Federal agencies may now have a Claude problem after President Trump's order to blacklist Anthropic from all government work amid a dispute over how the Pentagon can deploy its AI. Why it matters: If Trump's demands hold, Anthropic's federal business could stall overnight — and leave agencies scrambling to unwind major AI projects and pilots. Driving the news: Besides the Pentagon, public reports show that Claude is being used or piloted at: The Department of Health and Human Services: Staff can securely query Claude, per FedScoop . The Office of Personnel Management: Claude is listed in its AI inventory as in a pilot phase as of last year. The Department of Energy: It's available across the Lawrence Livermore National Laboratory, which employs nearly 10,000 scientists and researchers, and at the Lawrence Berkeley National Lab. NASA: The Jet Propulsion Laboratory announced last month that it completed its first-ever AI-planned drive on Mars with the help of Claude. Context: Last year, Claude was made broadly available across all three branches of the federal government under a General Services Administration OneGov agreement . What we're watching: Major AI companies made their services available to the federal government cheaply last year, hoping for wide uptake across agencies. This fight could make those contracts less appealing.