US appeals court upholds Pentagon's national security threat designation for Anthropic

September 26, 2026 at 10:27 AM2 min

This digest was compiled by AI from multiple sources — links to the originals are below.

A US federal appeals court in Washington upheld the Pentagon's designation of AI startup Anthropic as a national security threat, rejecting the company's challenge by a 2-1 vote. The ruling preserves the March decision that barred Anthropic from defense contracts and prohibited contractors from using its Claude models. The decision comes despite a California federal court last month finding the designation unlawful.

Key Facts

The US Court of Appeals in Washington upheld the Pentagon's national security threat designation for Anthropic by a 2-1 vote on September 25, 2026.
The Pentagon first designated Anthropic a national security threat in March 2026, barring the company from defense contracts and prohibiting contractors from using its Claude AI models.
Judge Gregory Katsas wrote that the Defense Department provided compelling evidence that further integration of Claude into its information systems poses a legally supported national security risk.
Judge Neomi Rao joined the majority opinion, while Judge Karen LeCraft Henderson dissented.
A California federal court last month ruled the designation unlawful, but the Washington appeals court's decision means the Pentagon's action remains in force.

Appeals Court Ruling

The US Court of Appeals in Washington upheld the Pentagon's national security threat designation for Anthropic by a 2-1 vote on September 25, 2026. The court rejected Anthropic's argument that the Defense Department's March decision was arbitrary, unlawful, and unconstitutional. Judge Gregory Katsas wrote that the department provided compelling evidence that further integration of Claude into its information systems, including by contractors, poses a legally supported national security risk. Judge Neomi Rao joined the majority, while Judge Karen LeCraft Henderson dissented.

Pentagon Designation

The Pentagon first designated Anthropic a national security threat in March 2026 amid conflict with the company's founders over use of its AI models for battlefield targeting and surveillance of US citizens. The designation stripped Anthropic of contracts with the US military and barred Pentagon contractors from using the startup's technology in their work. A California federal court last month ruled the designation unlawful, but Anthropic had appealed the Pentagon's decision in two separate venues. The Washington appeals court's ruling means the Pentagon's designation remains legally valid.

Political Context

Anthropic founder Dario Amodei's September calls for voluntary AI slowdown drew criticism from President Donald Trump, who excluded him from a government dinner with Chinese leader Xi Jinping this week. On Monday, Trump wrote on Truth Social that his administration stopped AI figures like Amodei when they tried to do bad or potentially bad things.