The US military is still using Claude — but defense-tech clients are fleeing

Anthropic's Claude AI models are currently being used by the U.S. military for targeting decisions in the ongoing conflict with Iran, even as the company faces a mass exodus of defense industry clients. The situation stems from conflicting government directives: President Trump ordered civilian agencies to discontinue Anthropic products, while the Department of Defense was given a six-month wind-down period. This timeline was upended when the U.S. and Israel launched a surprise attack on Tehran, leaving Claude integrated into operational systems. According to a Washington Post report, Claude works in conjunction with Palantir's Maven system to suggest hundreds of targets, provide precise coordinates, and prioritize them by importance for Pentagon planners.

Despite this active battlefield use, Anthropic's position in defense tech is collapsing. Major contractors like Lockheed Martin are already swapping out Claude models for alternatives, and venture firm J2 Ventures reports that 10 of its portfolio companies are in active processes to replace Claude for defense use cases. The central unresolved issue is whether Defense Secretary Pete Hegseth will formally designate Anthropic as a supply-chain risk, which would trigger legal battles. For now, the result is a stark paradox: one of the world's leading AI models is being systematically removed from the defense industrial base while simultaneously providing critical intelligence support in a live warzone, highlighting the complex and often contradictory pressures facing AI companies operating in the national security sector.