Defense Secretary Pete Hegseth demanded Anthropic grant the U.S. Military full access to its artificial intelligence model, Claude, by the end of this week, according to sources familiar with the discussions.
The ultimatum, delivered during a meeting at the Pentagon on Tuesday, threatens to escalate tensions between the Defense Department and the AI company, which secured a $200 million contract in July to develop AI capabilities for national security purposes. Hegseth reportedly gave Anthropic CEO Dario Amodei until Friday to provide a signed document authorizing complete military access to Claude.
At the core of the dispute are Anthropic’s safeguards designed to prevent the AI from being used for mass surveillance of American citizens or employed in autonomous weapons systems. Anthropic has repeatedly sought written assurances from the Pentagon that Claude would not be utilized for such purposes, a request Defense officials have dismissed as unnecessary, asserting that any leverage would be lawful. One source indicated Amodei is concerned Claude is not reliable enough to avoid potentially lethal mistakes in military operations without human oversight.
The Pentagon is prepared to invoke the Defense Production Act to compel Anthropic’s compliance, or designate the company as a supply chain risk, effectively hindering its ability to work with the government, sources said. The Defense Production Act allows the government to prioritize contracts and influence businesses in the interest of national defense. CNN reported Hegseth stated he would ensure the act is invoked if Anthropic does not “secure on board” by 5:01 p.m. On Friday.
A senior Pentagon official, speaking on background, claimed the dispute does not concern mass surveillance or autonomous weapons, stating the military is only seeking a license for lawful activities. However, Anthropic maintains its concerns regarding the potential for misuse.
The standoff comes as other AI companies appear more willing to cooperate with the Pentagon. According to sources, xAI, owned by Elon Musk, is already on board with being used in a classified setting, and other companies are reportedly nearing agreement with the Defense Department.
As of Wednesday, Anthropic had not publicly responded to the ultimatum. The company has until Friday to comply with the Pentagon’s demands or face potential repercussions, including the loss of its contract and a possible designation as a supply chain risk.
