Imagine a new country appears overnight.
50 million citizens. Each one smarter than any Nobel laureate. Capable of making ten moves for every one of yours.
That is the thought experiment Dario Amodei, CEO of Anthropic, posed last month. His conclusion is that a competent national security official would call it “the single most serious threat we’ve faced in a century, possibly ever.”
Now look at what’s happening quietly underneath all the US v China AI headlines.
→ Chile: LatAm-GPT → France: Mistral building in Sweden → UAE: Falcon → Saudi: ALLaM
Plus a dozen more.
▶️ We saw this up close last year in Sofia, where the conversation was already about data residency and linguistic independence (2 min video) ↘︎ https://lnkd.in/dxq_QGQn
None of these will beat GPT or Claude on raw benchmarks. That is not the point.
The point is that dependence on someone else’s intelligence is a national security problem. And enough capitals have decided they’d rather have a competent local model than a brilliant foreign one they don’t control.
This is the sovereignty layer of the AI race. It is quieter than the frontier lab wars, but it may matter more.
Because the real question isn’t who builds the smartest model. It’s who gets to decide what it thinks.
