Bernstein: Computing Power Racing Ahead, Electricity Taking the Baton, China's AI Cost Advantage Ignites Long-Term Opportunities in Grid Energy Storage

Bernstein's September 2026 research report focuses on the linkage between China's AI industry and its power system. The series aims to identify China's next-generation key industries and leading companies in the sector. The report notes that China has formed a structural competitive advantage in AI development by leveraging power costs, energy infrastructure, and policy coordination. The large-scale expansion of AI computing power will ignite data center electricity demand, bringing long-term investment opportunities to the grid and energy storage industry chain. Bernstein notes that China's artificial intelligence enjoys a structural cost advantage. China's leading large models, while competitive in intelligence scores, are often priced at only a fraction of comparable U.S. models, creating an attractive "intelligence per dollar" product advantage. This advantage stems from more efficient model architectures, lower infrastructure costs, and cheap electricity, enabling domestic service providers to offer AI inference services at significantly lower cost. As the focus of AI development shifts from training to large-scale inference, "intelligence per unit cost" will become a core competitive dimension, and low-cost computing power will become an important moat for domestic models to break through. The bank expects that over the next decade, electricity consumption by hyperscale internet companies will become one of the fastest-growing sources of power demand. With the spread of artificial intelligence, continued expansion of data center infrastructure, and China's goal of narrowing the computing power gap with the United States, the bank estimates that data center electricity demand will grow from 202 TWh in 2025 to 1,476 TWh in 2035, a compound annual growth rate of 24%. China's current AI computing power scale is only about 15% of that of the United States, but domestic AI chips and computing infrastructure are rapidly expanding. To match the United States at the computing power level, large-scale data center expansion will be required, which will bring enormous incremental electricity demand. Cheap, abundant, and scalable electricity supports China's AI development goals. China's total power generation is already more than twice that of the United States, with over 500 GW of new installed capacity added in 2025 alone. At the same time, China's industrial electricity prices are among the lowest globally; in core AI hubs such as Inner Mongolia, Ningxia, and Gansu, data center electricity costs are among the lowest in the world. China also combines low-cost renewable energy with cost-competitive nuclear power, providing highly cost-effective and reliable power supply for AI operations. Data centers will increasingly use renewable energy for power supply. Relying on initiatives such as "East Data, West Computing," green data center measures, and the emerging "computing-power-electricity coordination" framework, policy is driving coordinated development of computing power expansion and power infrastructure. New data centers at national-level computing hubs must have a green power share of more than 80%; leading operators have also set targets to achieve 100% renewable energy use by 2030. The deployment of computing power cannot be separated from grid and energy storage support. From 2026 to 2030, domestic grid investment is planned to exceed RMB 5 trillion, vigorously building ultra-high-voltage transmission lines to open channels for transmitting clean energy from the west to computing load centers. In energy storage, China's lithium battery capacity accounts for 80% of the global total. In 2026, domestic energy storage installations are expected to grow 95% year-on-year to 300 GWh. Renewable energy consumption combined with the need for reliable power supply for AI computing power will bring high growth to the energy storage industry. The bank believes the most investable directions are concentrated in grid infrastructure and energy storage. Ultra-high-voltage transmission and large-scale grid investment benefit equipment suppliers; a rising share of renewable energy combined with rising AI power demand further drives demand for battery energy storage. Among the names covered in the report, Contemporary Amperex Technology is well positioned to benefit from its leading position in grid-scale energy storage systems; Sungrow Power Supply is positioned across the entire track of renewable energy, power conversion equipment, energy storage, and grid integration.