Trump Cuts Ties With Anthropic; Pentagon Flags 'Risk'
By @sharedot · · 8 pages
- AI
- Anthropic
- Pentagon
- AI Policy
Trump ordered federal agencies to stop using Anthropic, with Hegseth designating the AI firm a Pentagon supply-chain risk.
What Happened at the Friday Deadline
After a 5 p.m.
Friday deadline passed without Anthropic agreeing to the Pentagon's terms, President Trump ordered every federal agency to 'IMMEDIATELY CEASE all use of Anthropic's technology,' promising a six-month phase-out and warning of 'major civil and criminal consequences' if the company was unhelpful.
Defense Secretary Pete Hegseth then moved to designate Anthropic a national security 'supply-chain risk' — a designation he said is usually reserved for foreign adversaries — barring defense contractors from using its AI, while allowing services to continue up to six months for a 'seamless transition.'
Anthropic's Two Red Lines
Anthropic CEO Dario Amodei had made clear the company refuses to drop two demands: no use of its AI for fully autonomous weapons, meaning AI rather than humans making final battlefield targeting decisions, and no mass domestic surveillance.
The company said it supports all lawful uses of AI for national security aside from those two narrow exceptions, but told ABC News that the Pentagon's latest contract language 'made virtually no progress' on those safeguards and added legalese that would allow the safeguards to be 'disregarded at will.'
Anthropic called Hegseth's designation 'legally unsound' and said it would challenge it in court.
Why the Escalation Is Surprising
The designation is extraordinary because supply-chain risk status has historically been reserved for foreign adversaries, not an American frontier AI company.
Adam Conner, vice president for technology policy at American Progress, wrote that the Pentagon could 'essentially declare war not on a foreign nation but on one of America's most successful frontier AI companies,' calling it an unprecedented peacetime move signaling that private firms must do the administration's bidding or face existential consequences.
That a company known for safety advocacy is now branded a security threat breaks the expected pattern entirely.
OpenAI Lands the Pentagon Deal Instead
In a striking counterpoint, OpenAI announced late Friday it had reached a deal with the Pentagon to deploy its AI models on the department's classified network.
Sam Altman said two of OpenAI's most important safety principles are prohibitions on domestic mass surveillance and human responsibility for the use of force, including autonomous weapon systems, and claimed the Pentagon 'agrees with these principles, reflects them in law and policy' in the agreement.
OpenAI asked the Department of War to offer the same terms to all AI companies, which it said everyone should be willing to accept.
Congress Tries to Intervene
Ahead of the deadline, top members of the Senate Armed Services Committee sent a private letter to Anthropic and the Pentagon urging them to extend negotiations and work with Congress.
The letter, obtained by ABC News, noted the Pentagon has stated it does not intend to conduct mass surveillance or use autonomous weapons without humans on the loop — positions Congress endorses — but acknowledged the question of 'lawful use' requires additional work, including whether new legislative or regulatory language is needed.
The standoff now lands squarely in a Congress already divided over AI oversight.
What Comes Next
Anthropic says it will challenge the supply-chain risk designation in court, setting up a legal fight over whether the government can punish a contractor for refusing terms it deems unsafe.
The six-month phase-out gives agencies and defense contractors time to transition off Anthropic's products, but the company has received no direct communication from the Pentagon or White House on negotiation status.
The dispute also shadows Anthropic's separate lobbying push for its draft federal AI safety bill, which would sharply limit government power to block frontier models — a proposal whose high intervention threshold critics continue to contest.