In Inner Mongolia’s grasslands, farmers now work beside a new kind of neighbor: massive data centers rising from the same fields once known only for potatoes as per recent report published by BBC
Ulanqab has long been called China’s “potato capital.” Now it’s also become a construction zone for the computing power behind China’s AI ambitions – timed almost exactly as Donald Trump and Xi Jinping prepare to meet and discuss AI safety.
Building at “China Speed”
Workers here move fast. Cranes rise above scaffolding as welders and engineers race through phase after phase of construction. “This is China speed,” one worker jokes, calling his country “an infrastructure maniac.” An electrical engineer on-site says this is only phase two – with many more still to come.
These data centers matter enormously to Beijing. They’re central to China’s plan to embed AI across 90% of its industries and society by 2030, and the Communist Party treats them as core to both economic growth and national security.
Talent Is Coming Home
China isn’t just building infrastructure, it’s building a workforce. Rising tensions with the US, along with tighter American immigration rules, are pushing young researchers back home.
Jack Zhang of the University of Science and Technology of China notes a clear shift: fewer students and researchers are heading to the US, while more are choosing to return from the US to China instead. Zhang himself followed that path, studying at Yale and Cambridge before coming back, drawn by a level of government support that’s harder to access abroad.
Why the US Won’t Slow Down Either
China’s rapid AI growth is part of why the US refuses to pause its own development, despite warnings from inside the industry. “Whoever wins AI, wins,” Trump said last week. Anthropic CEO Dario Amodei has pushed to keep US chip export bans on China in place, warning that a Chinese AI lead would be dangerous for the world.
Beijing sees it differently. State-run Global Times dismissed US concerns as a “Cold War playbook” meant to contain China’s rise.
Open-Source vs. Closed Models
China is also pushing a different AI strategy: cheaper, open-source models anyone can modify – a contrast to closed systems from OpenAI and Anthropic. Sharing the technology has let Chinese firms innovate quickly, helping offset their biggest handicap: limited access to advanced chips.
At the same time, Xi wants a bigger voice in setting global AI rules, repeatedly calling for international regulation and a shared framework.
The Real Cost, In the Villages
Not everyone is celebrating. China is building data centers across low-population regions- Ulanqab, Ningxia, Gansu, Guizhou – drawn by cheap land, cooler weather, and growing wind and solar capacity. But coal still powers parts of the grid, and water remains scarce in dry areas like Inner Mongolia.
Entire villages have been relocated, though details on numbers or compensation are hard to confirm – it’s a sensitive topic locals are wary of discussing. “This big data has absolutely nothing to do with ordinary people,“ one resident said anonymously. “What benefits has it brought us?” Few, he said, actually found jobs at the sites.
Others see it differently. A local resident, also named Zhang, stood near a group of potato pickers and said his children had found jobs through the AI boom. In his view, China now stands at the forefront of the world.
Not Everyone Is Along for the Ride
Away from the labs, many Chinese citizens face more immediate concerns a slow economy, youth unemployment, falling property prices, and an aging rural population.
One Inner Mongolian farmer summed it up simply while checking on his sheep: “Big data and all that pretty good but it has nothing to do with me.” He barely uses the internet, relying on a cracked phone screen and a basic data dongle.
That quiet skepticism may be Beijing’s biggest challenge not building the technology, but convincing everyday citizens it’s actually for them.
