Nuclear Fusion’s Critical Minerals: Where the Real Scarcity Lies

The energy transition took oil out of the long-range power story. Nuclear fusion, if it hits its milestones, could do the same to lithium’s battery narrative, and then instantly replace it with a completely different lithium narrative.

Fusion has moved past the stage of pure laboratory science. Government programmes worth hundreds of millions of dollars, private funding rounds above one billion dollars, and institutional roadmaps carrying explicit 2030s pilot targets have pushed it into a new category: a credible, long-duration demand signal for a handful of specific critical minerals. Most coverage treats fusion as an energy story. The commodity angle underneath it is where the more interesting questions sit.

What follows maps the specific minerals inside the fusion supply chain, identifies where the real constraints lie, and offers a clear-eyed read on where the investable thesis holds up and where it stays speculative.

Why lithium sits at the centre of the fusion fuel cycle

Start with the fuel. The most commercially viable fusion pathway currently in development is the deuterium-tritium (D-T) reaction, and its two ingredients could not be more different in supply terms.

Deuterium is effectively abundant; it can be extracted from seawater, so it never becomes a supply chain problem. Tritium is the opposite. It cannot be mined in any meaningful quantity, because it barely exists in nature.

That leaves fusion with a manufacturing problem: every reactor has to make its own tritium as it runs.

The fusion vs fission investment distinction matters here because fission reactors run on enriched uranium and generate their own fuel management challenges, while fusion’s lithium-dependent tritium breeding cycle creates a structurally different commodity demand profile with no direct fission analogue.

The solution is lithium. Inside the blanket assembly wrapping the reactor core, neutrons streaming out of the fusion reaction strike lithium-6 isotopes, and that collision breeds tritium continuously during operation. This is not a peripheral role. Lithium is wired directly into the fuel production cycle, which makes it a prerequisite for any operating fusion plant, not an optional input.

That single mechanism is why lithium keeps appearing in fusion supply chain analysis. Take it out and there is no fuel loop.

The blanket inventory problem

The blanket is not just where fuel gets made; it is a large upfront capital commitment of lithium metal that has to sit in place before the reactor produces a single watt.

That inventory also degrades. Neutron bombardment damages the blanket material over time, which means the lithium has to be replaced across the reactor’s life. For anyone modelling demand across a future reactor fleet, that creates two separate demand events: the initial fill and the ongoing replacement cycle.

How much lithium each reactor needs is genuinely unsettled, and the spread in the published estimates is the tell.

These figures are institutional estimates and remain unverified against independent confirmation. The variance is the point. Blanket architecture is a design variable, not a fixed input, so the commodity intensity of fusion swings with engineering choices that have not been made yet. Any investment thesis built on a forecast tonnage inherits that uncertainty directly.

The enrichment bottleneck that raw reserves cannot solve

The reassuring number comes first. ITER puts global proven lithium reserves at roughly 50 million tonnes, equivalent to around 3 million tonnes of lithium-6, and its staff have argued that supply would not be a constraint for a thousand years even at a fleet of 10,000 reactors. On raw tonnage alone, fusion looks trivially well supplied.

That comfort does not survive contact with the isotopic detail.

Natural lithium is only about 7-8% lithium-6. Fusion blankets typically need that concentration pushed up to 30-90% lithium-6, which means the useful material has to be separated out through specialised isotope enrichment. The bottleneck is not in the ground. It is in the processing.

And that processing barely exists. According to ITER engineering documents, worldwide production of enriched lithium-6 is effectively zero, subject to export controls, and dependent on methods that are neither scalable nor environmentally sound. The legacy mercury-based COLEX process, for instance, is not something a modern industrial base can rebuild at scale.

Laser isotope separation has emerged as the most credible alternative to mercury-based COLEX methods, with Silex Systems reaching a commercial milestone in 2025 that suggests a Western-built enrichment pathway may be closer than programme timelines formally acknowledge.

The gap is not theoretical. The EU TRANSAT project concluded in 2022 that commercial lithium-6 is sold only in tiny quantities at very high prices, falling far short of the volume a single DEMO reactor would need.

Three constraints define the problem:

• Isotopic concentration: natural lithium carries only 7-8% lithium-6, so most mined lithium is the wrong isotope for fusion.
• Absence of Western capacity: no operating Western commercial enrichment line runs at fusion grade, per a 2026 Kronos Fusion Energy editorial that ranked lithium-6 enrichment a top programme-level procurement risk.
• Regulatory environment: enriched lithium-6 sits under export controls, adding a policy barrier on top of the industrial one.

The absolute bottleneck Samuel Ward’s 2025 analysis labels lithium enrichment an “absolute bottleneck” for commercial fusion, a description that reframes where the scarcity actually lives.

These figures remain unverified. The scale of the requirement, though, sharpens the reasoning. R.J. Pearson’s 2020 fuel analysis estimated 400 tonnes of lithium-6 for the first 100 GWe of beryllium-blanket reactors, rising to 1,500-9,000 tonnes per year at industry maturity depending on architecture. Against near-zero current output, that is a supply chain built from scratch.

For a resource investor, the read is uncomfortable but clean. The valuable position in fusion’s lithium chain is not raw lithium mining, where supply is adequate. It is lithium-6 enrichment technology: a niche industrial process with no Western scale today and outsized value to whoever builds it first. The March 2026 supply deals between Molten Salt Solutions and Type One Energy, whose engineering director flagged large-scale enrichment as one of the most pressing bottlenecks, suggest the industry has started to notice.

Beryllium and helium-3: the supporting cast and its limits

Lithium is not the only mineral in the blanket. Two others deserve attention, and they sit on completely different clocks.

Beryllium in current blanket designs

Beryllium works as a neutron multiplier. The D-T reaction alone does not throw off enough neutrons to keep tritium breeding self-sustaining, so beryllium amplifies the neutron flux hitting the lithium-6. In current blanket designs that makes it functionally indispensable, more an engineering material than a bulk commodity.

The JASON technical report illustrates the scale: a beryllium-based design for a 3 GWth reactor calls for roughly 250 tonnes of lithium and 100 tonnes of beryllium. Omitting the multiplier and switching to a lead-lithium or pure-lithium blanket removes the beryllium but demands 3 to 10 times more lithium inventory overall. That is the design hedge, and it trades one constraint for another.

Beryllium carries its own baggage:

• Performance gap: measured neutron multiplication factors have come in roughly 25% below initial calculations, forcing careful optimisation.
• Toxicity: inhaled beryllium particulates can cause chronic beryllium disease, so facilities need strict industrial hygiene and specialised handling.
• Supply concentration: the United States holds an estimated 60% of known global beryllium reserves, a geopolitical exposure for everyone else.

These estimates are unverified, but the direction is clear. Beryllium is a near-term engineering constraint that will surface in reactor procurement this decade, not a tonnage story that moves commodity prices.

Helium-3 and the lunar mining thesis

Helium-3 is the seductive alternative. On paper it is a superior fusion fuel, producing far fewer high-energy neutrons than D-T and easing the brutal materials problems that plague blanket design. The catch is that D-He3 fusion needs much higher temperatures and more advanced confinement, which parks it well behind lithium-based systems in commercial readiness.

The bigger catch is supply. Terrestrial helium-3 barely exists, with global annual production estimated at around 40,000 litres, mostly a by-product of tritium decay in weapons programmes. Its price reflects that scarcity: roughly $15,000 to $19,000 per gram. At that level it is a niche quantum and scientific material, not a fusion feedstock at any foreseeable scale.

The lunar mining thesis is real in concept. Solar wind has implanted helium-3 into lunar regolith over geological time, and companies such as Interlune have designed concepts to process 100 tonnes of regolith per hour. But concentrations sit in the tens of parts per billion, so tens of millions of tonnes of regolith must be worked to yield tens of kilograms of the isotope.

The barriers stack up:

• Terrestrial scarcity: almost nothing available on Earth, and no cheap source.
• Processing economics: the capital and energy intensity of lunar extraction is extreme, and the payback is deeply uncertain.
• Reactor immaturity: D-He3 reactors trail D-T by multiple technology transitions.

Calibrate the horizons carefully. Beryllium belongs on a fusion-adjacent risk register now. Helium-3 belongs in a speculative optionality bucket, at minimum two technology transitions away from anything an investor could act on.

For investors wanting to model the gap between terrestrial and lunar sources in detail, our full explainer on the helium-3 supply chain covers extraction economics, concentration realities in lunar regolith, and the capital intensity of processing at scale, giving the quantitative foundation that contextualises why this remains a speculative optionality position.

Mapping the investable thesis: what the timeline actually permits

The bull and bear cases here are both serious, and they point in opposite directions.

The bull case leans on demand scale. Early forecasts by Hartley projected cumulative Western fusion-driven lithium demand of 750,000 tonnes by 2030, rising toward 12 million tonnes globally. A 2025 MIT Sloan thesis modelled fusion consuming nearly 30% of global lithium supply by 2050, with demand by 2100 potentially eight times current annual production. On those numbers, fusion is a genuine new demand vector.

The bear case answers on adequacy. Durham University modelling suggests accessible lithium resources could support 2,800 years of fusion power. Max-Planck Institute estimates put a mature fleet’s appetite at about 806 tonnes of lithium-6 a year, requiring roughly 10,050 tonnes of natural lithium annually, a rounding error against the 9.9 million tonne reserve base.

Both cases can be true at once, because they measure different things. The bull figures describe demand; the bear figures describe supply adequacy. Reconcile them and the credible near-term thesis is not raw lithium spot exposure. It is the enrichment technology layer, where the gap is real, current, and unambiguous.

The programme-level signal The UKAEA’s £220 million LIBRTI programme is building a testbed to validate lithium tritium-breeding blankets at commercial scale. In July 2026, Commonwealth Fusion Systems, fresh off a $1 billion raise, became the first international company to join, signalling that blanket and enrichment validation is now a funded priority, not a research curiosity.

Commonwealth Fusion Systems joining LIBRTI in July 2026, reported by the American Nuclear Society, confirmed that breeder blanket validation has crossed from funded research into active commercial procurement, with the programme specifically targeting prototype testing infrastructure that no fusion developer can bypass on the path to a pilot plant.

The timeline keeps everyone honest. First pilots remain a 2030s-2040s proposition, and the US DOE’s 2023-2024 roadmap treats breeder blanket validation as a gating milestone still ahead. Fusion is not a current price catalyst for lithium spot markets, but the capital flowing into blanket and supply infrastructure tells you where the smart early positioning sits.

Past performance does not guarantee future results. Financial projections are subject to market conditions and various risk factors, and the estimates cited above remain unverified and speculative.

Where the fusion supply chain thesis is credible and where it is not

Strip away the narrative and the verdict differs sharply by commodity.

• Lithium: the raw supply thesis is oversold for the near term, but the lithium-6 enrichment thesis is underappreciated and carries credible commercial logic tied to 2030s programme milestones. Watch this layer.
• Beryllium: a real engineering constraint in current reactor designs, with genuine supply concentration risk, but not a tonnage play that drives price formation in conventional mining analysis. Hold as a risk-register item, not a growth thesis.
• Helium-3: legitimate long-horizon optionality for space-resource and deep-technology investors, but multiple technology transitions away from being an investable fusion fuel. Treat as narrative, for now.

The 2026 milestones, LIBRTI, the CFS partnership, and the Molten Salt Solutions supply deals, are leading indicators that the enrichment infrastructure gap has entered commercial awareness. That awareness is early. The first commercial pilots in the 2030s-2040s mark the earliest point fusion demand could realistically appear in commodity balance sheets.

Set that against the Max-Planck steady-state figure of roughly 10,050 tonnes of natural lithium a year for a mature fleet, and the raw material impact looks modest against global lithium markets. The scarcity does not live in the rock.

So the correctly framed thesis is this: fusion critical minerals is a precision trade, not a broad commodity call. The investors best positioned are the ones who can identify and access the enrichment technology layer before it is priced, not those buying raw lithium or beryllium on a fusion headline.

Critical minerals investment positioning in 2026 has increasingly bifurcated between broad commodity exposure and technology-layer plays, a split that mirrors exactly the distinction the fusion supply chain thesis forces between owning raw lithium and owning enrichment process access.

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.