SCMP Tech→ original

SIA: China Supplies the US Just 5% of Chip Equipment, but Over 90% of Some Rare Earths

The Semiconductor Industry Association has published a map of the supply chain behind the US chip industry. By country of manufacture, China supplies about 5% of the equipment and materials that US chipmakers buy directly. They spend almost $27 billion a year on these items, and 95% of that goes to companies from the US, Europe and Japan. But for some rare earth elements needed in chip production, China's share exceeds 90%. • Direct supply from China: about 5%. • US chipmakers' spending: almost $27 billion a year. • Rare earths: dependence on China above 90%.

AI-processed from SCMP Tech; edited by Hamidun News
SIA: China Supplies the US Just 5% of Chip Equipment, but Over 90% of Some Rare Earths
Source: SCMP Tech. Collage: Hamidun News.
◐ Listen to article

China supplies about 5% of the equipment and materials sold directly to the US semiconductor industry, measured by country of manufacture. For some rare earth elements that chipmaking cannot do without, however, its share exceeds 90%. That is the conclusion of the Semiconductor Industry Association (SIA), which mapped the US supply chain.

The baseline numbers are these. US chipmakers spend almost $27 billion a year on manufacturing equipment and materials. About 95% of that money goes to companies headquartered in the US, Europe and Japan. At first glance, there is almost no dependence on Beijing. But the report's headline points the other way: China occupies far more of the supply chain than the direct statistics show. The available portion of the material breaks off at this point, so details of the SIA's methodology are not recounted here.

Why direct statistics understate the dependence

Accounting at the first link of the chain shows who billed the fab. It does not show where the supplier got its raw materials and components. A tool from a Japanese or European company may contain parts or materials that were mined and processed in China. Procurement records will still list it as Japanese or European.

That is why the 5% and 90% figures do not contradict each other. The first describes the final seller. The second describes the depth of the chain, where rare earth processing has historically been concentrated in a single country.

Where Beijing holds leverage

Rare earth elements are needed in small quantities, but they are nearly irreplaceable. China already has experience applying export controls in this market. Restrictions on gallium and germanium have been in force since 2023, and in 2025 Beijing introduced licensing for a number of rare earth metals. This is context from open sources, not data from the SIA report. But it explains why the industry now measures dependence several tiers deep.

Key metrics from the available part of the report:

  • China's share of equipment and materials supplied directly to the US chip industry, by country of manufacture: about 5%.
  • Annual spending by US chipmakers on manufacturing equipment and materials: almost $27 billion.
  • Share of suppliers headquartered in the US, Europe and Japan: about 95%.
  • China's control over certain rare earth elements used in chip production: more than 90%.

Practical takeaways

For the AI infrastructure market, the takeaway is simple: look for supply chain risk deeper than the first supplier. Building fabs in the US and diversifying equipment purchases do not, by themselves, remove dependence on raw materials.

For companies that buy chips and build data centers, this is a signal to plan for delays and rising materials prices. GPU delivery times depend not only on TSMC capacity and packaging, but also on metals that rarely make it into presentations.

For policymakers and investors, the priority shifts to rare earth processing and separation outside China. Mining without processing does not solve the problem. Watch how the SIA and US authorities break this data down in more detail, and watch for any new Chinese licensing rules. Those will show how real the threat is that the supply chain mapping has just outlined.

ZK
Hamidun News
AI news without noise. Daily editorial selection from 50+ sources. A product by Zhemal Khamidun, Head of AI at Alpina Digital.
✦ Unified AI Workspace

Access Frontier LLMs & Multimodal AI in One Workspace

Discover NomadNet AI (nomadnet.ai) — unified access to frontier language models, research synthesis, and multimodal workflows.

Try NomadNet AI →Launch NomadNet Agent
What do you think?
Loading comments…