US defence secretary Pete Hegseth has issued a stark ultimatum to Anthropic, warning the AI company it faces exclusion from the Pentagon’s supply chain unless it agrees by Friday to permit its technology to be used in “all lawful military applications.”
According to the Financial Times, Hegseth summoned Anthropic chief Dario Amodei to the Pentagon on Tuesday to press the issue, threatening to cut the company out of the defence supply chain or invoke the Defence Production Act to compel cooperation.
Core of the dispute
The standoff stems from a disagreement over conditional access to Anthropic’s Claude AI models. FT reports that Anthropic has resisted full Pentagon demands for unrestricted use — including for domains like mass domestic surveillance and lethal autonomous weapons systems.
Anthropic confirmed the meeting took place, saying the parties “continued good-faith conversations about our usage policy to ensure Anthropic can continue to support the government’s national security mission in line with what our models can reliably and responsibly do.”
The company has expressed particular concern about deploying advanced AI in scenarios where there is no human in the loop, arguing that even state-of-the-art models are not yet reliable enough to be trusted in those contexts.
Importance of Claude
Claude remains pivotal to Pentagon AI work. As FT notes, Anthropic’s model is currently the only one approved for classified US military missions, largely via a partnership with Palantir.
Anthropic reportedly has a $200m contract with the US defense department.
