Hours west of Beijing by train, travelers arrive at Ulanqab, a city nestled in the vast, arid landscapes of Inner Mongolia. This region, once dominated by sheep farming and coal mining, has recently become a critical epicenter for China’s burgeoning artificial intelligence (AI) data center industry. Since 2016, nearly 100 data centers have been established, a figure that has surged to 12.5 gigawatts of combined capacity, representing a significant portion of the total compute cluster in Asia, according to Goldman Sachs’ recent report. This rapid expansion, fueled by a confluence of factors, paints a compelling picture of China’s strategic push towards technological dominance within the AI sector. The city’s unique geographic location – situated at a high elevation on the Inner Mongolian Plateau, shielded from harsh winters – combined with proximity to Beijing and a comparatively low electricity cost driven by wind and solar energy, has attracted substantial investment. Chinese companies, including DeepSeek, ByteDance, Alibaba, and Xiaohongshu, are aggressively expanding their infrastructure, representing a significant shift from years of less intensive investment. The Ulanqab data center boom isn’t simply about capacity; it’s about strategic positioning – a move away from reliance on traditional cloud computing while simultaneously bolstering China’s energy security through renewable energy integration. However, this growth is shadowed by a critical challenge: the city’s extremely dry climate, receiving only roughly 14 inches of rainfall annually. The local government is actively addressing this scarcity, with recent water restrictions and a recent water company shutdown – a stark example of the environmental considerations driving the rapid development. The data centers are projected to require less water than the surrounding area, but the long-term environmental impact of the region remains a significant concern, particularly given the potential for increased dust and thermal pollution. Furthermore, the initial latency challenges posed by the data centers’ location, requiring substantial data transfer times, have been largely overcome by the increasing demand for AI-driven models, which can be trained almost instantaneously. The Ulanqab data center boom is a result of a complex interplay of economic incentives, government policy, and technological innovation, making it a crucial element in China’s broader vision for a digitally-driven future. The government sees it as a win-win: bolstering AI capabilities while simultaneously increasing the nation’s energy production. For example, the report indicates that the data centers are currently supporting approximately 37% of the city’s electricity, a figure that will increase as more AI-powered models are deployed. As Andrew Stokols, a professor at Singapore Management University, notes, this represents a key turning point for China’s compute infrastructure – a move that’s proving to be more adaptable and economically viable than previously anticipated. Despite the government’s enthusiasm, experts caution that the reliance on coal for power generation remains a persistent issue, particularly given the region’s historical ties to coal extraction. The full extent of this transition to renewables is still uncertain, and the data centers’ impact on regional ecosystems will require careful monitoring and mitigation strategies.”
Watch Related Video
Source: Wired
