A US federal judge in San Francisco has temporarily blocked the Pentagon’s designation of AI firm Anthropic as a supply chain risk, halting a directive that ordered federal agencies to stop using its Claude chatbot. Judge Rita Lin ruled that the government’s actions appeared arbitrary and potentially retaliatory, stating there was no legal basis to label an American company a national security threat for disagreeing with government policy.

The dispute began after negotiations collapsed over a Pentagon contract that would allow military use of Anthropic’s AI without restrictions. Anthropic opposed deploying its technology for lethal autonomous weapons and mass domestic surveillance, prompting the administration to move toward a government-wide ban. The court’s injunction allows Anthropic to continue operating with federal agencies while the lawsuit proceeds.