A US federal appeals court has allowed the Pentagon to keep Anthropic designated as a supply-chain risk, preserving a block on Claude across the military and related federal systems. The DC Circuit panel ruled 2–1 that the department had enough support for its national-security conclusion.
The dispute centers on restrictions Anthropic places inside Claude. The company has said its current models may not support autonomous weapons or domestic surveillance, while Defense Secretary Pete Hegseth treated those limits as a risk to government access and operations. Anthropic says it remains confident in its case and is considering further options.
Two separate supply-chain laws produced parallel challenges. A federal judge in San Francisco struck down one designation, but the DC ruling leaves the other in force indefinitely. Both cases can continue through appeals, potentially for years. Anthropic could ask the full circuit court to rehear the case or petition the Supreme Court.
The Pentagon has not detailed progress replacing Claude with alternatives from OpenAI, Google or xAI. The immediate effect is continued exclusion rather than a final nationwide precedent on whether a model provider may enforce its own use restrictions against a government customer.