U.S. President Donald Trump has softened his position on Anthropic after previously viewing the artificial intelligence company as a possible national security threat, signaling a shift in tone as the White House and the AI lab work through a dispute over access to advanced models.
Trump told Axios he might have viewed Anthropic as a national security threat last week but no longer does. In the original interview, Axios reported that Trump said “not now, but a week ago, maybe” when asked whether he viewed Anthropic or CEO Dario Amodei as a national security threat.
Dispute Centered on Foreign Access to AI Models
The issue stems from the Trump administration’s concern over foreign access to Anthropic’s most advanced artificial intelligence systems. Reuters reported that senior Anthropic technical staff were scheduled to meet Trump administration officials earlier this week over a dispute involving foreign access to the company’s Fable 5 and Mythos 5 models.
Anthropic disabled access to those models for all users after Trump ordered the company to block foreign nationals from accessing them.
Axios reported that the Trump administration restricted access for countries outside the United States and for foreign nationals inside the United States.
That move placed Anthropic at the center of a larger debate over how governments should control frontier AI models that may have national security, cybersecurity, and geopolitical implications.
Trump Says Amodei Responded Responsibly
Trump’s latest remarks suggest that the confrontation may be easing.
Reuters reported that Trump told Axios that Amodei responded to the administration’s export control directive very quickly and responsibly.
Trump said he left the G7 summit with the impression that Amodei was nice and smart. Trump said Amodei responded quickly because the issue carried tremendous liability.
The shift matters because Anthropic has been one of the most closely watched companies in the race to build advanced AI systems. A dispute with the White House could affect not only the company’s model access policies but also how the U.S. government handles other frontier AI developers.
Emergency Powers Still Remain Possible
Even with the softer tone, Trump did not fully close the door on stronger government action. Trump did not rule out using emergency powers under the Defense Production Act against Anthropic.
Trump said he had the power to use a lot of things under the Defense Production Act but was not sure he needed to do so. That statement leaves open the possibility that the administration could still use stronger legal tools if it believes an AI company is not complying with national security expectations.
Amazon’s Role Adds Another Layer
The dispute also appears to involve concerns raised outside the government. A report from Amazon detailing a vulnerability alarmed the administration and was brought to Anthropic leadership.
Trump said a competitor and part owner had turned Anthropic in because of concerns about what the company was doing. Trump and other G7 leaders had met with technology executives, including Amodei, at the summit in France.
Anthropic Says It Wants Resolution
Anthropic framed the situation as a matter of continued cooperation with the administration. An Anthropic spokesperson said the company was grateful for the administration’s ongoing partnership in working to resolve the matter quickly.
Anthropic’s statement that the company remains committed to working with the administration on protecting critical infrastructure and ensuring U.S. leadership in AI.
The episode shows how quickly advanced AI companies can become entangled in national security policy. Trump’s latest remarks suggest relations with Anthropic may be improving, but the broader question remains unresolved: how far the U.S. government is willing to go to control who can access the most powerful AI models.
