The debate over the military use of Claude AI is intensifying after US Defence Secretary Pete Hegseth reportedly summoned Anthropic CEO Dario Amodei to the Pentagon for high-stakes discussions.

According to a report by Axios, the meeting scheduled for Tuesday is expected to involve tough talks about how Anthropic’s AI models — particularly Claude — can be used by the US military. The report, published Monday and citing unnamed sources, suggests tensions between the Pentagon and the AI company are reaching a critical point.

The development highlights growing friction between government agencies seeking expanded AI capabilities and technology companies aiming to maintain safeguards around their systems.

Pentagon Push for Expanded AI Access

A separate exclusive report by Reuters earlier this month revealed that the Pentagon has been urging major AI companies, including OpenAI and Anthropic, to deploy their AI tools on classified networks. The proposal reportedly involves loosening many of the standard restrictions typically applied to users.

The push reflects broader defence efforts to integrate advanced AI into national security operations. However, the military use of Claude AI remains a point of contention, as Anthropic has insisted on maintaining limits on how its models are deployed.

Talks Near Breaking Point

In another recent report, Axios noted that the Pentagon had been considering cutting ties with Anthropic due to disagreements over usage restrictions tied to Claude AI. Defence officials reportedly believe negotiations are now on the verge of collapse.

A senior defence official told Axios the meeting is not a routine introduction, indicating the urgency surrounding the dispute over the military use of Claude AI.

Despite the tension, an Anthropic spokesperson said the company is engaged in “productive conversations, in good faith,” according to the Axios report.

Verification and Official Response

Reuters said it could not independently verify the Axios report. At the time of reporting, the Pentagon, the White House, and Anthropic had not immediately responded to Reuters’ requests for comment.

The outcome of the discussions could shape how AI companies collaborate with defence agencies and determine the future direction of the military use of Claude AI within classified environments.