Pentagon Upholds Blacklisting of AI Firm Anthropic

The Pentagon has upheld Anthropic's supply-chain risk designation. The decision follows a previous court ruling. Anthropic had faced challenges in the appeal, but ultimately lost its bid to block the blacklisting.

A federal appeals court in Washington, D.C. denied Anthropic's request to temporarily block the Department of Defense's supply-chain risk designation, effectively upholding the Pentagon's ability to blacklist the AI company while underlying litigation continues. The ruling follows a series of escalating legal clashes since the Trump administration formally designated Anthropic as a supply-chain risk in early March, barring defense contractors from using Claude models in military work and blocking Anthropic from new Pentagon contracts.

The dispute traces back to a $200 million contract Anthropic signed with the Pentagon in July 2025. Negotiations over Claude's deployment on the DOD's GenAI.mil platform broke down when the Pentagon sought unfettered model access for all lawful purposes, while Anthropic sought assurances its technology would not be used for fully autonomous weapons or domestic mass surveillance. A separate federal judge in California issued an injunction blocking portions of the Pentagon's measures, creating conflicting rulings across jurisdictions that will likely require higher court resolution.

The outcome has broad implications for the AI industry's government contracting ambitions. The case sets a precedent for how AI companies can negotiate usage policy limits with federal clients — and whether safety-oriented refusals to grant unlimited model access can be treated as grounds for supply-chain exclusion. Competitors including MSFT and GOOGL, which have deeper existing Pentagon relationships, could benefit if Anthropic's enterprise government pipeline is disrupted.