Gallium and Germanium: Why the Supply Gap May Last the Decade

Two of the smallest metals markets in global commodities, measured by the tonnage that changes hands each year, now sit at the centre of a strategic contest that far outweighs their physical footprint. Gallium and germanium barely register against copper or aluminium by volume, yet they are woven through AI data-centre optics, military infrared lenses, and the fibreoptic backbone of the internet.

That mismatch between apparent obscurity and actual strategic weight is the whole story. When a metal you have never heard of underpins the power electronics inside an AI server, its supply constraints stop being a niche concern.

China’s export controls have made this concrete. Active since August 2023 and tightened into a near-total ban on US-bound shipments in December 2024, the restrictions have pushed Western gallium prices to roughly 10 times 2023 levels. The demand story, meanwhile, is accelerating on its own timeline, and the two forces are now converging at once.

China’s broader export controls framework, established across rare earths and now applied to gallium and germanium, follows a consistent strategic template: licencing requirements arrive first, then quantity restrictions, then country-specific bans, a sequencing that allows Beijing to calibrate pressure without triggering WTO disputes immediately.

What follows here is a structured read on gallium and germanium demand for investors trying to decide whether supply plays represent a durable, multi-year opportunity or a control-driven price spike with a short shelf life. The aim is to separate the demand drivers that are structurally embedded from the substitution risks that are real versus overstated.

What is actually driving gallium and germanium demand, and how fast is it growing?

Start with the baseline, then watch the case build layer by layer. According to preliminary S&P Global estimates, global gallium demand sits at roughly 1,000 metric tons in 2025 and is forecast to grow at around 12% annually through 2030. Germanium is the slower runner, at approximately 343 tons in 2025 growing near 3.3% per year over the same window.

Those growth rates matter less than the reason they exist. Three distinct demand vectors are pulling on these metals simultaneously, and each one relies on a specific compound that is difficult to engineer around.

• AI infrastructure: gallium arsenide (GaAs) and gallium nitride (GaN) for high-speed optical interconnects, RF components, and the high-efficiency power stages inside AI servers
• Fibreoptic networks: germanium-doped fibre preforms that carry long-haul optical transmission across the internet backbone
• Infrared imaging: germanium lenses prized for their transparency across the thermal imaging range, central to defence optics and thermal cameras

Because these are additive rather than competing forces, the cumulative demand picture is stronger than any single application suggests.

The compound-semiconductor mechanics behind each demand vector

The AI layer is where the forecasts get their momentum. Yole Group projects indium phosphide (InP) substrates growing at more than 18% CAGR through 2031, driven by AI data centres and co-packaged optics. GaAs substrates are on a similar path, with one projection showing growth from US$1.33 billion in 2025 to US$1.48 billion in 2026 (a figure flagged as unverified in the underlying research).

Here is the mechanical detail investors should hold onto: because AI optics draw on both GaAs and InP, a shift toward InP does not eliminate gallium at all. It can instead strain indium supplies, moving the bottleneck rather than removing it.

The most aggressive demand estimate comes with a health warning.

Foreign Policy Analytics estimates that AI-specific needs could raise gallium demand by approximately 85% and germanium demand by approximately 37% by 2033 versus a non-AI baseline. This figure is flagged as unverified in the underlying research and should be read as a directional signal, not a settled forecast.

The reality check sits in the tonnage. USGS-derived consumption data suggests roughly 73% of gallium flows into integrated circuits and around 26% into optoelectronics, meaning AI mostly amplifies demand within sectors that already exist. Physical gallium consumed by AI data centres today is estimated at just 0.2 tons against roughly 900 tons of global primary output.

That gap is the point. The 12% gallium CAGR is a directional signal confirming the structural tailwinds are genuine, but the tonnage reality tells you the strategically sensitive tier is the high-purity refined segment where AI demand concentrates, not the headline volume.

Why the substitution playbook is harder than it looks

The bear case you will hear most often is substitution. If gallium and germanium are expensive and constrained, the argument goes, engineers will simply design them out. It is worth taking that thesis seriously, then watching where it holds and where it breaks.

It holds in the commercial and consumer tiers. InP is already replacing GaAs in specific-wavelength optics. In mass-market infrared, zinc selenide, zinc sulphide, silicon, and chalcogenide glass are gaining ground. Belgium’s Umicore is substituting germanium in optics using its proprietary GASIR chalcogenide glass and Tessella wafer-moulding, targeting a roughly threefold reduction in germanium intensity for certain optics and fibre-optic applications (a target flagged as unverified).

Technical analysis of chalcogenide glass as an alternative to germanium confirms that while the material shares a similar IR transmission profile, performance trade-offs in high-resolution defence optics remain significant enough to limit adoption to commercial and mass-market tiers rather than mission-critical systems.

The evidence that substitution is real, not theoretical, comes from the field. One anonymous optics company reported cutting germanium usage by roughly 50% over approximately 18 months by switching to substitute materials and sourcing from Western suppliers.

Then the thesis hits a wall. In defence and high-performance segments, substitution stalls, and the industry voice on this is blunt.

“There is no direct substitute for germanium” in many high-end infrared applications without significant redesign or performance degradation, according to Jessica DeGroote Nelson, Senior Vice President of Precision Optics at Edmund Optics.

The barriers stack up in a predictable sequence, and each one lengthens the substitution timeline:

1. Performance gap: alternatives typically underperform the incumbent material in demanding applications
2. Redesign cost: switching materials forces expensive product re-engineering
3. Qualification time: defence and high-reliability systems require lengthy re-certification
4. Supply access for alternatives: scaling substitutes creates its own sourcing constraints

The clearest read on market tightness is not a price chart. It is behaviour: buyers are reportedly purchasing germanium before their product designs are even finalised, simply to secure physical access. That tells you the binding constraint is availability, not cost tolerance, and availability constraints do not resolve when a substitute appears in one segment while defence demand stays captive.

The germanium supply crisis that preceded China’s formal export controls has its own internal logic, one shaped by the metal’s byproduct status, the concentration of refining assets, and the structural gap between mine output and purified feed available to Western manufacturers.

How the supply picture actually looks for investors assessing ex-China plays

The concentration is close to absolute. China supplied approximately 98.9% of primary gallium and about 68.6% of germanium in 2025, according to figures flagged as unverified in the research. Against that dominance, S&P Global expects ex-China gallium supply capacity to reach only around 20 tons by end-2026, leaving a near-term gap of roughly 678 tons.

That is the starting point for any supply thesis: the shortfall is stark, and the near-term relief is thin.

The pipeline of ex-China capacity

A build-out is underway, though it should not be oversold. At least eight announced gallium projects could lift ex-China supply to approximately 386 tons by 2030, from a base of around 5 tons currently. On the germanium side, new refining projects could expand ex-China capacity to roughly 126 tons by 2030 (all figures flagged as unverified).

Ex-China germanium supply projects increasingly rely on tailings reprocessing and coal byproduct recovery rather than primary mining, a structural distinction that changes both the capex profile and the permitting risk compared with conventional mine development.

The named projects give the pipeline shape:

• Wagerup, Australia (gallium)
• Clarksville, United States (gallium)
• New germanium refining capacity in Canada
• Germanium refining in South Korea

Allied governments are backing the effort. In April, Australia and the United States pledged over US$3.5 billion to support critical minerals projects, explicitly naming gallium and germanium, a signal of strategic commitment rather than pure market economics (figure flagged as unverified).

Note: the figures in this table are flagged as unverified in the underlying research and should be treated as directional.

Do the arithmetic and the investment window becomes visible. A 386-ton ex-China gallium target by 2030 against a current gap near 678 tons means even the most optimistic build-out leaves a structural shortfall through most of this decade. That shortfall, not the pipeline, is where the opportunity sits.

The role of recycling in bridging the near-term gap

Recycling is the fastest lever that does not require new mining. Umicore runs the most advanced ex-China model, reporting that more than 50% of its germanium feed comes from recycling scrap recovered from fibre-optics, solar cells, LEDs, and IR optics, with the EU selecting its germanium projects for strategic support in March 2025 (both flagged as unverified).

The headroom is substantial: global end-of-life germanium recycling is estimated at only around 20%. In the US defence sector, Lattice Materials supplies germanium crystals for fighter-jet and tank display systems and leans heavily on recycling, while cautioning that it does not anticipate near-term supply relief. For investors, that combination of low current recycling rates and explicit strategic backing marks where incremental supply is most likely to appear before new mines come online.

Where the demand bull case breaks down and what investors should watch

The risks here are not a footnote to weigh lightly. Each one deserves to be measured against the structural demand case already built, because that is how a calibrated position gets sized.

Three risk categories separate cleanly:

1. Policy and geopolitical: China’s US-specific export ban is only suspended, with the reprieve expiring 27 November 2026, and a sudden relaxation could trigger a sharp Western price correction while stalling ex-China investment economics.
2. Demand composition: physical gallium in AI data centres today is around 0.02% of global primary output, so a shift toward silicon-heavy architectures or an AI deployment plateau would erode the AI narrative faster than the headline CAGR implies.
3. Substitution scaling: if efficiency gains in consumer and mass-market segments accelerate, baseline volume demand could soften even while defence stays captive.

The policy risk carries the most immediate weight. USGS modelling suggests a total disruption of China’s net germanium exports (roughly 35 tons out of about 210 tons of world output) would lift prices only around 26% and cut non-Chinese supply by approximately 5.5%, which frames how policy shifts translate into price outcomes rather than catastrophe. China’s speculated purchase of about 100 tons of germanium in July 2024 shows the same lever working in reverse, driving domestic prices to record highs.

Price is the reference point for how far this has already run.

Reuters reported on 18 September 2026 that gallium prices were 9 to 10 times 2023 levels, with Rotterdam near US$2,101.6/kg in March 2026 against roughly US$247/kg inside China. These figures are flagged as unverified in the underlying research.

The bull case survives the honest weighing. Defence, fibre optics, and 5G/6G infrastructure represent embedded demand that analysts frame as moderators rather than reversals, and the supply gap persists on any realistic scenario. But the 27 November 2026 expiry is the single most time-sensitive variable in the near-term thesis, and any investor in ex-China supply plays should know how their position behaves under a relaxation scenario before that date arrives.

The structural case, weighed honestly for the investor

Pull the four layers together and a coherent read emerges. Demand is embedded in defence, fibre optics, and power electronics rather than resting on AI hype. Substitution is real but partial and slow, blocked precisely in the segments with the least price elasticity. And the supply math, a near-term gallium shortfall around 678 tons against an ex-China germanium refining target of roughly 126 tons by 2030, points to a gap that persists through most of this decade even on optimistic pipeline assumptions.

The IEA’s 2026 Outlook classifies both metals as highly exposed to supply vulnerabilities, an institutional validation of the structural thesis (flagged as unverified).

The IEA 2026 Outlook flags both gallium and germanium as highly exposed to supply vulnerabilities, citing high production concentration and limited substitutability. This classification is flagged as unverified in the underlying research.

What remains genuinely unknown matters just as much: the pace of ex-China project commissioning, China’s export posture after November 2026, and whether AI hardware sustains or trims its compound-semiconductor intensity. The US$3.5 billion allied commitment signals policy-backed supply is coming, but government backing does not guarantee commercial timing.

The fundamentals justify treating these as long-duration theses. Which part of the value chain, mining, refining, recycling, or component manufacturing, captures the most defensible margin is the question the demand analysis alone cannot answer.

Investors wanting to map how the US$3.5 billion allied commitment fits within the broader Western policy architecture will find our full explainer on critical minerals supply chain strategy, which covers implementation mechanisms, allied coordination frameworks, and the commercial conditions attached to government-backed project support.

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. Financial projections are subject to market conditions and various risk factors, and forward-looking statements are speculative and subject to change based on market developments.