Federal Court Vacates Pentagon’s Anthropic Supply-Chain Risk Designation

A U.S. District Court ruled that the Defense Department unlawfully labeled Anthropic a supply-chain security risk and barred the AI company from federal work. Judge Rita Lin found insufficient evidence that Anthropic’s models posed a meaningful national-security threat, including claims it could use a purported model “kill switch” to sabotage systems, and ordered the designation removed.

The court held that the action was unlawful retaliation for Anthropic’s protected objections to using its AI for mass surveillance and autonomous lethal weapons, finding violations of the company’s First Amendment and Fifth Amendment due-process rights. The designation followed a dispute over a $200 million Pentagon contract for AI on classified systems and had prohibited military contractors and suppliers from working with Anthropic; the administration may appeal, while a separate Anthropic case in the D.C. Circuit remains pending.