The phrase that stuck was "unplugged overnight." When Anthropic disabled Fable 5 and Mythos 5 on Friday to comply with a US export-control order barring foreign nationals, the practical disruption was real but contained; the company's other models kept running. The political reaction was something else entirely. Within hours, ministers and opposition leaders across Europe and India were treating the shutdown not as a product outage but as a live demonstration of how exposed they are.
France led the chorus. Bruno Retailleau, the former interior minister now running for president, wrote that Washington's decision "should serve as a wake-up call," because "a nation that depends on others for its technology is a nation that can be unplugged overnight." He pointed to France's own assets, decarbonised nuclear power and companies such as Mistral, OVHcloud and Scaleway, and called for an end to "naivety." Benjamin Haddad, the minister for Europe, described the move as "an accelerator of the geopolitical battle over AI." Former prime minister Edouard Philippe put it most starkly: AI is now "critical infrastructure, as essential as electricity," and infrastructure that others can unplug is not infrastructure you control.
Britain's response carried a note of grief. Al Carns, who resigned as armed forces minister this week, wrote that "the most advanced AI model on the planet got switched off by a foreign government" while "British researchers were studying it, British companies were testing it, British hospitals were piloting it. Not any more." His conclusion ran wider than technology: "This isn't an AI story. It's the story of every industry we used to lead." Even Geert Wilders in the Netherlands joined in, half-joking that he wanted his "Claude Fable 5 back" while arguing that AI is increasingly a question of national sovereignty.
India had a sharper reason to worry. Barely two weeks ago it became the only major non-aligned power admitted to Project Glasswing, Anthropic's tightly controlled circle of fifteen countries given early access to Mythos for cyber defence. That access now looks conditional. Zoho founder Sridhar Vembu called the episode proof that "technology is the ultimate weapon" and declared "globalization is dead," urging Indian organisations toward indigenous and open-source models, including Chinese ones. The irony was not lost on anyone: Anthropic's own chief technology officer, Rahul Patil, is an Indian national, and the order's language could in principle lock foreign employees out of the very systems they helped build.
The backlash arrived alongside, and amplified, a homegrown alternative. On the same day, the Indian startup Avataar.ai launched Varya, an indigenous video model built with subsidised national compute under the government's IndiaAI Mission. Its pitch was efficiency rather than scale: video generated at 0.48 rupees per second, which the company claims is up to ten times cheaper than leading global models. "For a country of 1.4 billion people, affordability is not a feature, it is a prerequisite," said chief executive Sravanth Aluru. The framing, sovereign capability built cheaply at home, suddenly read less like marketing and more like a hedge.
Markets found their own way to express the anxiety. Bittensor's TAO token jumped roughly 18 percent as traders rotated toward decentralised AI networks, which analysts pitched as harder for any single government to switch off. That is a speculative reading of a serious shift, but it points in the same direction as the politicians. The lesson of Friday, repeated in Paris, London, New Delhi and on the crypto exchanges, is that frontier AI has quietly become something a government can grant or revoke. The open question is whether Europe and India can build credible alternatives faster than that dependence deepens.