shovel here.

Investors wanting to build a framework around the supply chain thesis rather than reactor selection will find our full explainer on fusion’s pick-and-shovel investment case covers which mineral and infrastructure segments hold structural pricing advantages regardless of which reactor design ultimately wins.

Positioning your portfolio for the fusion fuel reality

Strip away the hype and the picture is clear. Tritium scarcity will act as a hard throttle on how fast, and how widely, fusion can actually deploy through the 2030s.

The vulnerability is concentrated in one window. ITER’s operations are projected to draw the global stockpile down toward 5 kg or less, while commercial breeding blankets remain unproven and are unlikely to contribute meaningfully before the 2030s at the earliest. That gap between drawdown and breeding maturity is precisely where both the risk and the opportunity live.

So make the audit practical. Before you back any fusion company, current or prospective, interrogate its tritium procurement contracts and the realism of its fuel cycle assumptions. A slick plasma story with no fuel strategy is a red flag, not a rounding error.

The open question is whether governments will step in. FAS and Eurofusion analysts argue that public leadership on dedicated production capacity may be required to stop fusion projects queuing behind one another for a tiny shared stockpile.

The DOE heavy water reuse initiative at the Savannah River Site signals that governments are beginning to treat existing heavy water stocks as a strategic fusion fuel asset, an early policy move that could prefigure larger dedicated production programmes if ITER’s drawdown accelerates the supply crisis.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions.

Past performance does not guarantee future results, and these forward-looking assessments are speculative and subject to change based on technology development and policy decisions.