China is rapidly expanding its data centre infrastructure across energy-abundant regions such as Inner Mongolia, aiming to leverage cheap electricity and land to gain a competitive edge in the global artificial intelligence (AI) industry. The government has identified AI development as a strategic national priority, seeking to enhance economic productivity, reinforce domestic political influence, and broaden China’s global technological reach.
One of the most notable examples of this expansion is Ulanqab, a city in Inner Mongolia known traditionally for agriculture but now emerging as a major data centre hub. Approximately 89 data centres are either operational or planned in the area, with committed computing capacity reaching about 15 gigawatts (GW), marking it as Asia’s fastest-growing data hub. This rapid development benefits from a combination of factors including abundant wind and coal power generation, inexpensive land, a cool climate that reduces cooling costs, and a supportive construction environment.
China currently operates around 24GW of data centre computing capacity—surpassing the combined total of the rest of Asia but remaining lower than the United States’ 56GW capacity. Additionally, nearly 50GW of capacity is either under construction or has been announced. In contrast, the US data centre boom is reportedly constrained by challenges in securing sufficient electricity supply and grid connections, whereas Inner Mongolia experiences significant electricity oversupply.
Inner Mongolia itself boasts approximately 117GW of installed wind power capacity as of June 2026, the largest in China and nearly four times that of the United Kingdom. The region also maintains around 130GW of fossil fuel power, predominantly coal. Experts have described the area as having one of the largest local electricity surpluses globally, with much of the electricity currently underutilized.
Several major Chinese technology firms and data centre operators are active in Ulanqab. Alibaba and Huawei have established multiple facilities, complemented by operators like ZData and VNET, which provide infrastructure leasing to technology companies. ByteDance has emerged as the largest domestic renter of computing services, while AI research labs such as DeepSeeker and Z.ai are also developing facilities there. Huawei’s chair, Eric Xu, noted the city’s appeal due to its low electricity costs and proximity to Beijing, only two hours away by rail, expressing surprise at the scale of the infrastructure developed during his visit in August.
Despite China’s progress in data centre construction, the country continues to face challenges in securing advanced semiconductor chips critical for AI applications. Nonetheless, the ability to mobilize resources such as land, energy, and construction capacity has allowed rapid expansion of the physical infrastructure required to support AI development.
Overall, China’s accelerating build-out of data centres, particularly in regions like Inner Mongolia, reflects a broader strategic effort to position the country as a leading player in AI technology by capitalizing on local energy advantages and supporting growing domestic and international demand for computing capacity.