Hegseth to Meet Anthropic CEO as Pentagon Threatens Supply Chain Ban

U.S. Defense Secretary Pete Hegseth is set to confront Anthropic CEO Dario Amodei in a closed-door meeting Tuesday morning at the Pentagon, amid escalating tensions over the military’s use of advanced artificial intelligence systems.

According to Axios, senior defense officials have characterized the meeting as an ultimatum rather than a routine discussion, signaling mounting frustration within the Department of Defense.

Pentagon Frustration Over AI Restrictions

At the center of the dispute is Claude, Anthropic’s flagship AI model and currently the only system integrated into the military’s classified networks. Defense officials argue that Anthropic’s safeguards are too restrictive and have demanded that AI labs make their systems available for “all lawful uses.”

Anthropic has indicated willingness to ease some restrictions but insists on maintaining two red lines: preventing mass surveillance of Americans and blocking the development of weapons systems capable of firing without human involvement.

A company spokesperson said discussions remain “productive” and conducted in good faith, emphasizing Anthropic’s commitment to supporting U.S. national security.

Threat of “Supply Chain Risk” Designation

The Pentagon has reportedly warned it could designate Anthropic a “supply chain risk.” Such a move would not only terminate existing defense contracts but also require other Pentagon contractors to certify they are not using Claude in related workflows.

Officials have described the potential offboarding of Anthropic as a massive operational challenge, given the company’s deep integration into classified systems and the absence of equally capable alternatives.

Ideological and Cultural Divide

The standoff reflects deeper philosophical differences over AI’s role in national security. Amodei has publicly positioned Anthropic as a safety-first AI developer, frequently warning of the dangers posed by unregulated artificial intelligence.

Defense officials, however, view some of the company’s positions as ideological. One senior official told Axios that the disagreement extends beyond technical terms to fundamental views about AI’s use in warfare and intelligence.

The tension intensified earlier this year following the reported use of Claude during a U.S. operation targeting Venezuelan President Nicolás Maduro, further straining relations between the Pentagon and Anthropic.

High-Stakes Meeting Ahead

Leading the Pentagon delegation will be Hegseth, Deputy Secretary Steve Feinberg and Under Secretary for Research and Engineering Emil Michael, who has overseen negotiations with Anthropic and other AI firms. Anthropic declined to disclose the composition of its delegation.

The outcome of Tuesday’s meeting could shape not only the Pentagon’s AI strategy but also set a precedent for how private technology companies navigate the balance between ethical safeguards and national security demands.