Anthropic Pulls Top Models After Trump Export Order Hits All Users
A US order forced Anthropic to cut Fable 5 and Mythos 5 access for every foreign national, even inside America. The fallout reaches India's developer base.
The News
Anthropic spent much of this week scrambling to put its newest AI models back online after the Trump administration abruptly ordered the company to cut off access for all foreign nationals. The instruction was sweeping enough to cover users sitting inside the United States and Anthropic's own staff, not just people overseas.
Faced with that demand, the company took the blunt route and pulled access to its flagship models, Fable 5 and Mythos 5, for everyone while it worked out who it was actually allowed to serve. The administration has not publicly set out the legal basis for the order, leaving the company and its customers guessing at the scope and the duration.
According to reporting by The Verge, the episode appears to be the first occasion on which US export controls have been turned against access to a specific AI model in this fashion. Until now, those controls have largely governed hardware rather than the software running on it.
Why It Matters
Export controls in the technology sector have, for years, been a story about chips. The most-cited precedent is the October 2022 round of US restrictions that barred advanced Nvidia accelerators from reaching China, later tightened to close workarounds. Those measures treated compute as the choke point. Applying the same machinery to who may log in to a model marks a different philosophy, one in which the model itself, and the people querying it, become the controlled item.
That shift carries weight well beyond one company. If access to frontier models can be switched off by executive order at short notice, every enterprise that has wired its operations to a single US provider now carries a continuity risk it had probably not priced in. The fact that even American-based foreign nationals and Anthropic's own employees were swept up suggests the order was written broadly and the company chose to over-comply rather than risk breaching it.
Markets have spent two years treating frontier models as utilities you simply rent. This week is a reminder that the tap has an owner, and the owner answers to a government.
Indian Angle
Few countries have more at stake here than India. The country supplies one of the world's largest pools of software engineers, a meaningful share of whom are foreign nationals working at US labs on visas, and an even larger base of developers and startups building products on top of American model APIs. A rule that targets foreign nationals, even those physically in the US, lands squarely on Indian talent inside firms such as Anthropic.
For Indian startups, the lesson is about dependency. A fintech in Bengaluru or a healthtech in Pune that routes its core reasoning through a single foreign model has just watched that model vanish for an entire class of users overnight. That strengthens the commercial case for domestic alternatives such as Sarvam and Ola's Krutrim, and gives fresh urgency to the government's IndiaAI compute push.
It also hands MeitY and policymakers a concrete data point. Sovereign capability has often been argued in the abstract; this week offers a live example of why over-reliance on externally controlled models is a strategic exposure for Indian banks, regulators and public services, not just a procurement preference.
FAQ
Why did Anthropic block its own models?
The Trump administration ordered the company to cut access for all foreign nationals, a category broad enough to include US-based users and Anthropic staff. Rather than risk serving someone it was barred from serving, Anthropic suspended Fable 5 and Mythos 5 access for everyone while it clarified the rules.
Has this happened to an AI model before?
The Verge reports it appears to be the first time US export controls have been used to restrict access to a specific AI model in this way. Earlier controls focused on hardware, such as advanced chips, rather than on who may use a model.
What does it mean for Indian developers?
Indian engineers who are foreign nationals at US labs, and Indian startups building on American model APIs, face direct disruption and continuity risk. It strengthens the case for diversifying across providers and for backing domestic models.
Where can I read the original report?
The Verge's coverage sets out the timeline and the unanswered legal questions in full. The link is in the attribution note below.
This story was reported by The Verge. Read the full original coverage at The Verge.