Balaji’s right that China dominates niche manufacturing through spatial concentration.

But the mechanism is the opposite of what policy makers assume.

Datang makes 35% of the world’s socks. Shengzhou makes 60% of the world’s neckties. Wenzhou once held 80% of the global lighter market. Qiaotou makes 60% of the world’s buttons. One town, one product, global domination.

But these clusters didn’t form because the Chinese government designated “lighter zones” or “sock cities.” They formed because poor farmers in mountainous Zhejiang province had almost no arable land, couldn’t farm their way out of poverty, and were too remote for state-owned enterprises to bother with.

In 1978, Wenzhou had less than half an acre of farmland per person. Locals had been peddling handicrafts for generations because there was no alternative. When Deng Xiaoping’s reforms hit, Wenzhou already had underground factories and shadow banking networks that exploded overnight. From 1980 to 1988, private industrial output in Wenzhou went from 1% to 41%.

The origin stories are almost comical. Qiaotou’s button empire reportedly started when three brothers spotted discarded buttons in a gutter and thought “there’s money here.” Shaodong’s lighter industry began when a couple bought 50 lighters, took them apart repeatedly, and taught themselves the manufacturing process. One town’s survival instinct became the world’s supply chain.

This tells you everything about why US “special economic zones” keep failing. Opportunity Zones. Enterprise Zones. Tax breaks in designated corridors. American policy makers keep offering incentives to attract capital to specific geographies.

China’s clusters weren’t built on incentives. They were built on constraint. No farmland meant you had to make things. No state support meant you competed on cost and speed. No capital meant you borrowed from your neighbors and ran family workshops until something worked.

Detroit’s drone zone initiative is interesting precisely because it’s building on an existing asset: automotive manufacturing know-how, supply chain infrastructure, skilled labor that’s been displaced by EV transitions. That’s closer to the Wenzhou model than dropping tax breaks on empty land.

The pattern that works: geographic constraint forces specialization, specialization creates knowledge spillovers, knowledge spillovers attract more specialists, the cluster compounds until it’s unbeatable on cost, speed, and expertise. Wenzhou’s lighter factories could go from design to production in 25 days when Japanese competitors needed 80.

What doesn’t work: picking winners from above, offering tax breaks without underlying capability, building zones disconnected from existing skills.

China’s manufacturing cities weren’t planned. They were accidents of geography, poverty, and survival instinct that crystallized into global monopolies. The US won’t replicate that by drawing new lines on maps.

The real opportunity is identifying existing clusters with latent potential and removing friction, which is what Detroit’s actually trying to do with drones and mobility.