The Trump administration is keeping export controls in place on Anthropic’s Claude Fable 5 and Mythos models over concerns about potential “jailbreak” risks, leaving them offline since June 12. Anthropic is holding urgent discussions with the White House and the Commerce Department’s Bureau of Industry and Security on a technical verification path, but there is still no clear timetable for restoring access. The episode does not directly affect traditional financial asset prices, but it signals tighter oversight at the AI infrastructure layer and reinforces expectations of scarcity in high-end AI chips and compliant cloud services.

Meta has reassigned about 7,000 engineers into newly created AI-focused teams as it tries to accelerate its push in the AI race, sparking an internal backlash that included employees publicly insulting executives and comparing the new unit to a “Gulag.” CTO Andrew Bosworth acknowledged the reorganization was “atrocious,” but some of the fixes discussed—such as improving office snack stations—were mocked by staff as out of touch. The rollout of Meta’s Muse Spark model has been delayed and early user response has been muted, adding to concerns about heavy spending without clear payoff.

A series of AI-related incidents is putting renewed focus on regulatory and compliance risks across the tech sector. Meta was reported to have embedded face-recognition code for its smart-glasses platform in an unreleased version on millions of phones, while Microsoft faced its biggest-ever Patch Tuesday tied to AI bug hunting and Oracle was hit via a zero-day exploited by ShinyHunters. OpenAI, xAI and Anthropic have also moved toward IPOs, drawing added attention to safety disclosures and reviews, though no specific new rules or enforcement actions were cited. The developments point to higher longer-term compliance costs for listed companies rather than an immediate financial impact.

The Trump administration imposed export controls on Anthropic’s Claude Fable 5 over concerns the model could be jailbroken. Emergency talks between Anthropic and the White House did not produce a resolution. Amazon CEO Andy Jassy separately alerted the Treasury Department to the alleged risk, underscoring Amazon’s close ties to Anthropic as its largest investor and AWS cloud partner. The episode has fueled investor worries that tighter AI regulation could hinder cloud providers’ AI commercialization, though there is no evidence of a direct financial hit or broader policy expansion.

Meta Chief Technology Officer Andrew Bosworth said the company did an “atrocious” job rolling out its newly formed AI division, a reorganization that badly hit morale among roughly 6,500 engineers and product managers, with some employees likening the group to a “gulag.” Meta is seeking to rebuild trust by reshaping management, improving communication, offering career development support, and upgrading office perks. The episode underscores a major setback in executing the company’s AI strategy that could slow model iteration and commercialization.