Governments Overlook a Third of the Global AI Supply Chain, Cambridge Report Finds

November 18, 2026
Governments Overlook a Third of the Global AI Supply Chain, Cambridge Report Finds
  • A University of Cambridge audit finds that governments are overlooking much of the global AI supply chain: after mapping 137 elements across 14 layers and comparing them with 117 policy documents from 43 countries and territories, the EU and OECD, researchers found that more than a third of the elements were not mentioned.

  • The blind spots cluster in the supply chain’s “physical middle”—including chip design, testing and packaging, semiconductor chemicals, design software and server hardware—even as policy discussions concentrate on models, data centres, funding and governance.

  • The report argues that AI production is globally interdependent, with influential capabilities spread well beyond the United States and China; no country controls the entire chain, making complete national self-sufficiency impossible.

  • Britain is linked to 41 supply-chain positions across 13 of the 14 layers, but the audit says the number alone does not establish leverage: a position matters most when it is scarce, controllable and recognized in policy.

  • UK policy documents mention only three of the 41 positions identified, overlooking potential strengths such as Arm’s chip-design intellectual property, AI security research, health and genomic data, and industrial gases used in semiconductor plants.

  • Arm’s processor IP is identified as Britain’s clearest difficult-to-replace capability, with about 41% of the semiconductor-IP market; however, Arm is headquartered in Britain, more than 80% owned by Japan’s SoftBank and listed in the US, limiting the UK government’s control.

  • The audit also flags chip design and industrial gases supplied by Linde to semiconductor plants, while its authors warn that broad references to supply chains can obscure the specific positions governments might protect or use as leverage.

  • Other overlooked chokepoints include electronic design automation software, dominated by Synopsys, Cadence and Siemens EDA, and advanced packaging, which connects chips into high-performance AI processors.

  • The report highlights dependencies outside the US and China, including Japan’s Ajinomoto, which supplies more than 95% of a material used to insulate advanced AI chips.

  • Leverage is distributed across other economies too: Belgium has strengths in recycled semiconductor chemicals and nanoelectronics research, Taiwan, Vietnam and the Philippines are central to advanced chip packaging, and Ireland can influence AI capacity through energy, planning and its role as a base for major technology companies.

  • AI minister Kanishka Narayan argues that Britain need not build its own frontier model if it can secure trusted access to other systems, while investing in chip design, compound semiconductors and scientific research; the strategy depends on maintaining access and developing capabilities other countries cannot readily replace.

  • The researchers developed a toolkit to help governments identify supply-chain dependencies, assess their leverage, diversify suppliers and decide when to build domestic capacity or move away from an outside supplier.

Summary based on 4 sources


Get a daily email with more AI stories

More Stories