Appeals Court Backs Pentagon's Blacklisting of Anthropic Over Claude

A divided D.C. Circuit panel ruled 2-1 on Friday to uphold the Defense Department’s designation of Anthropic as a supply chain risk, keeping the company’s Claude models barred from military systems and from defense contractors’ government work. The majority, written by Judge Gregory Katsas and joined by Judge Neomi Rao (both Trump appointees), found the Pentagon had sufficient grounds to treat Claude’s integration as a covered national-security risk; Judge Karen Henderson dissented. The ruling leaves Anthropic split across two courts: a San Francisco judge struck down one of the two designations underpinning the ban last month, while the D.C. panel has now sustained the other.

The designation traces back to a March labeling that followed the collapse of negotiations over deploying Claude on the Pentagon’s GenAI.mil platform, despite a $200 million contract signed in July 2025. The core disagreement was over control: the DOD sought unfettered access to the models for all lawful purposes, while Anthropic pushed for guarantees that its technology would not be used in fully autonomous weapons or domestic mass surveillance. Defense Secretary Pete Hegseth cast that stance as an attempt to seize veto power over military operations and raised concerns that overly constrained models could shut down unexpectedly or be manipulated—arguments the court credited by deferring to executive judgment on national-security tradeoffs.

The decision sits against a backdrop of open hostility between Anthropic and the administration, including repeated public attacks on CEO Dario Amodei. The panel stayed its ruling to give Anthropic room to seek a rehearing, an en banc review by the full circuit, or Supreme Court intervention, and the company signaled it is weighing all options. The dueling outcomes in the two cases set up a likely next round of litigation over how far agencies can go in cutting a major AI vendor out of federal systems on national-security grounds.