On August 22, 2026, the fourth annual Green Computing Power (Artificial Intelligence) Conference took place on the Cilechuan Grassland in Hohhot, Inner Mongolia, drawing attention to the city's pivotal role in China's computing power landscape. The event, themed "Building Synergy between Computing and Electricity, Exploring New Horizons in Token Economy, and Creating the Future of Artificial Intelligence Together," unveiled the Green Computing Power Development Research Report (2026) and highlighted pilot projects such as "millisecond-level urban computing." Major companies, including ByteDance and Cambricon, signed agreements valued at tens of billions of yuan to develop green intelligent computing centers and token factories.
The conference's location is no coincidence. Hohhot offers a unique combination of factors that make it an ideal hub for data centers. With an average annual temperature of just 7°C and over 200 days per year suitable for direct outdoor air cooling, the city achieves more than 20% savings in cooling costs. Electricity prices are also competitive at about 0.36 yuan per kilowatt-hour, enabling large-scale GPU clusters to save around 50 million yuan annually on power expenses. Currently, Hohhot's total computing capacity reaches 150,000 P, with green electricity usage exceeding 80%.
The city's green credentials are bolstered by an innovative "computing-electricity synergy" mechanism. Wind turbine blades on the grassland each generate 12 kilowatt-hours of green electricity per rotation, which is transmitted via a dedicated 41-kilometer line directly to data centers. China Huadian Corporation has constructed a 360,000-kilowatt integrated wind-solar-storage power station, combining 300,000 kilowatts of wind, 60,000 kilowatts of photovoltaic power, and energy storage capable of four-hour regulation. A collaborative control platform collects data on wind output, solar irradiance, storage status, and computing load at millisecond intervals to dynamically allocate power: when green energy is abundant, storage systems charge and delay-tolerant tasks are prioritized; when green energy is limited, storage discharges, the public grid supplements power, and flexible tasks are deferred. This bidirectional interaction between power generation and consumption is a key innovation.
On the consumption side, efforts to minimize waste are also underway. The Hohhot data center operated by China Mobile recovers waste heat from nearby power plants to provide cooling for machine rooms. The B02 machine hall has improved its Power Usage Effectiveness (PUE) from 1.315 to 1.209, resulting in annual electricity savings of 560,000 kilowatt-hours since commencing operations. Upon full operation of two machine halls, annual savings are projected to reach 10 million kilowatt-hours of electricity and 150,000 tons of water.
The significance of these developments extends beyond Hohhot. Training large-scale AI models utilizing tens of thousands of GPUs can consume over 100 million kilowatt-hours annually. For instance, the Hohhot Intelligent Computing Center, operated by China Mobile, incurred electricity costs of approximately 290 million yuan in 2025, accounting for more than 70% of its total operating expenses. In the competition for computing power, electricity costs prove to be a decisive factor. China's strategic "East Data West Computing" project redirects high computing demand from eastern regions to resource-rich western areas, with the Horinger data center cluster in Hohhot among the top ten clusters nationwide.
Today, 62 computing center projects have clustered here, including data centers of China's three major telecom operators, 16 financial institutions, and computing nodes from Huawei, ByteDance, Kuaishou, and others. Upstream, 45 manufacturing enterprises produce servers, cabinets, and sensors locally. According to the recently published three-year action plan, by 2028, Hohhot aims to establish 10 token factories capable of producing 20 trillion tokens daily. At that time, customers will procure not only cabinets and bandwidth but also directly accessible intelligent services customized to their requirements.
The prairie winds remain constant, as does the city's commitment to converting every kilowatt-hour of green electricity into stable and reliable computing power. This conference underscores Hohhot's emergence as a model for sustainable AI infrastructure, balancing environmental concerns with the growing demands of the digital economy.
