Hours before the U.S. bombing campaign in Iran began, President Donald Trump announced he was banning all government agencies from using Anthropic's AI tools, giving the Pentagon six months to phase them out. But the military will continue using Claude in the Iran campaign as it waits for a replacement, as reported by "Hvylya", citing The Washington Post.
The ban followed a bitter fight between Anthropic CEO Dario Amodei and the Trump administration over the terms of Claude's use in warfare. The dispute centered on two red lines drawn by Anthropic: the use of its tools for mass domestic surveillance and for fully autonomous weapons. Negotiations reached an impasse, and the administration moved to sever ties.
Yet the military's dependence on Claude has grown so deep that walking away is not an option mid-campaign. Commanders have become so reliant on the Maven system - which uses Claude to process classified intelligence, generate targets, and prioritize strikes - that the administration would use government powers to retain the technology if Amodei tried to shut it down. "Whether his morals are right or wrong or whatever, we're not going to let his decision making cost a single American life," one source said.
Amodei, for his part, framed the standoff as a principled position. "I believe deeply in the existential importance of using AI to defend the United States and other democracies, and to defeat our autocratic adversaries," he wrote in a blog post as talks collapsed. He confirmed that Claude was "extensively deployed" across the Defense Department and other security agencies.
The six-month phase-out window means Claude will remain active through the most intense phase of the Iran campaign. The Pentagon did not respond to a request for comment. Anthropic and Palantir both declined to comment.
Also read: OpenAI Rushed to Cut Pentagon Deal Hours After Anthropic Blacklisting, Acemoglu Reveals