Appeals court backs Pentagon in dispute with AI firm Anthropic

A three-judge panel's ruling has given the Defense Department a significant victory in its conflict with a prominent artificial intelligence firm. The decision may alter the future direction of US military AI oversight.

The appeals court, in a 2-1 vote, backed the Pentagon, affirming its move to bar Anthropic from all defense contracts.

Confirmed: @AnthropicAI = Supply Chain Risk.

The @DeptofWar does what is right for the Country and our Warriors. https://t.co/xERtGRh6Kc

— Pete Hegseth (@PeteHegseth) September 25, 2026

The Fight That Split Two Federal Courts

The conflict began in February, when Secretary of Defense Pete Hegseth ordered Anthropic to remove protections that prevented its Claude system from being used for fully autonomous lethal weapons and widespread domestic surveillance. The company rejected the demand on February 26, shortly before Hegseth's deadline.

On March 3, Hegseth officially labeled Anthropic a national security threat. This designation terminated the firm's current defense agreements and also blocked other Pentagon contractors from employing its technology.

The court's ruling noted that an Anthropic executive had independently raised concerns about a contractor's deployment of Claude in a particular military operation.

According to media accounts included in the case record, the mission in question was the January 3 operation that resulted in the capture of Venezuelan President Nicolás Maduro.

Pentagon officials pointed to that incident as proof that Claude's actions were still too erratic to be trusted in military contexts.

The majority opinion was authored by Judge Gregory Katsas, with Judge Neomi Rao concurring. Katsas stated that the Defense Department had 'ample support' for its determination. Judge Karen LeCraft Henderson opposed the ruling.

Why Anthropic Says This Fight Isn’t Over

Just last month, a different federal court in California arrived at a contrary judgment. Judge Rita Lin deemed a wider, parallel ban by the Trump administration unconstitutional, characterizing it as retaliation designed to turn Anthropic into 'a public example'.

Anthropic offered a strong response to the Friday verdict. A company representative, as cited by CNBC, stated: 'We respectfully disagree with the court’s decision. Another federal court has already held the government’s parallel designation unlawful.'

It remains uncertain whether Anthropic will request a rehearing or escalate the matter to the US Supreme Court, leaving the country's latest artificial intelligence legal struggle unresolved between two conflicting lower courts.