A U.S. appeals court in Washington ruled 2-1 on September 25 that the Pentagon can keep Anthropic on its national security supply-chain risk list. Judge Gregory Katsas, writing for the majority, found the March 2026 designation reasonable and lawful, and rejected Anthropic's claim that it was retaliation for the company's AI safety positions.
The dispute began when Anthropic refused to remove guardrails that block use of its models for fully autonomous weapons and mass surveillance. The majority argued that "overly constrained AI models shutting down unexpectedly" could "cause important military operations to fail." The blacklisting has reportedly cost the company billions in lost contracts ahead of its planned IPO.
The ruling means Claude stays barred inside the Pentagon. A separate San Francisco federal ruling in August, which struck down a parallel designation under a different law, still lets other agencies and contractors work with Anthropic. The company is weighing a request for review by the full appeals court.