A Friday afternoon that upended a launch
While much of the United States spent the weekend celebrating the country's first-ever World Cup win and the New York Knicks' championship, one of the world's leading AI companies was locked in a very different kind of fight. According to a report from The Verge, Anthropic received a US export control directive at 5:21 PM on a Friday ordering it to suspend access to its latest models — Mythos 5 and Fable 5 — for "any foreign national" inside or outside the country.
The wording was unusually broad. It applied not just to overseas users but reportedly to "foreign national Anthropic employees" as well, meaning the company's own staff could be barred from touching the very products they had just helped ship.
Why the company hit the kill switch
Faced with a directive that drew almost no distinction between customers, partners, and employees, Anthropic determined there was only one realistic way to comply: pull the plug entirely. The company moved to completely disable the two models — the same models it had spent the prior week hyping to the public and press.
For an industry built on momentum and launch-week buzz, going dark on a flagship release is close to a worst-case scenario. Models that had been positioned as the company's most capable yet were suddenly unavailable, and the rollout it had carefully staged evaporated in a matter of hours.
A trip to Washington
Rather than simply absorb the order, Anthropic took its case directly to the source. Company representatives travelled to Washington, DC, hoping to change President Donald Trump's mind and find a path that would let the models come back online without running afoul of the new controls.
The episode underscores how tangled AI and national security policy have become. Export controls have traditionally targeted physical goods — advanced chips, manufacturing equipment, military hardware. Applying them to software models that can be accessed from anywhere with an internet connection raises thorny questions about enforcement, and about whether a company can realistically wall off a product from every foreign national on earth, including its own workers.
The bigger picture
The standoff lands at a moment when governments around the world are scrambling to figure out how to regulate frontier AI. The US has leaned heavily on export controls as a lever to slow rivals' access to cutting-edge technology, but those tools were designed for an era of tangible exports, not cloud-delivered models that update in real time.
For now, the fight over Mythos 5 and Fable 5 is a vivid example of how quickly a celebrated product launch can collide with geopolitics — and how even the biggest AI labs remain at the mercy of a single government directive issued late on a Friday afternoon.
Source: The Verge — "Inside the fight over Claude Mythos 5."
