Gallium, Germanium and the Retaliation Toolkit

If rare earths are the tendons of the compute economy, gallium and germanium are its nervous system's specialty wiring. Neither is a household metal. Both are byproducts, won not from dedicated mines but as trace outputs of larger industrial processes: gallium mostly from the refining of bauxite into alumina, germanium from zinc smelting and coal fly ash. This byproduct nature is the key to their strategic character. You cannot simply open a gallium mine. Supply is chained to the economics of aluminium and zinc, which means capacity is slow to add and concentrated wherever those primary industries already run at scale.

Gallium's importance to compute is rising fast. Gallium arsenide and gallium nitride underpin the power electronics, radio-frequency components and high-efficiency amplifiers that a modern data centre and its telecom backbone depend on. Gallium nitride in particular is becoming central to power conversion, the unglamorous business of turning grid electricity into the clean, dense, efficient current that racks of accelerators consume. As AI clusters push power draw into the tens and hundreds of megawatts, the efficiency of that conversion stops being a footnote and becomes a design constraint. Germanium, meanwhile, is indispensable to fibre-optic cable, infrared optics and certain high-speed transistors, which places it squarely inside the connective tissue that links data centres to one another and to users.

In July 2023 China moved these two metals from the realm of quiet commerce into the centre of the compute contest. Citing national security, Beijing imposed export licensing requirements on gallium and germanium products, a direct answer to tightening American and allied controls on advanced chipmaking. The choice was deliberate and instructive. China accounts for the dominant share of global gallium and germanium refining, so a licensing regime granted Beijing a valve it could open and close at will. In late 2024 it went further, announcing outright bans on exports of these and related materials to the United States. The retaliation toolkit was no longer theoretical.

The genius, if that is the word, of targeting byproduct metals is that retaliation is cheap for the imposer and expensive for the target. Because gallium and germanium are small markets by dollar value, restricting them costs China little in lost revenue, yet the disruption ripples through high-value downstream products worth vastly more. It is asymmetric coercion: a pinprick that causes a systemic ache. And because Western capacity to refine these byproducts atrophied over decades of relying on cheap Chinese supply, there was no idle plant waiting to absorb the shock.

Rebuilding is possible but awkward. Gallium can be recovered from the alumina refining that already happens outside China, and germanium can be pulled from zinc operations and recycled from optical scrap, but the recovery circuits were dismantled or never built because the margins never justified them against subsidised imports. This is the recurring pattern of the entire strategic-minerals story: a rival state lets the West grow comfortable and dependent during years of low prices, then converts that dependence into leverage the moment the technology contest sharpens. The cheap supply was, in retrospect, a slow-acting instrument of policy.

Faraday's read is that gallium and germanium are the clearest proof of the course thesis in miniature. Export controls do not merely slow an adversary; they reshape where things get built. When Beijing signalled it would ration these metals, buyers did three things at once. They scrambled for stockpiles, they qualified alternative and recycled sources, and they began paying premiums that suddenly made previously uneconomic Western recovery projects pencil out. In other words, the control itself summoned new capacity into being, but only after a painful lag and at a higher permanent cost. The map redraws, but the redrawing is neither instant nor free.

For investors and infrastructure planners, the practical lessons are concrete. First, map your exposure to byproduct metals honestly, because the risk hides several tiers down the bill of materials, inside power modules and optical components rather than on the marquee chip. Second, assume that any material with a single dominant refiner is a candidate for weaponisation, and price that tail risk rather than ignoring it. Third, recognise that stockpiles buy time but not immunity; the durable hedge is diversified refining and a functioning recycling loop, both of which take years to stand up. The states and firms that internalise these lessons early will find themselves a step ahead when the next valve turns. Those that treated cheap supply as a permanent feature of the world will discover, at the worst moment, that it was a lever all along.