China’s AI Boom Rewires Power Grid, Pushes Hubs to Deserts
When Chinese markets reopened after the May Day holiday, state-owned utility Datang International Power Generation Co. Ltd. embarked on a dizzying stock rally. After hitting the daily trading limit six times and steadily climbing, its shares reached a record high of 9.92 yuan ($1.5) on June 3 — surging more than 130% in a single month, according to Caixin Global.
The catalyst? The market interpreted Datang’s launch of a 500-megawatt solar plant in Zhongwei, Ningxia, as China’s first large-scale green-energy direct-supply project for computing power. The company issued eight risk advisories stating it had no operational computing-and-power synergy projects, but the signal was clear: China’s artificial intelligence boom is fundamentally reshaping the country’s energy infrastructure.
The Scale of AI’s Energy Appetite
China’s computing electricity consumption has more than doubled in six years, rising from 82.4 billion kilowatt-hours in 2019 to 196 billion kWh in 2025 — growing from 1.3% to 1.9% of total power consumption. The China Academy of Information and Communications Technology projects this could reach 500 to 700 billion kWh by 2030.
To put that in perspective, China’s total electricity consumption surpassed 10 trillion kWh in 2025, making it the first country to do so. That figure more than doubles the United States and exceeds the combined consumption of the European Union, Russia, India, and Japan, as reported by the National Energy Administration (NEA).
The driving force behind this surge is staggering: daily AI token usage in China surpassed 140 trillion by March 2026 — a more than 1,000-fold increase from early 2024. As of 2025, China had built 42 intelligent computing clusters at a 10,000-card scale, each consuming enormous amounts of power.
The Technical Challenge: Not Volume, but Instantaneous Delivery
Contrary to fears of widespread blackouts, the core challenge is not total energy volume but the grid’s ability to deliver stable, instantaneous power to specific locations at specific times. Wang Zesen, deputy director of the power system research institute under State Grid Jibei Electric Power Co., explained in a Caixin report that AI inference workloads create a “peak upon a peak” effect, with AI servers causing instantaneous fluctuations of up to 10,000 kW.
“The core challenge is not total energy volume but instantaneous power delivery and grid stability,” Wang said.
This volatility is compounded by a severe timeline mismatch: data centers can be built in 8 to 24 months, while substations take 3 to 5 years. Opaque utilization rates — ranging from 10% to 80% — make grid planning extraordinarily difficult, risking either wasted capacity or stranded assets.
Gao Xing, chief utilities analyst at China Securities, offered a measured perspective in the same Caixin report. He argued that fears of AI causing widespread power shortages are overblown, noting that AI’s annual demand increase represents only about 1% of total load. The real transformation, he suggested, is a fundamental shift in how the grid must operate.
The Eastern Data, Western Computing Strategy
China’s answer to this challenge is the “Eastern Data, Western Computing” strategy, launched in 2022. The policy directs energy-hungry data centers to resource-rich western regions where renewable energy is abundant and cheap. Data centers in these hubs must source over 80% of their power from green energy.
Zhongwei in Ningxia exemplifies this model. Electricity costs there run at 0.36 yuan per kWh — 20% below the local average — thanks to Datang’s solar farm and a planned 1.5-gigawatt wind farm. Over 65% of electricity in the Zhongwei cloud computing park already comes from green sources.
Further north, Ulanqab in Inner Mongolia operates with 67% green power and aims to brand itself as China’s “Token Capital” by 2029, having contracted more than 5 million standard server racks. Meanwhile, Tencent deployed a small-scale computing experiment in a zero-carbon industrial park in Chifeng, Inner Mongolia in June 2025, operating entirely on green power. The tech giant aims to run its entire data center portfolio on 100% green energy by 2030.
Government Policy Response
The Chinese government has moved quickly to formalize the computing-power coordination agenda. In March 2026, it officially endorsed computing-power coordination in its work report. On May 8, four top government bodies — the National Development and Reform Commission (NDRC), the National Energy Administration (NEA), the Ministry of Industry and Information Technology (MIIT), and the National Data Administration — issued a joint action plan targeting a secure, green energy guarantee system for AI by 2027 and significant mutual empowerment between AI and energy by 2030.
On May 26, the NEA released the China AI Plus Energy Development Report 2026, unveiling 51 high-value AI-plus-energy application scenarios across eight sectors. Lin Boqiang, dean of the China Institute for Studies in Energy Policy at Xiamen University, told Caixin that by spelling out explicit application scenarios, “the government is pushing the industry from conceptual rhetoric to concrete implementation.”
China expects to invest more than 5 trillion yuan ($0.74 trillion) during the 15th Five-Year Plan (2026-2030) on upgraded power grid infrastructure alone, according to the NDRC.
Virtual Power Plants: A Glimpse of the Future
China is experimenting with virtual power plants (VPPs) to manage the volatility of AI-driven energy demand. As of April 2026, the country had 535 virtual power plants in operation.
In December 2025, State Grid’s Shanghai branch and China Telecom executed a historic test of rapid cross-provincial computing transfer, migrating AI inference tasks to Fujian within three minutes to shed 50 kW of local load. In May 2026, Guangdong province integrated three large telecom data centers into its electricity spot market via a virtual power plant, enabling flexible load management.
Guo Yunzheng, an executive at China Mobile Energy Technology Co., warned at the 14th Energy Storage International Conference in April that AI data centers will present three critical challenges: high power, high volatility, and high capacity, as Caixin reported. He proposed a three-tiered storage solution involving supercapacitors for millisecond-to-second peaks, batteries for second-to-minute fluctuations, and grid-side systems for longer-duration adjustments.
Local Grid Strains and the Dual-High Risk
The strain is already visible in several Chinese cities. In Datong, the computing sector consumed over 6 billion kWh in 2025 — a 40% jump that surpassed coal consumption — accounting for 26.2% of the local grid load. Similar surges are occurring in Zhangjiakou and Gui’an.
Regions with high renewable generation AND large data center loads face what experts call a “dual-high” stability risk. This was dramatically illustrated in July 2024, when a thunderstorm in northern Virginia, U.S., triggered a voltage disturbance that caused approximately 60 data centers to drop off the grid, losing 1,500 megawatts of load in seconds.
What’s Next: The Road to Integration
Sun Chuanwang, a professor at Xiamen University, observed that “the interplay of AI and energy has moved from one-way support to deep integration.” The path forward requires unifying fragmented power, computing, telecom, and carbon markets — a monumental coordination challenge.
Tech giants like Alibaba and Tencent are investing in next-generation nuclear technology, including small modular reactors and nuclear fusion startups, to secure stable, clean baseload power for their data centers.
Wang Yongzhen, a computing-power coordination expert, expects the experimental phase to last another three years. The ultimate goal, he said, is integrating security, green energy, and economic efficiency into a seamless system where computing and power work in harmony.
For now, China is racing to synchronize its digital and energy infrastructure before the power demands of AI outpace the grid’s capacity to deliver. The outcome of this race will not only determine the trajectory of China’s AI ambitions but could also serve as a blueprint for the world as AI reshapes global energy systems.