President Trump told CNBC on Tuesday that a deal allowing the Department of Defense to use Anthropic's models is 'possible,' reversing tone after months of public hostility. 'They came to the White House a few days ago, and we had some very good talks with them, and I think they're shaping up,' Trump said on Squawk Box. The thaw follows a Friday meeting between Anthropic CEO Dario Amodei, chief of staff Susie Wiles, and Treasury Secretary Scott Bessent that a White House spokesperson called 'productive and constructive.'
The context matters. In March, the Pentagon labeled Anthropic a supply chain risk, a designation that forces defense contractors to certify they do not use Claude in military work. Trump followed with a Truth Social post ordering all federal agencies to 'IMMEDIATELY CEASE' use of Anthropic technology and vowing never to do business with the company again.
Anthropic sued the administration in San Francisco and Washington, D.C. to reverse the blacklisting, and a federal judge temporarily blocked Trump's directive. The Pentagon has continued using Claude during the war with Iran in the meantime, an awkward fact given the public stance.
Key facts
- 01Anthropic. A key thread of reporting in this story.
- 02Trump. A key thread of reporting in this story.
- 03Department of Defense. A key thread of reporting in this story.
The original dispute was substantive, not political. Anthropic signed a $200 million Pentagon contract in July, but September negotiations over deployment on the DOD's GenAI.mil platform broke down over scope. The Pentagon wanted unfettered access across all lawful purposes. Anthropic wanted written assurance that its models would not be used for fully autonomous weapons or domestic mass surveillance.
“Anthropic wanted assurance its models would not be used for fully autonomous weapons or domestic mass surveillance. The DOD wanted unfettered access for all lawful purposes. Talks collapsed.”— Jaeden Schafer
What changed is Mythos, Anthropic's new model announced earlier this month and limited to a small group of companies due to its cybersecurity capabilities. Amodei joined an early-April call on AI cyber readiness with Vice President JD Vance, Bessent, and other tech CEOs, and the Mythos pitch appears to have reopened the door. The company said it has 'ongoing discussions' with U.S. government officials about the model.
The subtext is leverage. A frontier model with advanced offensive cyber capabilities is exactly the kind of thing the Pentagon cannot afford to be locked out of, and Anthropic knows it. The red lines on autonomous weapons and surveillance have not obviously moved, which means either the DOD is softening its demands or Anthropic is quietly adjusting what it will sign.
Either way, the framing has flipped from national security risk to potential partner in under two months. That speed tells you how much of the original blacklist was policy and how much was posture.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
