Different philosophies of the Sino-US Arms Race:

The old three T's of US-China relations — Tibet, Taiwan, Tiananmen — have been replaced. Taiwan remains, but the other two are now trade and tech. Technology has become the real driver of global fragmentation.

The US and China are running different AI races. America leads on the large language models — the expensive, data-hungry systems like Claude and ChatGPT. China leads on applied AI: robotics, drones, facial recognition, factory automation. As I put it: the US is better at the brains, China is better at the bodies.

America's strategy has been to trip its competitor rather than run faster — blocking ASML chip exports to slow China's development. Think of chips as oil: without access, your economy stalls. But DeepSeek proved the strategy has limits. China built a comparable model for a fraction of the cost, open source. That was AI's Sputnik moment — suddenly it was clear China wasn't as far behind as everyone assumed.

Taiwan's Silicon Shield — making itself so indispensable to global chip supply that invasion would be economic suicide — is also eroding. As the US builds its own fabs in Arizona, taking Taiwan becomes cheaper. And it's a Western fallacy to think Taiwan is optional for China. It sits so deep — emotionally, strategically, in terms of identity — that it simply isn't.

The bigger picture is a new triangular dynamic emerging between the US, China, and Europe. The transatlantic relationship is clearly splitting. Europe needs a stable external partner, and it won't necessarily be America. Perhaps the real advantage of being behind in the AI race is that Europe can watch the US experiment unfold and decide whether it actually wants to follow.