The short version
Our October 10 guide traced how ASML's parts-price hike flows from equipment vendors into foundry costs. This week's news goes one layer deeper — into the raw materials that make the equipment vendors' parts possible. On October 9, 2026, the Semiconductor Industry Association (SIA) published a supply-chain assessment that systematically measured what American chipmakers actually depend on — not by where their suppliers are headquartered, but by where things are actually made. The gap between those two measurements is the story. When counted by supplier headquarters, 95% of the roughly $27 billion that US chipmakers spend annually on equipment and materials comes from companies based in the US, Europe, and Japan. Counted by production location, the SIA estimates that figure falls to 68%. When counted by production site, China accounts for about 5% of the Tier 1 equipment and materials supplied directly to US chipmakers (1% by headquarters). And that is only the direct-purchase picture. The report's sharper finding is upstream: nearly one-third of the roughly 100 upstream categories the SIA analyzed rate as high-risk — and more than three-quarters of the Tier 1 categories it analyzed depend on at least one high-risk upstream input. The chip industry's China problem, it turns out, is not in the invoice. It is in the mine, the refinery, and the precursor plant.
What the report actually says — and what it doesn't
Tier the evidence before drawing conclusions. **Industry-association research (SIA, "Understanding Dependencies in the U.S. Semiconductor Supply Chain," published October 9, 2026):** Built with procurement and strategy leaders from chipmakers, fabless firms, packaging companies, equipment makers, and materials suppliers. The analysis covered 200 Tier 1 inputs and nearly 100 upstream categories (a list the report itself calls non-exhaustive). Each input was scored on five factors: supplier concentration, ownership, country concentration, exposure to high-risk countries and sub-tier dependency. The report's own figures — all of them SIA estimates, not independent measurements — include the 95%-by-headquarters versus 68%-by-production-location gap above; China at about 95% of gallium supply and about 85% of tungsten, and about 95% for the rare earths neodymium/praseodymium and terbium; Japan at about 75% of ABF substrate production and about 70% of photomask blanks (the blank plates masks are made from, used across DUV and EUV lithography); and the warning that building capacity for some materials "can take a decade or more," plus three to four years to qualify new suppliers and materials for chip production. What is *not* in the reporting: any US government endorsement of the findings, any linkage between these upstream risks and specific component part numbers or prices, and any prediction that a particular shortage *will* occur. The SIA's "decade plus three to four years" is a warning about lead times, not a forecast of an imminent cutoff. Keep that distinction — the analysis below is built on the report's structure, not on anyone's claim that your next shipment is at risk.
The four layers of upstream risk
The report's most useful contribution for buyers is structural: it shows risk concentrating at different depths of the supply chain. Think of it as four layers, from the ground up. **Layer 1 — Raw minerals.** The deepest and most concentrated. On the SIA's estimates, China holds roughly 95% of global gallium supply and ~85% of tungsten, and about 95% of supply of the rare earths neodymium/praseodymium and terbium. These are the inputs where alternative supply genuinely takes a decade to build — the SIA's starkest warning. **Layer 2 — Engineered precursors and chemicals.** One step downstream, and where chipmakers actually buy. Precursor chemicals for deposition, photoresists, ultra-high-purity gases, sputtering targets. The geography here is Japan, Taiwan, and South Korea for many of these — but those producers themselves depend on Layer 1 minerals. A US-headquartered materials supplier making photoresist in Japan from Chinese-sourced inputs is exactly the headquarters-vs-origin gap the report exposes. **Layer 3 — Equipment and its critical sub-components.** Lithography, etch, and deposition tools — and, deeper still, the custom components inside them: turbo pumps, dry pumps, vacuum robots, transfer valves — components the report's sub-tier analysis flags as frequently single-sourced. This is the layer our ASML guide covered from the pricing side; the SIA report covers it from the supply side. A tool you cannot get parts for and a tool whose parts have no second source are the same risk wearing different clothes. **Layer 4 — Packaging and interconnect.** ABF substrates (~75% Japan by production location, per the SIA report), photomask blanks (~70% Japan — the blank plates, not finished masks). Closer to the finished chip, less about China directly — but the same concentration logic applies, and the same single-supplier fragility. The buyer's insight: most procurement teams stop their risk mapping at Layer 3 or 4 — the supplier they have a contract with. The SIA report says the real exposure sits at Layers 1 and 2, where your supplier's supplier lives. You cannot fix what you cannot see, which is why the checklist below starts with visibility.
The November 10 catalyst — read carefully
One set of dates in the reporting deserves a calendar entry, handled precisely — because there are three different measures here, with three different dates, and they are often merged into one. Beijing imposed export-license requirements on gallium, germanium, and graphite in 2023, escalated to bans in 2024, and added seven medium and heavy rare earths in April 2025. Separately, China suspended its 2024 US-specific ban on gallium, germanium and antimony exports in November 2025 — a suspension reported (Mining Technology, November 10, 2025) to run until November 27, 2026, with the underlying licensing requirements still in place. Separately again, the broader set of controls announced last October was suspended until November 10, 2026. And the Sasakawa Peace Foundation notes the US describes the truce as running to January 10, 2027, with the two sides reading implementation differently — three measures, three clocks, no single expiry date. What happens on November 10 is *not* predictable from the reporting: the broader suspension could be extended, narrowed, or allowed to lapse. What is predictable is that it is a decision point both governments are watching — the White House said after the September 24 summit, as quoted by the Sasakawa Peace Foundation, that both sides 'continue to work on U.S. concerns regarding supply chain shortages related to rare earths and other critical minerals, with the goal of ensuring shipment levels return to appropriate levels.' Treat November 10 as a watch date for your procurement calendar, not as a forecast of disruption. The buyers who get hurt by policy dates are the ones who first hear about them afterward.
A buyer's upstream-exposure checklist
This is the actionable core — five things a component buyer can do with this report this quarter. **1. Ask for origin-of-manufacture disclosure, not headquarters.** The report's central methodological point is directly usable: when you qualify a supplier, ask where the critical inputs are *produced*, not where the supplier is *based*. A one-page "origin exposure statement" for your top five suppliers' key inputs is a reasonable ask — and the suppliers who cannot produce one are telling you something. **2. Map your BOM against the four layers.** You do not need the SIA's full input model. Take your top 20 components by spend and ask, for each: which layer holds its scarcest input? Anything touching Layers 1–2 (gallium in compound semis, tungsten-derived precursors, rare-earth magnets in equipment-adjacent systems) goes on the watch list. **3. Write price pass-through terms into contracts now.** If upstream costs are the new volatility source — as the ASML hike and this report jointly suggest — then fixed-price multi-year agreements without adjustment clauses transfer that volatility to your supplier, who will either refuse or price it in. Negotiate indexed or reviewable pricing for the exposed categories instead of pretending the cost base is stable. **4. Qualify second sources for the concentrated items first.** The report's concentration figures are a prioritization tool: start alternate qualification where concentration is highest and substitution hardest (single-supplier sub-components, Japan-concentrated substrates), not where it is easiest. **5. Note what you cannot stockpile.** The report is candid that inventory is not a universal answer: the SIA suggests chipmakers hold strategic inventories of up to 12 months for some priority materials where feasible — 9 to 15 months for scandium metal — but that is not possible for every input. Your mitigation plan must distinguish stockpilable from non-stockpilable exposure — they need different strategies.
Upstream risk triage: the four layers at a glance
| Layer | What concentrates there | Buyer exposure signal | Your move | |---|---|---|---| | 1. Raw minerals | Gallium ~95%, tungsten ~85% (China, SIA estimates); Nd/Pr and terbium ~95% (China) | Any component whose chain touches these minerals; decade-scale replacement lead time | Origin disclosure from suppliers; longest-horizon planning | | 2. Precursors & chemicals | Photoresists, UHP gases, sputtering targets (production in JP/TW/KR, inputs often Chinese) | Supplier HQ in a "safe" country, production elsewhere | Ask where inputs are made, not where the supplier is based | | 3. Equipment sub-components | Turbo pumps, vacuum robots, transfer valves — often single-supplier | A tool with no second-source parts path | Map single-supplier sub-components; tie to equipment-maintenance contracts | | 4. Packaging & interconnect | ABF substrates ~75% Japan; mask blanks ~70% Japan | AI-accelerator and high-density packaging exposure | Alternate substrate/mask qualification; watch packaging capacity adds | *All concentration figures are SIA-report or media-report figures as cited above, not independent measurements.* ---
Frequently asked questions
**What did the SIA report actually find?** That the US chip industry's supply-chain exposure to China is larger than headquarters-based data suggests: 95% of equipment-and-materials spend goes to US/European/Japanese-headquartered suppliers, but only 68% by production location, on the SIA's estimates. Upstream, nearly a third of the ~100 upstream categories it analyzed rate high-risk, and over three-quarters of the Tier 1 categories it analyzed depend on at least one high-risk upstream input. It is an industry-association research report, not a government assessment. **Where can I read the SIA supply chain report?** It is titled "Understanding Dependencies in the U.S. Semiconductor Supply Chain," published October 9, 2026 by the Semiconductor Industry Association. Search that exact title — and read the headquarters-vs-origin methodology section first; it is the part most coverage skips, and it is the part you can reuse directly in supplier qualification. **Which chip materials is China most dominant in?** On the SIA's estimates: gallium (~95% of global supply), tungsten (~85%), and the rare earths neodymium/praseodymium and terbium (about 95% each). **What happens when China's export-control suspension expires on November 10, 2026?** Unknown — and note there are three separate measures: the suspension of the broader October controls runs to November 10, 2026; the separate suspension of the 2024 US-specific gallium/germanium/antimony ban was reported to run until November 27, 2026; and the earlier licensing requirements on gallium, germanium, graphite, and the seven rare earths added in April 2025 remain in place throughout. The November 10 date is a policy decision point to watch — the White House says both sides are working toward restoring shipments — not a forecast of disruption. **Can the US build alternative supplies for these materials?** The SIA's answer is measured: for some inputs, yes, with sustained investment and allied cooperation (it cites offtake agreements, the Pax Silica initiative, and R&D programs). But it warns capacity for some materials can take a decade or more to build, plus three to four years of supplier and material qualification. That is a lead-time warning, not a prediction of shortage. **How should component buyers respond?** Start with visibility: ask suppliers for origin-of-manufacture disclosure on critical inputs, map your top-spend BOM against the four upstream layers, write price-review clauses into contracts for exposed categories, and qualify second sources where concentration is highest. And distinguish stockpilable from non-stockpilable exposure — they need different plans. **Will this make specific chips more expensive?** No source in the reporting connects upstream concentration to any specific part number's price. The mechanism is directional — upstream cost and disruption risk compound through equipment and foundry layers before reaching components — which is why the response is contractual and structural (clauses, second sources, disclosure), not a price forecast. ---
Sources
Use the latest guidance for the exact part and manufacturer when deciding acceptance criteria.
- SIA: Understanding Dependencies in the U.S. Semiconductor Supply Chain (9 Oct 2026)
- SIA: SIA Report Highlights Strategies for Strengthening Supply Chains Supporting the U.S. Chip Ecosystem (9 Oct 2026)
- Sasakawa Peace Foundation: Why China's Rare-Earth Leverage Does Not Wear Out (Oct 2026)
- Mining Technology: China suspends ban on gallium, germanium, antimony exports to US (10 Nov 2025)