Anthropic export ban on Fable 5 sparks global AI sovereignty push

Anthropic being forced to pull Fable 5 and Mythos 5 offline just three days after launch is the clearest demonstration yet of how national security fears can override commercial AI roadmaps. The US export‑control directive didn’t target chips or data centres—it targeted model access itself, and in practice it was so sweeping that Anthropic disabled the models for everyone, including paying US customers and its own foreign‑national staff. That’s a template every frontier lab now has to price into its deployment strategy.

The downstream reaction, particularly in India, is telling. Commentators and founders are openly arguing that this is proof countries cannot rely on rented foreign models for critical infrastructure, and that sovereign models and domestic compute must move from rhetoric to funded programs. Similar debates are starting in Europe and parts of Asia. In the short term, this could slow diffusion of the most capable systems outside the US; in the medium term, it’s likely to accelerate parallel AGI efforts in other jurisdictions that don’t want their access switch sitting in Washington. For Anthropic, already labelled a “supply‑chain risk” by the Pentagon, the episode may complicate its planned IPO and cement a perception that it sits at the uncomfortable intersection of national security, civil liberties, and frontier AI.