Anthropic just lost its bid to get back into the Pentagon's good graces.
A federal appeals court on Friday upheld the Defense Department's designation of Anthropic as a supply chain risk, a label that blocks the military and its contractors from running Claude models. The DC Circuit Court of Appeals ruled 2-1 against Anthropic, rejecting the company's petitions to overturn the designation. Circuit Judge Gregory Katsas wrote the majority opinion. The ruling keeps the Pentagon's restriction in place, at least for now.
This is a competitive problem, not just a paperwork one. Federal contracts are becoming a serious revenue line for AI labs eager to prove their models belong in security-sensitive environments, and a supply chain risk label is exactly the kind of flag that shows up in due-diligence checklists for years. Losing this appeal makes it harder for Anthropic to compete for federal business against rivals that carry no such designation.
The ruling doesn't establish that Claude is unsafe. It establishes that the Pentagon gets to make that call, and a divided court wasn't willing to overrule it.