The Fable 5 export control order that Commerce Secretary Howard Lutnick issued on June 13, forcing Anthropic to pull two of its newest models from global access, began almost accidentally. Amazon researchers stress-testing Fable 5 found a jailbreak sequence that they documented as enabling access to cyberattack-relevant information. Amazon CEO Andy Jassy was already on a pre-scheduled call with White House officials on June 11 about a separate matter. He raised the vulnerability. White House officials encouraged him to flag it directly to Treasury Secretary Scott Bessent. Jassy did. Bessent, who had been managing the administration’s response to Anthropic’s Mythos model separately, moved quickly. By Friday, Anthropic received a directive to either fix the jailbreak or remove the models. Since American “deemed export” rules prevent any foreign national, including Anthropic’s own employees, from accessing the restricted models, the company concluded it had no option but to pull them entirely.
The chain of events, from Amazon’s test to Jassy’s call to the export control order, is what NewsTrackerToday traces as the mechanism that raises the most analytical questions. Amazon is both a major investor in Anthropic, with up to $20 billion committed, and the company whose CEO triggered the regulatory action that restricts Anthropic’s flagship products. Amazon stated publicly that Jassy raised the concerns as a national security matter and also flagged worries about AI capability across multiple frontier labs, not just Anthropic. More than 150 cybersecurity experts signed an open letter asking the president to revoke the order, saying the jailbreak Fable exhibited was not unique to Anthropic’s models – that similar vulnerabilities exist in other publicly available systems including OpenAI’s GPT-5.5 – and that pulling Fable and Mythos removed essential cybersecurity defense tools from U.S. network defenders.
Daniel Wu places the regulatory action in historical context: “The U.S. has used export controls as a geopolitical tool against technology companies since the semiconductor wars of the 1980s. What is different here is the target: a domestic frontier AI company, not a foreign one. The stated rationale is national security. The circumstantial pattern – the company with the strained relationship with the administration, the CEO of its largest investor escalating the vulnerability, the NSA reportedly pushing for controls before Fable launched – creates an environment where the technical and political rationales are impossible to cleanly separate. That ambiguity is the environment every frontier AI company now operates in.”
The G7 summit in Evian-les-Bains, France, on June 18, produced one of the most visually telling moments in the story. Trump sat at the head of the table at a roundtable for business and government leaders. Sam Altman sat directly at his side. Dario Amodei sat across the room, next to French President Macron. Amodei reportedly told the group to “resist the temptation to splinter” in how they handle AI regulation. The administration’s approach – selectively applying export controls to one company whose CEO has publicly clashed with the White House – is the specific version of splintering Amodei was describing. That visual and that argument together are what NewsTrackerToday reads as the contrast between Anthropic’s diplomatic posture and its actual regulatory position.
The Department of Defense had already labeled Anthropic a “supply chain risk” in March, requiring defense contractors to certify they will not use Claude models in military work. The export control order in June extended that adversarial relationship into the commercial product space. Anthropic said in a blog post that it believes the government should have the ability to block unsafe deployments – but through a process that is transparent, fair, clear, and grounded in technical facts. The Friday directive that arrived without prior communication or published criteria was not that process, and the “supply chain risk” label that preceded it is what News Tracker Today holds as the designation that explains why Jassy’s vulnerability disclosure escalated into a policy action where the same disclosure about a different company might have generated a quiet advisory.
Anthropic and the government reportedly began working on Thursday, June 19, toward a framework for judging when future model jailbreaks pose a serious danger. A deal to restore Fable and Mythos access on modified terms appears to be in progress. Whether the regulatory precedent set by the export control order – a Commerce Secretary letter to a CEO, Friday afternoon, no published criteria, no appeals process – becomes the template for future AI enforcement or a one-off driven by specific relationship dynamics is the open question that Anthropic’s filing of an IPO prospectus makes urgently commercially relevant. Can a company with a near-trillion-dollar valuation go public while an unresolved government designation of “supply chain risk” sits on its record? That is what NewsTrackerToday closes on as the question the ongoing deal negotiation will resolve or complicate.
