Geopolitical Shifts in Advanced Compute Supply Chains

Executive Summary

The global semiconductor supply chain is no longer a commercial concern. It is a national security architecture. The materials required to manufacture advanced compute hardware — gallium, germanium, high-purity polysilicon, cobalt, and neon gas — are concentrated in a handful of geographies, many of which are either adversarial to Western interests or politically unstable. This concentration creates chokepoints that can be weaponized.

In July 2023, China imposed export restrictions on gallium and germanium, two materials essential for advanced chip manufacturing. The restrictions were not embargoes — they were licensing requirements, adding bureaucratic friction to every transaction. The effect was immediate: spot prices for both materials increased by 27% within 72 hours. Lead times for gallium-based compound semiconductors extended from 8 weeks to 22 weeks.

This was a signal, not a weapon. The weapon is the underlying reality: China controls 98% of global gallium production and 60% of germanium production. No amount of stockpiling or recycling can alter this dependency in the short term. The question is not whether these materials will be further restricted, but under what geopolitical conditions.

I. The Gallium Bottleneck

Gallium is not mined directly. It is extracted as a byproduct of aluminum smelting — specifically from bauxite ore processing. China dominates global aluminum production, and consequently, gallium extraction. There is no alternative supply chain that can be activated quickly. Building gallium extraction capacity from scratch requires investment in aluminum smelting infrastructure, which takes 4-7 years to bring online.

The compound semiconductors that depend on gallium — particularly Gallium Nitride (GaN) and Gallium Arsenide (GaAs) — are not optional components. GaN is the foundation of 5G base station power amplifiers. Without it, the rollout of advanced telecommunications infrastructure stops. GaAs is essential for satellite communications, radar systems, and certain categories of photovoltaic cells used in military applications.

Japan and South Korea have begun emergency programs to develop gallium recycling from electronic waste. The recovery rates are promising (approximately 85%) but the total volume recoverable from e-waste represents less than 12% of annual global demand. The arithmetic is unforgiving: recycling is a mitigation strategy, not a solution.

Canada possesses significant bauxite reserves and existing aluminum smelting capacity. Rio Tinto's Kitimat facility in British Columbia could theoretically be retrofitted for gallium extraction within 18-24 months. As of this writing, no public announcement has been made, but procurement activity consistent with such a retrofit has been detected in supply chain databases monitored by ANN.

II. The Neon Crisis Nobody Remembers

Before Russia's invasion of Ukraine in 2022, approximately 54% of the world's semiconductor-grade neon gas was produced in two Ukrainian cities: Odesa and Mariupol. Neon is essential for the excimer lasers used in extreme ultraviolet (EUV) lithography — the process by which the most advanced chips in the world are manufactured. Without semiconductor-grade neon, ASML's lithography machines cannot function.

The war disrupted these supplies catastrophically. Neon prices increased by 900% in the weeks following the invasion. TSMC, Samsung, and Intel activated strategic reserves. The crisis was contained — barely — because these reserves existed and because alternative neon purification facilities in South Korea and the United States were accelerated into production.

But the structural lesson was clear: the semiconductor supply chain's dependence on single-source materials produced in conflict zones is not a theoretical vulnerability. It is a demonstrated one. And yet, as of early 2026, the diversification of neon sources remains incomplete. South Korean facilities now supply approximately 30% of global demand, up from 8% pre-war. The remainder still flows through Russian intermediaries.

III. The Xinjiang Polysilicon Problem

Approximately 45% of the world's polysilicon — the foundational material for both semiconductor wafers and photovoltaic cells — is produced in China's Xinjiang Uyghur Autonomous Region. The concentration in Xinjiang is driven by cheap coal-fired electricity, which is the primary cost input for polysilicon production.

The Uyghur Forced Labor Prevention Act (UFLPA), signed into law in the United States in December 2021, creates a rebuttable presumption that any goods produced in Xinjiang involve forced labor. This has created a paradox: the United States is simultaneously attempting to build a domestic clean energy industry (which requires massive quantities of solar panels, which require polysilicon) while blocking imports from the region that produces nearly half the global supply.

The result has been price inflation, supply uncertainty, and a quiet circumvention economy. Polysilicon produced in Xinjiang is shipped to facilities in Southeast Asia, where it is "processed" and re-exported with certificates of origin that obscure its provenance. Everyone in the supply chain knows this is happening. Enforcement is selective and politically motivated.

For semiconductor-grade polysilicon specifically, the stakes are even higher. The purity requirements for chip manufacturing (eleven-nines purity: 99.999999999%) limit the number of facilities worldwide capable of production to fewer than a dozen. Of these, four are in China.

IV. The Reshoring Illusion

The CHIPS and Science Act allocated $52.7 billion to incentivize domestic semiconductor manufacturing in the United States. TSMC is building facilities in Arizona. Intel is expanding in Ohio. Samsung is constructing a $17 billion fab in Texas. These are real investments with real timelines.

But fabrication is only one link in the chain. The raw materials — the gallium, germanium, neon, cobalt, and polysilicon — still originate from the same concentrated sources. Building a chip fab in Arizona does not reduce dependence on gallium from China. It moves the vulnerability one layer upstream, where it is less visible but no less acute.

True supply chain security requires vertical integration from mine to fab. No nation has achieved this. The closest candidate is China, which is precisely why it holds such extraordinary leverage.

V. Strategic Implications

The compute supply chain is not a market. It is a strategic asset that happens to be organized as a market. The distinction matters because market logic assumes fungibility and price-driven substitution. Neither applies when a single nation controls 98% of a critical input.

For institutions exposed to advanced compute — which in 2026 means every institution — the risk model must be updated. Semiconductor availability is no longer a procurement concern. It is a geopolitical variable with binary outcomes: either the materials flow, or they don't. Hedging strategies must account for both scenarios.

The next 36 months will be decisive. If alternative sources of gallium, germanium, and polysilicon are not brought online by 2028, the window for supply chain diversification will close. The geological and industrial realities are fixed. The political realities are accelerating.