The standoff between the U.S. Department of Defense and AI giant Anthropic is escalating into an unprecedented battle over technological sovereignty. U.S. Secretary of Defense Pete Hegseth recently issued a stern warning to Anthropic: if the company does not agree to the military's terms by this Friday's deadline, the Pentagon will invoke the Defense Production Act (DPA) and demand the sharing of innovative technologies on grounds of national security.

The core of this controversy lies in the "boundaries of AI technology use." AIbase learned that as the first company to integrate its technology into the U.S. military's classified network, Anthropic co-founder Dario Amodei has taken an extremely firm stance. He clearly stated that his Claude series large models must not be used to power autonomous weapons systems or for mass surveillance of the public.

Their tense relationship has further escalated in recent discussions about specific military operations. The Pentagon argues that its requirements are entirely within legal bounds and do not involve illegal use. Defense officials have also openly warned that if Anthropic continues to impose restrictions, it will be seen as a "supply chain risk," which could lead to the company and its partners losing all future government business opportunities.

Currently, Anthropic has shown a "break or die" attitude, stating that if concerns about technology abuse cannot be resolved, it is prepared to exit a $200 million defense contract. Although legal experts question whether the Defense Production Act has the legal authority to force companies to abandon their technical restrictions, the "5:01 p.m. Friday" deadline set by the Pentagon is rapidly approaching. This clash between "tech company values" and "national security will" may redefine the collaboration model between Silicon Valley and the military.