A chip is not made in one town. It is made in a relay of cities, each owning a single layer of the industry: the foundry that prints the wafer, the fabs that make the memory, the plant that packages and tests the die, the makers of next-generation power devices, and the toolmakers behind them all. Here are six — what each one makes, why it is there, and why this particular map has become the most fought-over in global technology.
Unlike almost everything else in this atlas, a chip is not made in one place. It is made in layers, and each layer has its own city. A design becomes a wafer in a foundry; that wafer becomes memory in a different kind of fab; the finished die is packaged and tested somewhere else again, on tools built somewhere else still. So China’s chip map is not a single cluster but a relay: foundry and design in Shanghai, NAND in Wuhan, DRAM in Hefei, packaging in Wuxi, compound power devices in Xiamen, and the equipment in Beijing.
What concentrates them is not terroir, or even the supplier-next-door density that builds an appliance town. It is state capital and talent. Since 2014 the national “Big Fund” has poured well over a hundred billion dollars into the industry across three rounds, and cities like Hefei built their fortunes on aggressive local-government chip bets. China is the world’s largest chip market — more than half of global consumption — yet has long had to import most of its advanced silicon, for years spending more on imported chips than on oil.
That gap is now the most contested line in world technology. From 2022 the United States and its allies cut China off from the most advanced tools — above all the EUV lithography machines that only the Dutch firm ASML makes, the single chokepoint of the entire industry. China’s answer has been to build a domestic full stack at speed: SMIC reached “7nm” without EUV and made Huawei’s comeback chip, Hua Hong became a second advanced foundry, YMTC and CXMT pushed into memory, and NAURA and its peers raced to replace the banned machines. The controls were meant to slow China down; by many accounts they instead triggered an investment surge that would not otherwise have happened. The old self-sufficiency target of 70% was missed — the real figure is nearer 30% — but the map keeps filling in.