China Races Ahead in Computing Hubs as United States Expansion Stalls
Beijing, Friday, 13 November 2026.
Building facilities at one-fifth of United States costs, China is rapidly constructing massive computing hubs, leveraging double America’s power generation to overcome microchip trade barriers.
China’s AI Data Center Surge: Strategic Infrastructure at Scale
As the United States grapples with local resistance to AI data center expansion, Chinese developers are accelerating the construction of high-capacity computing hubs across the country, particularly in Inner Mongolia’s Ulanqab region. This strategic push, backed by state-level support and minimal regulatory friction, is enabling China to rapidly close the technological gap with Western competitors. The contrast is stark: while U.S. tech giants face community opposition over grid constraints and rising electricity tariffs, Chinese firms benefit from abundant energy resources, cold climates for efficient cooling, and significantly lower construction costs. A November 11, 2026 report by CBS News highlighted Ulanqab’s transformation into a sprawling AI data center hub, with dozens of facilities under construction, driven by cheap and abundant energy from fossil fuels, wind, and solar sources [1]. The region’s energy costs are among the lowest globally, with electricity priced at approximately 0.055 USD/kWh [3], further enhancing its competitive edge.
Cost Advantage: China’s Data Centers at One-Fifth of U.S. Expenses
The cost disparity between Chinese and U.S. data centers is a defining factor in the current AI infrastructure race. According to a report by SemiAnalysis, a U.S.-based semiconductor research firm, constructing a 100–120 MW data center in China costs less than 300 million USD, excluding IT equipment—a figure representing just one-fifth of comparable U.S. projects [3]. This cost advantage is attributed to lower land and labor expenses, streamlined construction processes, and economies of scale. For instance, the Galaxy Campus in Ulanqab, developed by Envision Group, delivered its first 120 MW phase in under a year, with construction costs significantly undercutting U.S. benchmarks [3]. SemiAnalysis noted that peak construction periods in China employ 3,000 to 4,000 workers, three to five times the workforce typically seen in U.S. projects of similar scale [3]. The report also highlighted that China’s data center rental rates hover around 40 USD per kW per month (excluding electricity), aligning roughly with the annual operational costs driven by ultra-low energy prices [3].
Energy and Vertical Integration: The Backbone of China’s AI Expansion
China’s AI infrastructure expansion is not merely about building facilities but leveraging vertical integration to optimize energy and operational efficiency. Envision Group’s Galaxy Campus exemplifies this approach, combining wind energy farms, battery storage systems, and modular substations to create a self-sustaining ecosystem. The campus sources power from two proprietary wind farms located 80 kilometers away, with a combined capacity of 420 MW—nearly 1.75 times the initial 240 MW load requirement [3]. A 4-hour battery storage system ensures continuous power supply during low-wind periods, with over 60% of the first two phases’ energy needs met by on-site wind power [3]. This model, termed ‘source-grid-load-storage’ integration, allows energy companies to convert electricity into a tradable commodity, where data centers serve as a high-demand off-taker for excess renewable energy. The strategy aligns with China’s broader goal of reducing reliance on imported chips while capitalizing on its strengths in energy infrastructure and industrial policy.
Global Implications: China’s Infrastructure Edge in the AI Race
The rapid deployment of AI data centers in China is reshaping the global competitive landscape, particularly as energy and infrastructure constraints become critical bottlenecks in the U.S. and Europe. While U.S. firms continue to dominate in high-end AI chip development, China’s focus on scale, cost efficiency, and energy resilience is narrowing the gap. According to IDC’s 2026 Global Data Center Facilities Tracking Report, global data center rack deployments are projected to grow from 25 million in 2026 to 58 million by 2030, with annual energy consumption rising from 519 TWh in 2025 to 1,308 TWh in 2030 [5]. China’s installed smart computing power reached 2,185 EFLOPS (FP16) as of June 2026, marking a 177% year-over-year increase and reflecting the country’s aggressive infrastructure build-out [5]. Analysts suggest that power availability may become the decisive factor in AI competitiveness, with China’s installed power generation capacity exceeding twice that of the U.S. [1]. However, challenges remain, including the need to balance rapid expansion with grid stability and environmental sustainability. As Henry Wang, founder of the Center for China and Globalization, noted, ‘Chips may not be the latest diving criteria. It may be that power is the bottleneck, and whoever goes through that first gets the advantage. And right now, China has the edge’ [1].