- Anthropic CEO Dario Amodei met with White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent—after the Pentagon blacklisted the company as a ‘supply chain risk.’
- A federal judge called the ban an ‘attempt to cripple Anthropic,’ and the company filed its own political action committee to fight back.
- Trump told reporters he had ‘no idea’ the meeting was happening, even as both sides labeled it ‘productive and constructive.’
Dario Amodei walked into the West Wing on Friday morning to sit down with the two people who could decide whether his company survives its fight with the federal government. White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent met with the Anthropic CEO for what both sides later described as a “productive and constructive” discussion, the New York Times reported.
The meeting comes three weeks after a federal judge temporarily blocked the Pentagon from labeling Anthropic a “supply chain risk”—a designation that would cancel the company’s government contracts and those of its contractors. “It looks like an attempt to cripple Anthropic,” said Judge Rita Lin of the Northern California district court. It was the first time a US company had ever received such a designation.
The conflict centers on Anthropic’s refusal to let its AI be deployed for weapons systems without human oversight or for mass domestic surveillance. The Defense Department argued that Anthropic’s restrictions would undercut its “ability to control its own lawful operations.” An appeals court later reversed Judge Lin’s temporary block, leaving Anthropic blacklisted while the case proceeds. The company has since filed its own political action committee to push back on the administration’s stance.
Why Anthropic’s Pentagon Standoff Matters Beyond Courtrooms
The White House meeting signals something bigger than a legal dispute. It’s the first concrete sign that the Trump administration is willing to negotiate with an AI company it spent months trying to punish. Axios first reported that the meeting was scheduled, framing it as a thaw in relations that had grown bitter enough to land in federal court.
Meanwhile, parts of the federal government haven’t waited for the legal dust to settle. The Cybersecurity and Infrastructure Security Agency, operating under the Department of Homeland Security, along with segments of the intelligence community, have already been testing Anthropic’s Mythos model—the very AI whose cybersecurity capabilities sparked the Pentagon’s fears in the first place. Breitbart reported that the model’s ability to breach cybersecurity defenses has some experts calling it potentially dangerous.
That’s the irony sitting at the center of this whole standoff. The government banned Anthropic from its contracts because the company refused unrestricted deployment—but agencies are using the technology anyway because nothing else on the market does what Mythos does. The Atlantic Council called the situation “a larger crisis of trust over AI,” which is a polite way of saying the government can’t decide whether to fear Anthropic’s technology or buy it.
Trump, for his part, told reporters he had “no idea” the meeting was happening. Whether that’s strategic distance or genuine surprise is anyone’s guess—but it leaves Wiles and Bessent as the actual power brokers in whatever deal emerges.
The Pentagon dispute is just one front. The Fed and Treasury summoned Wall Street’s biggest CEOs earlier this month specifically to discuss Mythos and its implications for financial sector cybersecurity. Anthropic’s model has become the object of simultaneous fascination and fear across every branch of government—a company that 18 months ago was best known for being the cautious, safety-focused AI lab.
The appeals court case continues. Anthropic’s PAC is spending. And the White House just opened a door that the Pentagon tried to bolt shut.