Trump: Anthropic Talks Went Well; DOD Deal Possible

President Donald Trump said on Tuesday that talks with AI company Anthropic at the White House “were very good” and that a Department of Defense agreement for the firm’s models “is possible.” He said Anthropic representatives visited Washington recently and that he believes their technology “can be of great use.”

In March the Defense Department labeled Anthropic a supply chain risk, requiring contractors to certify they were not using the company’s Claude models in military work. The administration ordered agencies to immediately cease using Anthropic technology and avoid future contracts. Anthropic sued in federal court to overturn the designation, and a judge issued a temporary block on parts of the directive while legal challenges proceed.

The Pentagon maintained some operational use of Claude during the fall and winter. Anthropic signed a $200 million contract with the Defense Department in July, but negotiations over deploying Claude on the department’s GenAI.mil platform collapsed in September, contributing to the dispute.

Anthropic chief executive Dario Amodei met with senior administration officials last Friday to discuss Mythos, the company’s new model. White House chief of staff Susie Wiles and Treasury Secretary Scott Bessent attended the meeting, which a White House spokesperson described as “productive and constructive.” Amodei also joined an early April call with Bessent, Vice President J.D. Vance and other technology leaders on AI cyber readiness.

Anthropic limited Mythos to a small group of firms because of the model’s cybersecurity capabilities. The company plans to expand access to financial institutions in Europe and is including security checks as part of the onboarding process. People familiar with the plans said some banks could gain access within days or weeks.

Scott Keipper, Americas Financial Services Technology Consulting Leader at EY, said: “The speed of the technology is outrunning the governance, operating models, and control systems most banks were built to handle.” He urged banks to build AI into ongoing risk-management systems rather than rely on one-time cybersecurity fixes.

Anthropic says it is continuing discussions with U.S. government officials about Mythos and its limits. The meetings and the president’s comments follow months of legal and policy disputes between the company and the administration.