Is the strongest model about to be unsealed? Trump says after G7 talks, he no longer considers Anthropic a threat

According to Beating Monitoring, U.S. President Donald Trump said in an Axios interview that he no longer regards the artificial intelligence company Anthropic as a national security threat. Just a week earlier, he had characterized Anthropic as a potential threat. The reversal came from a discussion during the G7 summit luncheon, when Trump met with Dario Amodei, Anthropic’s CEO, who had been seated out of favor.

Earlier, over the course of a week, the two sides had disagreed over the severity of the “jailbreak” security vulnerabilities in the Fable 5 and Mythos 5 models. After the negotiations fell apart, on June 12 the U.S. Department of Commerce issued an export control order, intercepting the exports of the Claude Fable 5 and Claude Mythos 5 models and requiring that foreign technical personnel access be approved. This left Amodei facing an awkward situation at the G7 luncheon, with his seat placed at the far end of the long table.

After the crisis broke out, the talks shifted to jointly setting technical standards. The White House, the Department of Commerce, the National Cyber Director, and Anthropic’s senior leadership held multiple rounds of communication, and dispatched security experts to Washington to build a technical benchmark framework for assessing the severity of vulnerabilities in large-scale models and the boundaries of government intervention. Trump said that Amodei responded quickly to the remediation requirements.

The shift in attitude directly removed policy obstacles to Anthropic’s plans to go public and expand. As of June 2026, Anthropic’s annualized recurring revenue had reached $47.0 billion, its valuation was $965.0 billion, and it had filed for a listing application. If it continued to be positioned as a national security threat, Anthropic’s push to go public would face substantial policy uncertainty.

On its regulatory stance, Trump reiterated that he does not want to restrict the domestic AI industry by shutting down or taking over companies, and that the top priority is maintaining a competitive edge against China. However, coercive regulatory options using emergency executive powers are not ruled out. Trump said that it is not yet certain whether the Defense Production Act needs to be initiated, but if necessary, he would still exercise that power.