Nearly 90 data centers are slated for construction in Ulanqab, a city of fewer than two million people in Inner Mongolia, as China races to build the infrastructure backbone for AI dominance. The scale is staggering. Most of the densely packed Bayin Big Data Industrial Park took shape in just two years, turning a remote region into a computational nerve center with ambitions that rival Silicon Valley’s.
What’s driving the frenzy? Beijing announced a $560 billion infrastructure plan last month to rapidly expand computing facilities over the next five years. The government sees AI as a potential engine of growth for an economy described as slowing and facing a shrinking population. Chinese leader Xi Jinping told a global artificial intelligence conference in Shanghai this past July that “China has embraced AI with open arms,” pointing to an initiative to bring AI into sectors ranging from health to home goods. Official statistics say more than 700 million people in China already use AI.
But there’s a catch. Stanford University research found that the United States had roughly 10 times more AI data centers than China as of last year, while other estimates show it had vastly more computing power. American technology firms also have access to the most advanced chips, which the research describes as the single most important ingredient for developing AI. China, meanwhile, faces U.S. export bans that restrict access to advanced chips from companies like Nvidia and AMD.
The Infrastructure Boom’s Human Cost
Local residents have experienced rapid economic change, though not without friction. Ma Rong, 56, who has run a grocery store and guest house in Ulanqab for 30 years, said construction brought steady business. “Before, we’d just sit there and wait for customers. Now we’re constantly busy—picking up orders, doing this and that. We barely get a break. Our business has gotten much better than before.” Restaurant owner Liu Baoping, 58, said his daily clientele had doubled alongside the data-center buildout, while acknowledging a tradeoff: “The environment isn’t as clean as before, but people’s incomes and the economy—it’s way better.”
These accounts expose a fundamental tension in China’s development model: rapid growth carries environmental and quality-of-life costs that aren’t shared equally. Workers gain employment, but the infrastructure itself generates pollution and disruption. Public opposition to development is virtually nonexistent under the Communist Party’s control, leaving affected communities without an institutional mechanism to negotiate those tradeoffs.
China’s Chip Disadvantage and the Energy Workaround
With high-end hardware in short supply, China is pursuing a different strategy. “If you look at the overall production cost of building a data center, chips are the most important factor, and they just don’t have enough,” said Nick Corvino, a researcher at the Carnegie Endowment for International Peace. Investment by listed Chinese companies is expected to double this year to around $139 billion, according to research firm Rhodium Group—still less than one-fifth of the $800 billion their U.S. peers are investing in AI data centers.
Instead, Beijing is betting on what it has: cheap, abundant power. Between 2022 and 2025, China added electricity capacity equivalent to that of industrial power sources across the entire United States as of the end of 2025. In the first six months of 2026, China wasted enough excess green energy to power Mexico for a year, according to findings from the Centre for Research on Energy and Clean Air and Global Energy Monitor.
China launched its “Eastern data, Western compute” strategy four years ago. The country has designated eight national computing hubs, including several in the west, and 10 national data-center clusters. Its latest five-year national economic blueprint calls for linking computing infrastructure with green energy. The aim is to absorb excess renewable energy and support the green technology sector—a rational use of stranded power generation, at least in theory.
Environmental Promise and Performance Gap
Inner Mongolia, once a major coal-producing region, has become a green-energy hub. Solar panels cover fields, wind turbines rise over grazing sheep, and nearly half of Ulanqab’s electricity grid draws from wind and solar. Under construction several miles from Bayin industrial park, the Envision “Galaxy Campus” is slated to rank among the world’s largest data centers by area when complete. The company says more than 80% of the facility’s power will come directly from company-operated wind farms, bypassing the grid; battery systems will store wind power, while a solid-state transformer adjusts supply according to demand.
The numbers also conceal a darker reality. Chinese chip designers such as Huawei combine clusters of chips to approximate the computing power of more advanced technology. Analysts say the approach can work for some tasks, but it uses significantly more electricity than cutting-edge chips. Put plainly, China’s chip disadvantage forces it to burn more energy for the same computational results, a hidden cost masked by the apparent abundance of renewable power.
“China’s trying to take advantage of the fact that it has this muscle memory for building energy infrastructure quickly,” said Kyle Chan, a fellow at the Brookings Institution. “Whereas for the US, we haven’t done this kind of grid buildout in a long time and so we are struggling on the power generation and grid connection front.”
Structural Limits and Ghost Infrastructure
The strategy has built-in limits. Some western computing hubs sit too far away to handle certain tasks at the speeds urban centers in the densely populated east require. Observers have also raised concerns about “ghost data centers”—rural projects built to meet broad government initiatives but lacking the hardware or users to justify the investment.
Optical cables can transmit data from Ulanqab to Beijing, about 200 miles away, faster than the blink of an eye, and major companies are committing resources. According to general manager Ricky Zheng, the first phase of Envision’s planned 2-gigawatt facility was fully booked. He said the company was building additional phases for AI and cloud services.
This is a state-driven industrial policy racing against time and technological disadvantage. China is deploying massive capital, energy resources, and labor to build infrastructure that the more fragmented, market-driven U.S. system has struggled to scale. Yet centralization, along with the absence of public accountability mechanisms, leaves rural communities with little say as they absorb the environmental costs and distributional consequences.
Why This Matters:
China’s AI infrastructure push marks a critical shift in how state power can overcome market and technological disadvantages. The strategy shows both the capacity and limits of centralized planning: China can build faster and deploy abundant cheap energy, but it can’t easily overcome the structural barriers created by international export controls. More concerning is the model’s human geography. Development benefits concentrate in urban AI and tech sectors, while rural communities bear environmental and quality-of-life costs despite having no responsibility for the decisions. Without democratic deliberation or local veto power, these communities absorb pollution and disruption as the price of national technological competitiveness. Meanwhile, the efficiency losses from using inferior chips to compensate for chip scarcity suggest that China’s resource advantages can’t fully close the technological gap; they can only postpone the reckoning by burning more energy. The question isn’t whether China will catch up to the U.S. in AI capacity, but who pays the price as it does.