Trump Ousted: US Advanced Weapons Supplier Ready to Challenge the Government
President Donald Trump has publicly barred federal agencies from using AI tools developed by Anthropic. The hardline stance follows a dispute between Anthropic and the Department of Defence (DoD) that failed to reach agreement. Anthropic is the DoD’s official supplier of AI tools for military needs, notably in defence; in 2025 it secured a substantive contract valued at US$200 million.
Tensions intensified when Anthropic refused to permit its AI to be used to develop autonomous weapons or to facilitate domestic surveillance of US citizens. The DoD argued that the tool should be usable for actions that are legal.
Anthropic’s refusal drew a strong response from Trump, who labelled Anthropic a left‑wing, ‘woke’ company that threatens national security and endangers American lives. He also rejected the idea that Anthropic should dictate how the military fights and wins on the battlefield.
After banning Anthropic’s use, the Wall Street Journal reported that the United States’ strikes against Iran were still using Claude AI for intelligence assessments, target identification and battle‑scenario simulations.
Anthropic Ready to Fight the Government
On Tuesday 4 March, Anthropic received a letter from the DoD confirming that the company had been placed on the national security supply‑chain risk list. In a statement on its official site, CEO Dario Amodei responded: ‘As we have written before, we do not believe that taking this action [placing Anthropic on the supply‑chain risk list] is legal. We have no choice but to fight this in court,’ he wrote.
Amodei noted that the language used by the DoD in the letter aligned with Anthropic’s own statement from the previous week, that most customers remain unaffected by the supply‑chain risk designation. He stressed that the designation, even if disagreed with, applies only to uses of Claude directly connected to the DoD contract.
He said productive discussions with the DoD had taken place in recent days, including talks about ways the company could serve the DoD under two narrow exceptions they had proposed, and about how to ensure a smooth transition if those two exceptions proved unfeasible.
‘We are very proud of the work we have done with the Department, supporting soldiers on the front line with applications such as intelligence analysis, modelling and simulation, operational planning, cyber operations, and many more,’ Amodei wrote.
‘As we have stated before, we do not believe, and have never believed, that Anthropic’s or any private company’s role is to be involved in operational decision‑making, as that is the military’s responsibility. Our only concerns are the two exceptions for autonomous weapons and mass domestic surveillance, which relate to high‑level use and are not about operational decisions,’ Amodei emphasised.
Finally, Amodei said the company’s top priority is ensuring that soldiers and national‑security experts do not lose important tools during large‑scale combat operations. Anthropic will provision its models to the DoD and the national‑security community at a nominal cost, with ongoing support from its engineers, for as long as required to complete the transition and as long as it is allowed to do so.