Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.

iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more

The unlikely place at the center of China’s AI boom is Ulanqab, a city in the Inner Mongolia Autonomous Region. Cold winters, relatively inexpensive electricity and improved fiber links have made it attractive for data centers—but large project pledges are not the same as operating capacity, and coal dependence and water pressure complicate the growth story.

Where is Ulanqab, and how large is the buildout?

Ulanqab is in Inner Mongolia, within reach of Beijing compared with China’s more remote western regions. WIRED reported on August 21, 2026, that nearly 100 data centers had opened or begun construction there since 2016. That figure combines facilities at different stages; it does not mean all are complete or running.

The same WIRED report cited a Goldman Sachs estimate of 12.5 gigawatts in combined pledged project capacity, with more than 70 percent of those commitments announced in the preceding year. These are estimates and commitments, not a tally of commissioned power or live computing capacity. WIRED put the city’s population at about 1.5 million.

Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.

Why are AI companies building there?

Cold winters can reduce cooling needs

Data centers produce substantial heat, so a cold climate can reduce the work required to keep equipment within operating temperatures. Ulanqab’s long winters are an advantage, though they do not eliminate cooling requirements throughout the year.

Electricity is relatively inexpensive

Power is a major operating cost for computing facilities. WIRED describes Ulanqab’s electricity as relatively cheap, helping explain why operators may find the location attractive. Low prices alone, however, do not establish that the electricity is renewable or that supply is free of reliability constraints.

Fiber links help overcome distance

Ulanqab is not next door to every user or business it may serve, but distance matters differently for different AI workloads. WIRED reported that two dedicated fiber-optic cables, built in 2017 and 2019, brought average latency below five milliseconds. The report does not specify the measurement method or endpoints, so that figure should be understood as its reported average rather than a guarantee for every connection.

Why does AI make remote data centers more useful?

Workload What it does How sensitive it is to latency
Model training Uses computing resources over extended runs to develop or update a model. Often more tolerant of network delay than a live user interaction, making remote facilities viable for some training work.
Inference Runs a trained model to produce an answer or other output for a user. More sensitive to responsiveness because users are waiting for a result; network quality and proximity can matter more.

That distinction helps explain how a city outside China’s largest user centers can still attract AI infrastructure. As Andrew Stokols, a professor at Singapore Management University who studies China’s compute infrastructure, told WIRED: “With the rise of AI in 2022, there was the realization that actually, those remote data centers could be well-utilized for model training.” Better connectivity can also support uses that need more responsive exchanges, but latency still depends on the route and service, not just the city’s reported average.

Free tools Windows power users keep installed

One-click scans. No signup required.

Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.

Which companies are associated with projects there?

WIRED associates DeepSeek, ByteDance, Alibaba and Xiaohongshu with projects in Ulanqab and describes AI companies investing directly in infrastructure rather than relying only on rented cloud computing. Project announcements and plans should not be read as proof that each company has completed or operates a facility there; the reported buildout includes projects at different stages.

Does Ulanqab’s AI growth run on clean energy?

Not simply because wind and solar are expanding. WIRED reported that coal still supplied about 37 percent of Ulanqab’s electricity, citing research by Andrew Stokols. Damien Ma, director of Carnegie China, described Inner Mongolia as “the West Virginia of China. It’s a coal country,” in the report.

Renewable generation can grow without supplying every hour of a data center’s demand. Computing facilities need dependable power around the clock, so announced renewable projects or expanding wind and solar capacity do not by themselves show that the existing data-center load is powered by renewables. Ma also said that “in three years, maybe it will be completely powered by renewables.” That was a conditional forecast, not a confirmed outcome.

Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Support on Ko-Fi

Why is water a concern?

Ulanqab’s climate and power prices come with a local resource constraint. WIRED reported that the city receives roughly 14 inches of rain a year and described pressure on water service even before all planned data-center capacity is operating. The report said the local water company had shut off several waterworks for seven hours each night during the previous month to reduce peak demand.

Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.

Cooling demand is seasonal: according to local government weather data described by WIRED, data centers need additional cooling water during two months of the year. That does not mean water use is constant year-round, but neither does a short cooling season resolve the wider question of how new infrastructure affects a water-stressed city.

What Ulanqab’s rise does—and does not—show

  • It shows why location economics matter: climate, electricity costs and network links can combine to make a city attractive for compute.
  • It shows why project totals need context: pledged capacity and facilities under construction are not equivalent to completed, operating data centers.
  • It does not establish a clean-energy success story: coal remains a reported part of the electricity mix, and renewable growth is not the same as reliable renewable supply at every hour.
  • It does not make water a minor issue: seasonal cooling needs coexist with broader reported pressure on local water service.

Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.