The AI nationalization risk nobody is pricing

What happens to NVDA, MSFT, hyperscalers, and model companies if the US government decides frontier AI is too dangerous to remain fully private?
Jordi Visser raised that question on Full Signal, and it is the kind of tail risk the market is barely pricing.
The logic is simple: when a technology becomes powerful enough to be compared with nuclear weapons, governments usually do not leave control entirely in private hands.
You do not need to believe in 2027 AGI to update the risk model.
The trajectory alone is enough to matter.