Trump Anthropic national security threat
“Well, not now, but a week ago, maybe.” That is how President Trump answered Axios correspondent Marc Caputo when asked, in an interview aired June 19, whether he viewed Anthropic or its CEO Dario Amodei as a threat to national security. The softer tone arrived two days after Trump met Amodei in person at the G7 summit lunch we covered last week, the same lunch where Amodei and Google DeepMind’s Demis Hassabis pitched a US-led coalition to govern global AI access.
The interview also surfaced a detail that changes how the original story should be read.
| VERDICT: The public tone has warmed considerably. The actual restrictions have not moved at all. Trump described Amodei as ‘nice’ and ‘smart’ and said the company responded responsibly to the government’s request. But the Commerce Department’s June 12 order requiring federal approval before foreign nationals can access Anthropic’s Mythos 5 and Fable 5 models remains fully in force, and the Pentagon’s separate March 3 supply-chain risk designation, barring federal agencies from using Anthropic technology at all, has not been rescinded either. Softer words and unchanged policy are both true at the same time. |
What Trump Actually Said
Asked directly whether Anthropic’s models posed a security risk, Trump told Caputo the concern had passed: he no longer viewed the company or Amodei as a threat. He described walking away from the G7 lunch with the impression that Amodei was personable and sharp, and praised how the company had handled the government’s original request, saying the company had behaved very responsibly.
On the seriousness of the original concern, Trump was blunt. “People get put in prison immediately for that,” he said, describing the underlying issue as a tremendous liability that could not be played with. He did not elaborate on the specific technical vulnerability, consistent with Anthropic’s own public position that it was not given a detailed account of the threat the government was responding to.
The New Detail: Who Actually Reported It
This is the part that updates the story meaningfully. Trump told Axios that the original alarm came from what he described as a competitor and part owner of Anthropic, someone who did not like what the company was doing and brought the concern to the administration directly. Axios reports that detail matches Amazon, which has invested 8 billion dollars in Anthropic and simultaneously operates competing AI models through its own AWS Bedrock platform, a direct commercial rival sitting inside Anthropic’s own capitalisation table.
According to Axios, Amazon identified a vulnerability in the Mythos model and brought it to the administration, which then approached Anthropic’s leadership. Trump’s characterisation was direct: it was a competitor and a part owner that turned Anthropic in, and that party was, in his telling, very concerned about what the company was doing.
This reframes the original export control order in a way the initial coverage could not, since at the time neither Anthropic nor the government publicly identified who raised the alarm. A company that is both a major investor in and a direct competitor to Anthropic flagging the concern that triggered a federal export control order is a materially different story than an unspecified internal government finding, even though the technical substance of the vulnerability itself may be unchanged.
What Has Not Changed, and Why That Matters More Than the Warmer Tone
Despite the friendlier public exchange, the actual mechanics of the restriction are untouched. The Commerce Department’s June 12 directive suspending foreign access to Mythos 5 and Fable 5 remains active. Separately and independently, the Pentagon’s March 3 supply-chain risk designation, which bars federal agencies from using any Anthropic technology, has also not been lifted, and the Pentagon’s 180-day timeline to fully remove Claude models from Department of War systems continues to move forward despite a federal judge’s temporary restraining order issued in late March.
Trump was also asked whether he would use emergency powers under the Defense Production Act if AI labs did not, in his words, get in line. His answer left the door open without committing to it: he said he has the power to use a lot of things, but is not sure he has to use it.
Why This Matters for the UAE Specifically
We made the case in our companion piece that the UAE’s entire AI strategy, G42’s Microsoft partnership, Stargate UAE’s reliance on Nvidia chips and US partners, is built on the assumption that access to American AI infrastructure remains stable. This week’s interview is a useful, concrete test of that assumption, and the honest reading cuts against the reassuring headline.
The friendlier rhetoric between Trump and Amodei is real, and it may well smooth the relationship between Anthropic and Washington going forward. But the underlying mechanism, a federal export control order that can suspend access to advanced AI models with days of notice, triggered in this case by a rival company’s report rather than a transparent government process, remains exactly as available to use again as it was three weeks ago. Nothing about this week’s warmer exchange removes that lever from the table. For any country or company whose strategy depends on continuous access to that infrastructure, the lesson from three weeks ago still holds regardless of how cordial the lunch was.
Anthropic’s Response
Anthropic issued a brief statement following the interview, expressing gratitude for the administration’s ongoing partnership in resolving the matter and reaffirming its commitment to working with the government on protecting critical infrastructure and maintaining US leadership in AI. The company did not comment on the Amazon detail specifically.
Sources
• Axios: Exclusive, Trump tells The Axios Show that Anthropic was a national security threat — https://www.axios.com/2026/06/19/trump-anthropic-national-security-the-axios-show
• Axios: Read the transcript of Trump’s interview with The Axios Show — https://www.axios.com/2026/06/19/trump-axios-show-interview-transcript-marc-caputo
• Cryptopolitan: Trump says Anthropic is no longer a national security threat after G7 lunch with Amodei — https://www.cryptopolitan.com/trump-says-anthropic-is-no-longer-a-national-security-threat/
• WION: ‘Not now, but a week ago, maybe’ — Trump says Anthropic not a national security threat — https://www.wionews.com/world/trump-anthropic-dario-amodei-not-national-security-threat-1781701239194
• The Epoch Times: Trump says he no longer views Anthropic as national security threat — https://www.theepochtimes.com/tech/trump-says-he-no-longer-views-anthropic-as-national-security-threat-6050558
Robius.news — Dubai, UAE — 2026 | Built to be first. Built to be trusted.
