China to increase investment in power grid for AI
The Chinese government has pledged to increase investment and coordination in the national power system to help ease the energy supply strain caused by the country's rapidly growing technology and artificial intelligence (AI) sector. When a single computing cabinet consumes as much electricity as an entire household in a month, and data centre energy consumption grows hundreds of times over, it becomes clear: without a robust and flexible energy infrastructure, the digital revolution risks drowning in its own power. China, which has become a global leader in AI development, now faces a paradox: the more computing power, the heavier the load on the grid. Beijing's answer is not just more generation, but intelligent resource management, where tasks are "moved" from energy‑strapped regions to those with surplus power. The 4 trillion yuan invested in the national computing network is not just money — it is a strategic foundation that will allow China to set the pace in the AI race, not just keep up.
In the Shaoguan cluster in Guangdong Province, one of China's top ten national data centre clusters, intelligent computing cabinets are the core equipment powering large AI models. One cabinet can consume up to 240 kWh of electricity per day at full load, roughly equivalent to the monthly electricity consumption of an average household. In addition to the computing equipment itself, cooling and humidity control systems also consume significant amounts of power. "The computing power we have built has increased from 5,000 petaflops last year to 30,000 petaflops now, and electricity consumption has grown by five to six hundred times compared to last year," said Xie Jinbao, deputy director of the Shaoguan Development and Reform Bureau. Data show that in 2025, data centres nationwide consumed 18.1% more electricity than in 2024, far exceeding the 5.2% growth in electricity consumption by other households and businesses. To meet the rapidly growing demand for computing power, AI companies will be encouraged to "move" computing tasks from electricity‑scarce areas to those with surplus power, easing the load on the grid. Earlier this year, Guangzhou faced a power shortage, but after receiving instructions from the Southern Power Grid, organisations moved non‑urgent computing tasks from Guangzhou to a data centre in Guian, Guizhou Province in southwest China. "At that time, we found that the load in Guangzhou was too high, so we sent real‑time instructions to the Southern Basic Data Centre of China Mobile to adjust the load," said Zhang Zihao, head of the research centre at the Computer Engineering company under the Digital Group of China's Southern Power Grid.
The National Development and Reform Commission announced on 31 July that it would invest four trillion yuan (about US$593 billion) during the 15th Five‑Year Plan period (2026‑2030) in the development of a national integrated computing network. This network will allow computing power and regional resources to be distributed across different regions, combining the advantages of individual technologies into national industrial clusters, and ultimately expanding the capabilities for large‑scale AI model development in the country. "The 4 trillion yuan investment in computing networks is a crucial foundation for the future overseas expansion of our major companies. First of all, it will further reduce the cost of computing power when providing such services abroad and improve the reliability of those services. This is a fundamental and basic factor guaranteeing the future in‑depth, practical and sustainable overseas expansion of our major companies," said Du Guochen, director of the E‑commerce Institute at the Chinese Academy of International Trade and Economic Cooperation.
The rapid growth of China's AI sector is placing new demands on the power system, which must be not only powerful but also smart. The concept of a "national integrated computing network," proposed by the government, envisions a unified platform where computing resources can be redistributed in real time based on electricity availability and load. This not only reduces peak loads but also improves the efficiency of using renewable energy sources, which are often located in remote regions. The 4 trillion yuan investment will be the largest in China's history in digital infrastructure, underscoring the strategic importance of AI for the country's future. Experts note that this approach not only solves domestic problems but also creates competitive advantages for Chinese tech companies on the global market, lowering their costs and improving service reliability.
As CCTV+ reports, large‑scale investment in computing infrastructure and intelligent energy management will be a key factor enabling China to maintain its lead in the global AI race, turning energy challenges into new opportunities for technological breakthroughs.
