President Donald Trump told Axios that he no longer considers AI company Anthropic a national security threat — a dramatic reversal from just a week earlier. The turning point came during a G7 summit lunch where Trump had a face-to-face talk with Anthropic CEO Dario Amodei, who had been seated at the far end of the table.
The two sides had been sparring over the severity of "jailbreak" security flaws in the Fable 5 and Mythos 5 models. After talks broke down, the Commerce Department imposed export controls on June 12, blocking shipments of Claude Fable 5 and Claude Mythos 5 and requiring approval for foreign tech workers to access them. That left Amodei in an awkward spot at the G7 lunch.
Once the crisis hit, negotiations shifted toward joint development of technical standards. The White House, Commerce Department, national cyber director, and Anthropic executives held multiple rounds of talks. Safety experts were dispatched to Washington to build a framework for assessing the severity of large model vulnerabilities and the boundaries of government intervention. Trump said Amodei responded quickly to the demands.
The reversal clears a major policy roadblock for Anthropic's IPO and expansion. As of June 2026, the company had $47 billion in annualized recurring revenue, a $965 billion valuation, and had filed to go public. Being tagged as a security threat would have created huge uncertainty for those plans.
On regulatory philosophy, Trump reiterated that he doesn't want to hobble the domestic AI industry by shutting down or taking over companies; his top priority is maintaining an edge over China. But he didn't rule out using emergency executive powers. Trump said he's not sure yet if he'll need to invoke the Defense Production Act, but he'd use it if necessary.