Hidden Depths in the Semiconductor Supply Web

A Surface Reading and Its Limits

American semiconductor manufacturers report spending close to 27 billion US dollars annually on the equipment and materials that keep fabrication lines running. Headline figures from the Semiconductor Industry Association show 95 per cent of that outlay flowing to suppliers headquartered in the United States, Europe, and Japan - a picture that suggests diversification and resilience. Yet when analysts shift the lens from corporate domicile to country of manufacture, and then drill further into specific material categories, the portrait changes sharply.

China accounts for roughly 5 per cent of equipment and materials when measured by where those goods are physically made, according to SIA data. That share sounds modest until the analysis moves to rare earths and specialty materials: for certain elements essential to chip production, Chinese control exceeds 90 per cent. The gap between aggregate spending patterns and pinch-point dependencies is the story that matters, and it is one the industry has begun mapping with greater precision.

Headquarters Versus Factory Floor

The 95 per cent figure counts suppliers by the location of their global headquarters. A Japanese equipment maker, a German materials firm, and a US tooling company all fall into the "allied" column, even when their manufacturing footprint spans multiple continents. That accounting convention makes sense for trade-policy discussions and alliance frameworks, but it obscures the physical geography of production.

When the same spending is re-tabulated by country of manufacture - where lathes turn, chemicals are synthesised, and substrates are polished - China's share remains single-digit but its role in specific categories grows visible. The SIA analysis, released in October 2026, represents one of the first systematic efforts by a US industry body to publish both views side by side, a response to questions raised in congressional hearings and supply-chain task forces over the past two years.

At Opentechwire, we have tracked similar mapping exercises across memory, logic, and compound-semiconductor segments. Each reveals the same pattern: aggregate figures flatten the peaks and troughs that define vulnerability.

The Rare-Earth Choke Point

Rare-earth elements - scandium, yttrium, and the fifteen lanthanides - are not uniformly rare in crustal abundance, but their extraction, separation, and refining are capital-intensive and environmentally complex. Over the past two decades, production has concentrated in a handful of provinces, and China now controls more than 90 per cent of global processing capacity for several elements used in semiconductor deposition, etching, and doping.

Europium and terbium oxides, for example, are critical to certain photoresist chemistries and phosphor layers. Neodymium and dysprosium underpin the high-strength magnets in vacuum-deposition equipment. Substitutes exist for some applications, but they typically entail performance trade-offs or require multi-year qualification cycles that fab managers are reluctant to initiate unless compelled.

The concentration is not accidental. It reflects decades of policy continuity, vertical integration between mining and refining, and willingness to absorb environmental costs that other jurisdictions have externalised. Western rare-earth projects - Mountain Pass in California, Nolans in Australia, several ventures in Canada - have come online or restarted in the past five years, but their combined output remains a fraction of the volumes processed in Jiangxi and Inner Mongolia.

Policy Implications and Industry Response

Export controls introduced by the United States in October 2022 and expanded in 2023 and 2024 focus on advanced logic nodes, high-bandwidth memory, and the lithography and deposition tools that enable them. Those measures target the frontier of chipmaking; they do not directly address the material substrate. China has responded with its own controls: in July 2023, export licences were imposed on gallium and germanium, and in December 2024, restrictions were extended to several rare-earth processing technologies.

The result is a layered chess game in which each side holds asymmetric pieces. The US and its partners dominate tool design and advanced packaging; China holds sway over upstream materials and mature-node capacity. Neither position confers outright advantage, but both create friction.

American semiconductor firms have begun dual-sourcing programmes for critical materials, sometimes at significant cost. One executive at a leading foundry, speaking on background at an industry conference in Arizona in September 2026, estimated that qualifying a second rare-earth supplier added eight to twelve months to a process-development timeline and required duplicate inventory to hedge against batch variation. "We do it because the risk committee insists," the executive said, "but it is not free."

Beyond Rare Earths: The Broader Material Map

Rare earths are the most discussed category, but they are not the only one where concentration creates exposure. High-purity silicon wafers, specialty gases such as electronic-grade hydrogen fluoride and nitrogen trifluoride, and polishing slurries all exhibit regional clustering. Japan dominates wafer production; Taiwan and South Korea are major gas suppliers; slurry formulations are split between US and Japanese chemistry houses, with some precursor materials sourced from China.

The SIA report breaks spending into eleven categories, from photomasks and substrates to metrology tools and clean-room consumables. In seven of those categories, no single country accounts for more than 30 per cent of supply by manufacture. In the remaining four - rare earths, certain gases, graphite components, and bulk chemicals - one or two countries exceed 50 per cent.

Industry analysts have begun using the term "material sovereignty" to describe efforts by governments in the US, EU, Japan, and South Korea to build or rebuild domestic capacity in these concentrated segments. The EU's Critical Raw Materials Act, signed in March 2024, sets targets for domestic extraction and processing. South Korea's Materials, Parts, and Equipment 2.0 initiative, launched in 2025, allocates subsidies to firms that localise supply. The US CHIPS and Science Act includes a smaller materials-focused appropriation, though most funding has flowed to fab construction.

The Cost of Resilience

Diversification is not synonymous with efficiency. Rare-earth refining, for instance, generates tailings laced with thorium and other low-level radioactive isotopes; managing those tailings to Western environmental standards raises unit costs by an estimated 30 to 50 per cent compared with current Chinese operations. Gas production requires continuous demand to justify capital investment in distillation columns and purification trains; splitting orders across multiple suppliers can push plants below efficient scale.

Chip executives and policymakers now speak openly about a "resilience premium" - the incremental cost of supply-chain redundancy and geographic diversification. Estimates vary, but industry surveys suggest the premium adds 3 to 7 per cent to the bill of materials for a leading-edge wafer. That burden is unevenly distributed: fabless designers can pass it downstream, while integrated manufacturers absorb it directly.

At the same time, concentration risk carries its own price. A month-long export delay, a quality excursion, or a regulatory shift in a dominant supplier country can idle billion-dollar fab lines. The 2021 shortage of automotive-grade chips, though driven more by demand whiplash than material constraint, demonstrated how quickly bottlenecks propagate.

Forward View

The SIA analysis arrives as the US and Chinese governments enter another round of semiconductor negotiations, this time focused on mature nodes and the legacy equipment that produces them. Material dependencies will likely feature more prominently than in earlier talks, if only because both sides now have published data to reference.

Industry insiders expect further mapping efforts. The International Semiconductor Manufacturing Initiative, a consortium of fabs and equipment makers, is developing a material-traceability protocol that would track components from ore to wafer. The protocol remains voluntary, but participation has grown as customers - particularly in aerospace and defence - demand visibility into second- and third-tier suppliers.

Whether these initiatives reduce concentration or simply render it more transparent is an open question. Building a rare-earth refinery in Wyoming or a high-purity gas plant in Saxony does not guarantee those facilities will run at competitive cost or quality. It does, however, establish optionality, and in a world where supply chains have become instruments of statecraft, optionality commands a premium all its own.

The semiconductor industry has long prided itself on globalisation - on the ability to source the best technology and the lowest cost from wherever they emerge. That era has not ended, but it is bending under the weight of strategic calculation. The SIA report is one more data point in that shift, a reminder that the most critical dependencies often lie not in the headline figures but in the fine print of the periodic table.