Four Survival Strategies, One Bubble
The fit is the trap.
Last week I argued the four AI strategies fit together by accident. I was right about the fit. I was wrong about what it means.
Puzzles are static. What I described wasn't a puzzle. It was a feedback loop.
London's Nobel proves AI solves real problems — which justifies the capital flowing into San Francisco. SF's models make Chinese factory automation work — which undercuts European manufacturing on cost — which forces Paris to write rules — which demands more research from London to define what "trustworthy AI" even means.
Each node validates the next. None of it stands alone.
And none of the countries can stop. The US needs AI to deliver the productivity numbers its service economy has already priced in. China's demographic cliff turns automation into survival, not strategy. The UK has one frontier-grade asset — DeepMind — and stepping back means conceding the last lane where British inputs still matter. Europe's regulatory play only works if the rest of the world keeps building things worth regulating.
Stop participating and the whole logic collapses.
The money follows. OpenAI just raised $122 billion at an $852 billion valuation. Stargate commits $500 billion for data centers by 2029. Europe pledged €200 billion to InvestAI. China's semiconductor fund sits at $47.5 billion. Nearly a trillion dollars committed before the returns materialize.
That's not faith in a product. It's faith in a structure where everyone assumes everyone else can't afford to stop.
Each country thinks it's hedged. It isn't. If US deployment stalls, UK research loses its downstream buyer. If China can't automate, factory AI has no market. If Europe's regulation collapses, Mistral loses the category it competes in. If London stops producing breakthroughs, SF's models lose their scientific floor.
Interdependence isn't resilience. It's shared fate.
One axis breaking doesn't leave three standing. Four survival strategies. One bubble — not because anyone mispriced a model, but because the structure required it.
Those are the ones that always pop.
