A top lawmaker on the Senate Armed Services Committee has requested a briefing from the Pentagon on its escalating feud with AI startup Anthropic, which has prompted the Trump administration to threaten to declare the company a supply chain risk.
“I’ll withhold judgment until I’ve had my briefing,” Sen. Mike Rounds (R-S.D.), chair of the panel’s cybersecurity subcommittee, told reporters Tuesday, in response to questions from POLITICO about the unprecedented move.
A supply chain risk designation would result in Anthropic being blacklisted from government contracts — something typically reserved for firms with ties to U.S. foreign adversaries. The technology company is expected to challenge such a designation in court.
The Defense Department and Anthropic did not immediately respond to requests for comment.
Negotiations with the Pentagon over use of Anthropic’s AI systems hit a roadblock last week when the firm refused to lift restrictions on the military’s unfettered access to its technology. The startup had sought to stop its systems from being used in fully autonomous weapons systems or for domestic surveillance.
President Donald Trump, roughly one hour before the deadline to meet the Pentagon’s demands, directed all federal agencies to cease use of Anthropic’s AI and declared a six-month phase-out period in a post on Truth Social. Trump threatened “major civil and criminal consequences,” and Defense Secretary Pete Hegseth in an X post accused the company of endangering national security.
Last week, key defense policy lawmakers on the Hill urged a ceasefire as tensions between Hegseth and Anthropic CEO Dario Amodei escalated. Top Senate defense appropriators Mitch McConnell (R-Ky.) and Chris Coons (D-Del.), as well as Senate Armed Services Committee leaders Roger Wicker (R-Miss.) and Jack Reed (D-R.I.), sent a letter to the two men expressing “concern over the escalatory direction of negotiations between the Department of Defense and Anthropic.”
Sen. Thom Tillis (R-N.C.), who is retiring at the end of his term in 2027, also weighed in Tuesday, telling reporters: “They’re telling Anthropic that they should compromise their code of conduct to facilitate whatever it is Hegseth or somebody wants.” He added that this will give other AI upstarts the “green light” to cast aside imposing their own codes of conduct.
