filed a petition on October 13, 2026 asking the to its appeal of the 's blacklisting of its AI models.
The label Anthropic wants undone is , applied by the Pentagon in March 2026. It blocked defense contractors from using Anthropic's technology on agency work.
The designation followed collapsed negotiations over how the military could use Anthropic's Claude models. Anthropic says it declined to let the models be used for mass surveillance of Americans or , and explained those AI-safety commitments publicly.
In September 2026 the D.C. Circuit upheld the blacklisting on a , rejecting Anthropic's argument that the ban was arbitrary, unauthorized and unconstitutional.
A divided panel is the standard ground for asking the full court to reconsider. A unanimous loss would have left Anthropic with almost nothing to argue.
Anthropic's attorneys wrote in the petition that the court ignored straightforward evidence the designation was retaliation for the company's speech about AI safety.
A San Francisco federal judge ruled in August 2026 that the Pentagon had violated Anthropic's First Amendment rights, and that the supply chain risk designation amounted to illegal retaliation.
The dispute began after Anthropic signed a $200 million contract with the Pentagon in July 2025. Talks collapsed in September 2025 as the company negotiated Claude's deployment on the department's platform.
Anthropic declined to comment beyond the filing. A Pentagon spokesperson said the agency does not comment on pending or ongoing litigation.