Trump orders all federal agencies to stop using Anthropic AI

President Trump ordered every U.S. federal agency to "IMMEDIATELY CEASE all use of Anthropic's technology" via a Truth Social post on February 27, 2026, calling the company "Leftwing nut jobs" who made a "DISASTROUS MISTAKE." The directive follows Anthropic CEO Dario Amodei's public refusal to comply with Pentagon demands that Claude -- the AI model deployed on U.S. classified military networks -- be made available without restrictions on mass domestic surveillance and fully autonomous weapons.

The ban terminates Anthropic's $200 million Pentagon contract, under which Claude became the first frontier AI model cleared for use on classified U.S. military systems. Despite the "immediate" language, agencies already deploying Anthropic products -- including the Pentagon -- receive a six-month phaseout window; all others must stop at once. This is the first time a U.S. president has directed federal agencies to abandon a commercial AI vendor over the vendor's refusal to remove its own safety limits. xAI's Grok, which signed an "all lawful use" deal with the Pentagon on February 23, is positioned as the administration's preferred replacement on classified networks -- though defense officials have acknowledged it is unlikely to serve as a like-for-like substitute. Note: this publication is built on Anthropic's Claude; we disclose that relationship and it does not affect our coverage.