Anthropic's forced shutdown sharpens India's sovereign-AI case
Washington made Anthropic pull Fable 5 and Mythos 5 offline for all foreign nationals. For India, the lesson about depending on borrowed frontier models is hard to ignore.
The News
Anthropic took its two most powerful artificial-intelligence systems offline over the weekend after the White House ordered the company to cut off access for every foreign national, including its own non-American staff. The models, Fable 5 and Mythos 5, had been generally available for barely three days when the instruction arrived.
According to reporting by The Verge, the directive reached Anthropic at 5:21 PM on Friday and demanded that both systems be suspended for any foreign national inside or outside the United States. Because the order also covered foreign-national employees, the company concluded that the only workable response was to pull the models down completely rather than try to fence off individual users.
Anthropic had launched the pair on 9 June. It described Fable 5 as a system whose "capabilities exceed those of any model we've ever made generally available," while Mythos 5 runs on the same underlying architecture with certain safeguards relaxed. The shutdown landed while the company was already locked in a separate dispute with the Pentagon, opening a second front in its troubles with the federal government.
Why It Matters
The episode is a blunt reminder that frontier AI is not just an American industry but, increasingly, an instrument of American statecraft. Washington has spent three years tightening its grip on advanced computing. When it restricted exports of Nvidia's most capable chips to China in October 2022, the lever was hardware. This time the lever is the model itself, and the reach extends inside the company's own walls to its foreign employees.
That shift collapses a comfortable assumption: that a commercially available model is a dependable utility, like electricity or cloud storage. It is not. Access can be revoked over a weekend by a government that is not your own, for reasons unrelated to your contract. For any business that has wired a US frontier model into its core workflow, that is a continuity risk hiding in plain sight.
The broader signal is geopolitical. The United States dominates frontier AI, and its government has now shown it will decide who may use the best of it. For rivals and allies alike, the case for building at home has rarely looked stronger.
Indian Angle
For India, the timing is pointed. New Delhi has staked a deliberate bet on sovereign capability through the IndiaAI Mission, the multi-thousand-crore programme that is funding shared GPU capacity and home-grown foundation models. Sarvam AI was picked to build an Indian foundational model under that effort, and Bhavish Aggarwal's Krutrim has pushed its own large models. An incident like this one is the clearest argument yet for why that money is being spent.
The foreign-national clause should also concentrate minds in Bengaluru and Hyderabad. Indian engineers form a large share of the research talent at American labs, and a rule that locks out non-American staff treats some of the field's best minds as a compliance problem. It strengthens the pitch for Indian researchers to build within Indian institutions instead.
There is a regulatory thread too. MeitY and the bodies weighing AI governance now have a live example of supply-side fragility to factor into procurement thinking. An Indian bank, insurer or government department that routes sensitive workloads through a single overseas model should ask what happens if that tap is turned off without notice.
FAQ
What exactly did the White House order?
The directive instructed Anthropic to suspend access to its Fable 5 and Mythos 5 models for any foreign national, inside or outside the United States, and explicitly included the company's foreign-national employees. Anthropic determined that full compliance effectively meant taking the models offline.
When did this happen?
The two models launched on 9 June. The order arrived at 5:21 PM on a Friday, and Anthropic took the systems down over the following weekend, days after they had become generally available.
How does this affect Indian developers and firms?
Any Indian business relying on these specific models would have lost access without warning. More broadly, it underlines the risk of building critical systems on a single foreign frontier model and strengthens the case for sovereign alternatives.
Where can I read the original report?
The Verge reported the details of the export-control order and Anthropic's response. The link to its full coverage is in the paragraph below.
This story was reported by The Verge. Read the full original coverage at The Verge.