Study Finds AI Supply Chain Gaps in Policy

Cambridge researchers report that more than one-third of 137 AI supply chain components go unaddressed in current government policies.

Updated on Nov. 22, 2026 in Artificial Intelligence

Study Finds AI Supply Chain Gaps in Policy

A University of Cambridge Bennett Institute for Public Policy research team compared 137 AI supply chain components against government policy documents from multiple countries and published its findings on November 18. The report concludes that South Korea's influence in the AI supply chain has been underestimated.

Why it matters

Midstream elements such as semiconductor wafers, packaging and networking are largely absent from policy, even though a small number of companies can control supply of specific components. Governments pursuing complete AI sovereignty may need to account for these chokepoints to reduce dependence on foreign suppliers.

The analysis mapped 137 components of the AI supply chain, with more than one-third left unaddressed in current policy frameworks. Advanced packaging technologies were flagged as a chokepoint that has largely escaped policy attention, and the report does not yet quantify how much supply risk these gaps carry.

The players

University of Cambridge Bennett Institute for Public Policy

A research institute within the University of Cambridge that studies public policy questions and publishes policy reports.

Jensen Huang

The chief executive of Nvidia, a leading designer of AI processors, who has frequently spoken publicly about the structure of the AI industry.

The details

Key AI processor manufacturing processes are concentrated in Taiwan, Vietnam and the Philippines, while South Korea's memory semiconductor strength depends on suppliers in Taiwan, the Netherlands, Singapore, Israel, the United States and Japan. The report also notes that several countries are pursuing complete AI sovereignty by operating the entire supply chain independently.

Timeline

  1. January 2026: Jensen Huang described the AI industry as a five-layer cake at Davos.

  2. November 18, 2026: The Cambridge Bennett Institute published the AI supply chain report on its website.

The Tech Race

Jensen Huang's five-layer description of the AI industry, offered at Davos in January 2026, frames the sector as a stack of interdependent layers. The Cambridge findings show that policy attention has concentrated on some of those layers while midstream chokepoints such as advanced packaging remain largely unaddressed.

Gaps in AI supply chain policy can affect the availability and pricing of AI-enabled devices and services that depend on semiconductors and advanced packaging. Users in countries that rely on imported components may see these pressures reflected in product costs over time.

The takeaway

AI capability depends on a chain of components from many countries, so policy that covers only one or two links leaves the rest exposed. Tracking midstream suppliers such as packaging and wafer makers offers a clearer view of where supply risk sits.

Further reading

For more on how governments are approaching the sector, see Artificial Intelligence coverage on Wisevoter.

Source note: This article includes information reported by Dongascience.