China’s Green Computing Transformation Shows Early Gains
China’s computing power industry is demonstrating tangible progress in its full-chain green transformation, according to a new report released at the 2026 Green Computing Power (AI) Conference in Hohhot, Inner Mongolia. The Green Computing Power Development Research Report (2026), unveiled by Ao Li, Deputy Director of the China Academy of Information and Communications Technology (CAICT), highlights significant advances in energy efficiency, infrastructure modernization, and the integration of computing with renewable power systems.
A Rapidly Expanding Digital Backbone
The report reveals that China’s intelligent computing capacity reached 2,185 EFLOPS (FP16) by the end of June 2026, with western regions accounting for 32.6 percent of the national total. By the end of 2025, the country had more than 13.73 million standard racks in operation, 42 ten-thousand-card-level intelligent computing clusters, and the eight national hub nodes concentrated over 80 percent of the nation’s intelligent computing power.
This expansion is accompanied by a significant push toward energy efficiency. More than 160 data centers across China have now been rated as 4A-level or above green data centers, reflecting a broader industry shift toward sustainable operations. New technologies including liquid cooling, high-speed optical interconnects, and advanced storage architectures have become essential components of high-density intelligent computing centers.
The Energy Challenge Behind AI Growth
The scale of the challenge is considerable. In 2025, China’s computing facility electricity consumption reached 170 billion kilowatt-hours, a year-on-year increase of about 30 percent, according to the National Energy Administration. The explosive growth of artificial intelligence is a primary driver, with National Energy Administration Director Wang Hongzhi noting that generating a five-second high-definition video through AI consumes as much electricity as charging ten mobile phones.
The global context underscores the urgency. The International Energy Agency projects that data centers worldwide will consume 945 terawatt-hours of electricity annually by 2030. In China, data center electricity consumption doubled from 82.4 TWh in 2019 to 166 TWh in 2024, with electricity accounting for 60-70 percent of data center operating costs.
Computing-Power-Electricity Coordination: From Concept to Practice
A defining feature of this transformation is the elevation of “computing-power-electricity coordination” (算电协同) to national strategic priority. For the first time, the concept was included in the 2026 Government Work Report, marking its formal recognition as a key pillar of new infrastructure development.
As China Daily reported, the coordinated development of computing capacity and electricity supply was highlighted alongside hyperscale intelligent computing clusters as a critical area of new infrastructure. The approach involves using digital technology, intelligent algorithms, and information networks to deeply integrate computing infrastructure with power systems, enabling dynamic matching and optimal allocation of resources.
Inner Mongolia has emerged as a pioneer in this space. The region’s total renewable energy technically developable capacity stands at 1.88 billion kilowatts, and the power supply reliability rate in Hohhot and Ulanqab has reached 99.999 percent. Hohhot has built a “2-5-20” millisecond latency network: 2 milliseconds covering the Hohhot-Baotou-Ordos-Wulanchabu metropolitan area, 5 milliseconds to Beijing-Tianjin-Hebei, and 20 milliseconds to the Yangtze River Delta and Greater Bay Area.
The Token Economy Takes Shape
A striking indicator of the AI boom is the explosive growth of the “token economy.” Daily token call volume in China surged from approximately 100 billion at the beginning of 2024 to about 140 trillion by March 2026 — a growth of over 1,000 times in just over two years. Tokens, the smallest unit of measurement for AI information processing, are increasingly being traded like commodities, with users paying precisely for the number consumed.
The conference saw the launch of the nationwide first green computing power full-stack AI token trading platform in Hohhot, as well as Inner Mongolia’s first trusted data space and the National Video Large Model Computing Base. As Xinhua Finance reported, these platforms are designed to accelerate the shift of AI capabilities from technology supply to industrial application.
Major Investment Signals Confidence
The conference also delivered a strong market signal with 13 major projects signed, representing a total investment of 136.1 billion yuan (approximately $19 billion). The projects are distributed between Hohhot (8 projects) and Ulanqab (5 projects), covering chip testing, equipment manufacturing, computing centers, and token factories. Key signings include a Volcano Engine gigawatt-level park, a Cambricon chip laboratory, and a GCL Energy Technology computing-power-electricity coordination integrated project, according to Sina Finance.
These projects represent a full-chain approach to the computing power industry, addressing upstream hardware gaps, midstream computing infrastructure, and downstream AI applications. The investments signal confidence in Inner Mongolia’s evolution from a traditional data center hub to a comprehensive computing power ecosystem.
Expert Perspectives and Future Directions
Ao Li, who released the report, suggested four priorities for further developing green computing: perfecting the green computing power system, deepening computing-power-electricity coordination planning, improving industry standards, and strengthening international cooperation. He noted that driven by the doubling of AI accelerator chips, global AI capacity is expected to grow sevenfold from 2025 to 2027.
Hu Maosheng, General Manager of Volcano Engine Public Service Industry Technology, emphasized that “green computing power is becoming important infrastructure for AI development, and the token economy is accelerating the shift of AI capabilities from technology supply to industrial application.”
The conference, which attracted over 500 attendees including academicians, experts, and industry leaders, reflects China’s broader strategy of positioning green computing as a cornerstone of its digital economy. With the “East Data, West Computing” project launched in 2022 designating eight national computing hub nodes, and computing-power-electricity coordination now embedded in national policy, the foundations are being laid for a more sustainable and efficient computing future.
What to Watch Next
As China continues to scale its AI infrastructure, several key developments bear watching: the implementation timeline for the 136.1 billion yuan in signed projects, the operational rollout of the green computing power token trading platform, and how the computing-power-electricity coordination model expands beyond Inner Mongolia to other western regions. The integration of renewable energy with computing infrastructure will be critical not only for China’s digital ambitions but also for the global push toward sustainable AI development.
With data center electricity consumption projected to grow exponentially worldwide, China’s approach to green computing transformation offers a potential template for balancing technological advancement with environmental responsibility. The initial results are promising, but the full impact of this transformation will unfold over the coming years as these projects come online and the model scales across the country.