Judge rules Pentagon’s blacklisting of Anthropic illegal

The ruling concludes that Defense Secretary Pete Hegseth’s decision to label the Claude maker a supply chain threat was arbitrary and capricious. The conflict originated from Anthropic’s refusal to remove safety guardrails that would have allowed the Pentagon to utilize its models for autonomous weapons and domestic surveillance. Judge Lin noted that the government’s own actions contradicted the risk designation, pointing to the Department of Defense’s continued pursuit of contracts with the company and their collaboration on the Mythos cybersecurity model.

Lin highlighted that the government’s behavior appeared driven by a desire to punish the company for its public criticism of federal policy. She dismissed the Pentagon’s claims of potential backdoor access, noting the lack of evidence supporting such concerns. While the Department of Defense retains the authority to choose its AI vendors, the court ruled that national security cannot be used as a pretext to retaliate against private firms. Anthropic welcomed the decision, while a separate legal challenge filed by the company remains active in Washington, D.C.