Dario Amodei, the CEO of AI company Anthropic, walked into the White House on Friday for a high-stakes sit-down with Donald Trump‘s chief of staff. It was the first serious attempt at diplomacy after months of bruising conflict between the two sides.
Amodei met with White House Chief of Staff Susie Wiles, with Treasury Secretary Scott Bessent also joining the discussion. CNN confirmed the meeting, describing it as a sign that the administration is trying to find a way out of a feud that has since spilled into the courts. The talks centered on Anthropic’s powerful new AI model, Mythos, and what role it could play for the federal government.
“This is a big problem. Everyone’s complaining. There’s all this drama. So this got elevated to Susie to hear Dario out, determine what is bullsh-t and start to plot a way forward,” a Trump adviser told Axios.
How the feud started
The fallout began after Anthropic refused to let the Pentagon use its Claude AI for what the military described as “all lawful purposes,” a term that included fully autonomous weapons and mass surveillance of Americans. Defense Secretary Pete Hegseth responded by moving to declare Anthropic a “supply chain risk,” a designation previously reserved for companies tied to foreign adversaries. It would have effectively banned Anthropic from doing business with the government.
Trump then took to social media in February 2026, announcing the administration “will not do business with them again!” Anthropic pushed back hard, filing lawsuits in two federal courts. In March, US District Judge Rita Lin sided with Anthropic, blocking the enforcement of Trump’s social media directive against federal agencies outside the Defense Department. The government has since appealed.
The fight was made more complicated by a second ruling from the DC Circuit Court of Appeals, which declined to block the Pentagon’s ability to cut ties with Anthropic while that legal challenge plays out. The court said forcing the military to keep an “unwanted vendor” during an ongoing military conflict would not be appropriate.
As the House Democrats moved to challenge Trump’s authority through the 25th Amendment, the administration has also been facing pressure on multiple fronts over how it handles its relationships with institutions both domestic and abroad. Pope Leo XIV also recently fired back at Trump, saying he has “no fear of the White House.”
Mythos changes the calculus
The dynamic shifted when Anthropic unveiled Mythos, an AI model that experts have described as a potential watershed moment for cybersecurity. The tool can identify and exploit software vulnerabilities at a level that no previous model has matched, according to the Washington Times.
Anthropic co-founder and policy chief Jack Clark said the company is releasing Mythos only to a select group of major organizations. “We’re releasing it to a subset of some of the world’s most important companies and organizations so they can use this to find vulnerabilities,” Clark said. He also warned that more models with similar capabilities will follow, from both US and Chinese labs, within the next year or two.
That warning appears to have gotten through to some in the administration. “It would be grossly irresponsible for the U.S. government to deprive itself of the technological leaps that the new model presents,” a source close to the negotiations told Axios. “It would be a gift to China.”
The White House confirmed the broader engagement in a statement to CNN: “The White House continues to proactively engage across government and industry to protect the United States and Americans. This includes working with frontier AI labs to ensure their models help secure critical software vulnerabilities.”
The Office of Management and Budget has already signaled it is preparing to give federal agencies access to Mythos for security assessments, according to a Bloomberg report cited by CNN. Whether Friday’s meeting moves the two sides closer to formally settling their legal dispute remains to be seen.
