Trump Says Anthropic Is Not a Security Threat: What It Means

Trump says Anthropic is no longer a national security threat, but nothing about your access to its models has formally changed yet.

In a pretaped Axios interview reported on June 19, the president was asked whether he still considers Anthropic a threat and replied, "Well, not now. But a week ago, maybe," adding that the company has "behaved very responsibly," according to The Next Web. That is a sharp tonal reversal, and it is also where the practical caveat starts.

Does the reversal restore access to Fable 5 and Mythos 5?

Not on its own. The warm words are a political signal, not a rescinded rule.

The two sources agree on the core fact: the Commerce Department's June 12 directive ordering Anthropic to seek US approval before foreign nationals access its most powerful models, Fable 5 and Mythos 5, has not been formally withdrawn. The Pentagon's supply-chain risk designation from March 2026 also remains in place. Trump said he would consider easing the restrictions but hedged, "I would, but I'm not sure I have to do that."

The catch is procedural. Rolling back a formal Commerce directive requires bureaucratic steps that a single interview cannot shortcut, and Commerce operates with considerable independence on export controls. So the most you can conclude today is that escalation has paused, not that the order is gone.

Where the two reports agree and where they diverge

Both outlets cover the same event but emphasize different stakes, which is the most useful way to read this story.

The conflict worth naming is on cause. The Next Web frames the dispute around Anthropic's refusal to strip surveillance and autonomous-weapons guardrails from military products. TechCrunch reports the proximate trigger as a jailbreak that Amazon researchers allegedly found in Fable 5, and notes Anthropic itself said the same jailbreaks exist in other models. Both can be true, but they point to different reasons the restriction might or might not lift.

Why the tone shifted now

The reversal followed a specific meeting, and the context around it explains the timing.

Trump met Anthropic CEO Dario Amodei on Wednesday at the G7 Summit in Evian-les-Bains, France, an encounter that appears to have moved his stance. At that summit, Amodei and Google DeepMind CEO Demis Hassabis jointly pitched a US-led AI coalition to G7 leaders, positioning Anthropic as a cooperative partner rather than a regulatory adversary.

The pressure had also been building from outside. The Next Web reports that Commerce Secretary Howard Lutnick had earlier sent a letter threatening criminal charges, a move that drew industry criticism and prompted allied governments, including the UK, to lobby for exemptions. Trump described the situation as creating "tremendous liability" for the administration, which reads less like a security reassessment and more like backlash management.

What this means for everyday Claude users and businesses

For most individual users, day-to-day access was never the thing in question. The Commerce order targets foreign-national access to the two most powerful models, not casual consumer use of the broader product line.

For businesses, the live risk is access continuity, not the headline. If your team includes foreign nationals or relies specifically on Fable 5 or Mythos 5, the directive still governs you until it is formally rescinded. A friendlier presidential tone does not change a compliance obligation.

This is also why a backup plan still earns its keep. During the original ban we argued the case for not depending on a single AI vendor in our breakdown of single-vendor AI risk, and the reversal does not retire that logic. The whole episode, from order to interview, played out in roughly a week, which is exactly the volatility a redundancy plan is meant to absorb.

Should you resume normal use or keep a backup?

Resume normal use where you were never restricted, and keep redundancy where the order still binds you.

• Individual consumer use of mainstream Claude products: continue as before, since the directive did not target it.
• Workloads tied specifically to Fable 5 or Mythos 5, or involving foreign-national access: treat the restriction as live until Commerce acts, and keep an alternative provider configured.
• IPO-sensitive or procurement decisions: note that de-escalation may stabilize investor confidence, but the formal constraints remain the binding reality.

The limitation to acknowledge is timing. Neither source reports a date for lifting the order, so any plan that assumes the restriction disappears soon is betting on a step that has not happened.

References:

• The Next Web, Trump says he no longer views Anthropic as a national security threat after G7 meeting with CEO. Accessed on Jun 20, 2026
• TechCrunch, Is the US government’s Anthropic ban accidentally helping the brand?. Accessed on Jun 20, 2026