A three-judge panel of the U.S. Court of Appeals for the District of Columbia declined to overturn the Department of Defense’s supply-chain risk designation on Anthropic, allowing the Pentagon’s ban on the company’s Claude models to remain in effect. The ruling, issued on a Friday, affirmed the lower-court decision that one of two designations stays in place indefinitely, while the other was previously dismissed by a San Francisco judge.
Earlier this year the Pentagon invoked two separate supply-chain statutes to compel the removal of Anthropic’s Claude AI from military and other federal systems by the end of the month. Department of Defense officials argued that Anthropic’s refusal to permit its models for autonomous weapons or domestic surveillance posed a significant national-security risk, prompting the agency to label the firm a supply-chain liability.
Anthropic’s spokesperson, Danielle Cohen, said the company remains confident and is evaluating further legal avenues, including a petition to a broader panel of the D.C. Circuit or a petition for certiorari before the Supreme Court. The firm maintains that its stance reflects ethical limits rather than opposition to broader governmental AI regulation, a point the judges found insufficient to overturn the designation.
Following the designations, Anthropic reported a loss of business as customers hesitated to engage with a company flagged as a government pariah. While the firm has not disclosed precise financial effects, it continues to cite rising sales and is moving toward a potential initial public offering later in the year, a timeline now complicated by the ongoing legal and procurement constraints.
The Pentagon has not released detailed plans for replacing Claude, but officials have mentioned alternatives such as SpaceX’s Grok, Google’s Gemini, and OpenAI’s GPT series. Employees at Google and OpenAI have voiced ethical concerns about their companies’ military contracts, yet corporate leadership has emphasized the strategic importance of supporting U.S. defense initiatives despite internal dissent.
During the hearing, the judges split 2-1, with the majority rejecting Anthropic’s due-process and free-speech arguments and characterizing the dispute as a routine contract negotiation. The opinion noted that the department’s exclusion of Anthropic stemmed from the company’s refusal to accept a contract term deemed essential, not from any broader regulatory stance, leaving the ban in place while further appeals may unfold.