Monday, November 16, 2026 Independent reporting on artificial intelligence Newsletter
AI News Beats
AI for Business

US and nine partners agree to coordinate AI supply chain protection

The San Francisco Declaration calls for five working groups covering policy, adoption and infrastructure security, but does not set out a timetable.

A close view of a chip and components on an electronics board
Photo: Ivan Chumak / Pexels

The United States and nine partner countries signed a declaration on November 15 to coordinate protection of technology and infrastructure essential to advanced AI. The agreement sets a common scope for cooperation, from the components that go into AI systems to the ports, power networks and software platforms that support them.

The US State Department named Australia, India, Israel, Japan, South Korea, Qatar, Singapore, the United Arab Emirates and the United Kingdom alongside the United States. The governments released the text as the San Francisco Declaration on the Secure Foundations of Enduring Prosperity. It was signed at a retreat of Pax Silica, the US-led technology and economic security initiative.

What the countries want to protect

The declaration distinguishes between sensitive technology and the infrastructure it depends on. The first category includes certain goods, software, materials and knowledge, as well as the inputs and production capabilities behind them. The test is not simply whether something is used in AI. The text ties sensitivity to the harm that theft, unauthorized transfer or misuse could cause, including whether access could give a hostile actor significant leverage over the participants.

The infrastructure category extends beyond the facilities that train or run AI models. It covers physical and digital systems that the participants rely on for essential services and economic activity. The declaration identifies ports, terminals, communications and logistics networks, energy infrastructure and the platforms that help coordinate them. In practical terms, a disruption to the movement of equipment or the supply of electricity can matter alongside access to an individual chip or model.

The text uses the term super intelligence, with a footnote saying some participants use it to mean artificial intelligence. The declaration does not depend on a finding that a particular model has reached a new level of capability.

Working groups are the next step

The State Department outlined five proposed working groups. They would cover policies intended to encourage innovation, adoption of AI in key sectors, cooperation across the supply chain, safe harbors and supply chain security hubs, and broader security and resilience. The groups are intended to turn a shared definition of what’s at risk into work governments can pursue together.

The new agreement builds on the original Pax Silica declaration. Japan’s Ministry of Foreign Affairs reported that seven countries signed that document at a Washington summit in December 2025, where governments discussed strengthening AI-related supply chains. The San Francisco text reaffirms cooperation on how countries enforce their existing policies and adds a structure for joint work on specific topics.

Neither the declaration nor the accompanying media note provides a schedule for the groups, names the first projects or announces a new export control. For companies that make chips, operate data centers or supply logistics services, the next indication of a practical change will be whether governments publish more specific criteria for what they intend to protect and how they plan to coordinate.

Stay on top of AI

Get the stories that matter in artificial intelligence, straight to your inbox.

Free. One email with the stories that matter. Unsubscribe any time. See our privacy policy.