Despite being labeled a supply-chain risk by the Pentagon earlier this year, Anthropic's relationship with the broader Trump administration appears to be warming. CEO Dario Amodei recently met with Treasury Secretary Scott Bessent and White House Chief of Staff Susie Wiles in what the White House called a productive and constructive introductory meeting.
The Meeting at the White House
Axios reported Friday that Amodei sat down with senior administration officials to discuss collaboration on cybersecurity, America's lead in the AI race, and AI safety. The White House confirmed the meeting took place, describing it as focused on shared approaches and protocols for addressing challenges associated with scaling AI technology.
Anthropic issued its own statement confirming that Amodei met with senior officials and that the company is looking forward to continuing these discussions. The tone from both sides was notably cooperative — a sharp contrast to the hostile relationship between Anthropic and the Department of Defense.
The Pentagon Dispute
The thaw comes against the backdrop of a bitter fight between Anthropic and the Pentagon. The dispute originated when negotiations over military use of Anthropic's models broke down. Anthropic refused to allow its technology to be used for fully autonomous weapons and mass domestic surveillance — conditions the Pentagon was reportedly unwilling to accept.
In response, the Department of Defense declared Anthropic a supply-chain risk, a designation normally reserved for foreign adversaries. The label could severely limit the use of Anthropic's models across the entire federal government. Anthropic is currently challenging that designation in court.
OpenAI moved quickly to capitalize on the situation, announcing its own military deal — a move that triggered some consumer backlash, with Claude briefly rising to number two on the App Store as users signaled their preference for Anthropic's stance.
Mythos Changed the Conversation
The turning point appears to have been Mythos, Anthropic's latest model that the company deemed too powerful to release publicly due to its unprecedented cybersecurity capabilities. Instead of releasing Mythos broadly, Anthropic briefed the Trump administration directly and shared the model with select corporations including JPMorgan Chase and Goldman Sachs.
Anthropic co-founder Jack Clark confirmed at the Semafor World Economy summit that the company spoke to the administration about Mythos and plans to continue briefing them on future models. He characterized the Pentagon dispute as a narrow contracting disagreement that should not overshadow Anthropic's genuine commitment to national security.
The Mythos strategy — withholding a dangerous model from the public while proactively sharing it with the government — appears to have bought Anthropic significant goodwill with parts of the administration that are not aligned with the Pentagon's position.
Every Agency Except the Pentagon
Perhaps the most revealing detail from the Axios report was a quote from an administration source claiming that every agency except the Department of Defense wants to use Anthropic's technology. That framing suggests the Pentagon's hostile stance is an outlier within the administration, not a consensus position.
For Anthropic, that distinction matters enormously. The company has warned that the supply-chain risk designation could cost it billions in lost enterprise revenue, with over 100 enterprise customers reportedly expressing doubts about continuing their contracts. If the broader administration signals that working with Anthropic is acceptable despite the Pentagon's label, those customers may feel more comfortable staying.
What It Means Going Forward
The meeting between Amodei and senior White House officials does not resolve the Pentagon dispute, which will likely play out in court over the coming months. But it does signal that Anthropic is not being frozen out of Washington — far from it.
The company is simultaneously suing the government and briefing it on its most sensitive AI research. That balancing act — challenging the Pentagon in court while collaborating with the Treasury Department and White House — reflects the increasingly complex relationship between AI companies and the federal government.
For now, Anthropic appears to be winning the political game outside the Pentagon's walls. Whether that translates into a resolution of the supply-chain risk designation remains the biggest open question for the company's government business — and for the more than 100 enterprise customers watching closely from the sidelines.
