The US Department of Defence has ceased using Anthropic's AI tools, a Pentagon official confirmed to the BBC on Monday, following the company's designation as a national security supply chain risk in February.
Defence Secretary Pete Hegseth ordered the phase-out by late August after Anthropic refused to remove safety guardrails from its Claude model and grant the military unrestricted access to its products. Anthropic cited concerns about mass surveillance and autonomous weapons development. The company has sued the Trump administration, calling the designation "unprecedented and unlawful."
The timing of the actual cessation remained unclear. Multiple sources told the BBC that Claude was still widely deployed across Pentagon operations as recently as last week, embedded in Maven Smart System, Palantir's primary intelligence platform. The system processes satellite imagery, drone footage and other visual data, which analysts fed into Claude for tasks including identifying potential military targets and compiling intelligence briefs.
Lauren Kahn, a senior research analyst at Georgetown's Center for Security and Emerging Technology and former US defence official, said the delay reflected the difficulty of removing deeply integrated systems. "Once they become integrated it can be painful to remove them," Kahn told the BBC.
Claude had been in government use since 2024, making Anthropic the first advanced AI company to deploy tools in classified US defence work. Pentagon employees continued using it throughout the dispute, including for intelligence gathering and analysis, according to sources familiar with defence operations.
The Pentagon has since signed contracts with Google, xAI and OpenAI. OpenAI tools have seen wider adoption in some military departments in recent months, two sources said.
The conflict between Anthropic and the Pentagon occurs as relations between the company and the White House appear to improve. Trump met with Anthropic CEO Dario Amodei twice at the White House last week. Following one meeting, Amodei said concerns about AI risks "are still under discussion" but expressed confidence that working with the president could address them safely. Trump praised Amodei as "great" and "fantastic."