Anthropic asks D.C. Circuit to rehear Pentagon blacklisting appeal

Anthropic has asked the D.C. Circuit to rehear its Pentagon blacklisting appeal, seeking to overturn a September ruling that allowed the Defense Department to exclude its Claude AI models from the department’s supply chain. For investors watching Anthropic’s path toward a potential public offering, the filing keeps a major dispute over government business and AI-use limits active, but it does not change the existing court ruling.

The company filed its petition Tuesday, according to CNBC. Anthropic says the Pentagon’s designation as a national-security supply-chain risk was retaliation for the company’s public AI-safety positions, including its refusal to permit Claude’s use in lethal autonomous warfare or domestic surveillance.

The D.C. Circuit has not indicated that it will grant a rehearing. Anthropic declined further comment beyond its filing, CNBC reported, and the Pentagon did not immediately respond to the outlet’s request for comment.

What did the D.C. Circuit decide in Anthropic’s Pentagon case?

In a 2-1 decision dated Sept. 25, the U.S. Court of Appeals for the District of Columbia Circuit rejected Anthropic’s challenge to the department’s action. The case concerned an exclusion of Claude under the Federal Acquisition Supply Chain Security Act of 2018.

That law allows agencies to take specified procurement actions, including barring agency contracts with a supplier and subcontracts that use that supplier for agency work, after making required national-security and supply-chain-risk findings. CNBC reported that the Pentagon’s label blocked defense contractors from using Anthropic technology in work for the department.

The majority, in an opinion by Judge Gregory Katsas, found sufficient support for the department’s conclusion that Claude posed a covered national-security risk. The court pointed to Anthropic’s built-in model restrictions, instances in which those restrictions stopped tasks requested by government users, and a disagreement over an overseas military operation that left the department uncertain whether Claude would perform as intended.

The panel also rejected Anthropic’s arguments that the action exceeded the department’s legal authority and violated due-process and First Amendment protections. The majority said the exclusion rested on Anthropic’s refusal to accept a contract term the department considered essential, rather than on the company’s policy views about AI regulation.

Judge Karen LeCraft Henderson dissented. According to ABC News, she said a contractor’s transparent enforcement of its restrictions did not amount to the sort of supply-chain risk covered by the statute.

Why is Anthropic seeking another hearing?

Anthropic’s lawyers argue in the new petition that the appellate court did not account for evidence that the designation retaliated against the company’s speech on AI safety, CNBC reported. Anthropic has said AI is not yet reliable enough to use safely in autonomous weapons and has opposed domestic surveillance, Reuters reported.

The disagreement followed failed negotiations over military deployment of Claude. Anthropic had signed a $200 million Pentagon contract in July 2025, CNBC reported, before talks over deployment on the department’s GenAI.mil platform later broke down.

A separate court case remains distinct

The D.C. Circuit case concerns one of two supply-chain designations issued under different laws. In separate litigation in San Francisco, a federal judge invalidated the parallel designation and blocked a government-wide ban on Anthropic and an order barring military contractors from doing business with the company, Reuters reported.

That separate ruling remains different from the designation upheld by the D.C. Circuit. Anthropic’s latest petition asks the Washington appeals court to reconsider only the decision in its own case.

This story draws on original reporting from CNBC.