China’s Data Centre Electricity Consumption to Quadruple to 774 TWh by 2030: Wood Mackenzie

China’s data centres are projected to quadruple their electricity consumption to 774 TWh by 2030, driven by the rapid expansion of artificial intelligence (AI) workloads, according to a new report from Wood Mackenzie.

The report, titled “Rewiring China’s Grid for the AI Era,” forecasts that data centres could account for 17% of China’s total electricity consumption by 2060, highlighting the growing impact of AI-driven computing demand on the country’s power system.

“Data centres are becoming an increasingly important part of China’s energy system,” said Wanting Zhao, Research Analyst, Asia Pacific Power and Renewables Research at Wood Mackenzie. She noted that the sustained growth in computing demand is increasing the importance of reliable, cost-competitive and lower-carbon electricity in determining where new data-centre capacity is developed.

AI Growth Drives Power Demand

According to Wood Mackenzie, China’s data-centre expansion is being driven by both increasing AI workloads and government policy.

AI training requires increasingly energy-intensive computing infrastructure, while the rapid expansion of AI inference is creating a growing base of sustained electricity demand. These trends are expected to keep data-centre power consumption rising even as broader electricity demand growth in China moderates.

Government policy is also encouraging closer coordination between computing infrastructure and electricity supply. China’s 15th Five-Year Plan calls for the coordinated development of computing infrastructure and green electricity, while recent energy policies are promoting direct green power supply and more flexible computing loads.

Shift Towards ‘Compute Follows Power’

Historically, data centres have been concentrated around major cities and technology hubs to access customers, skilled workers, network infrastructure and low-latency connectivity.

However, the growing electricity requirements of AI are creating a shift towards a “compute follows power” model, particularly for workloads that can tolerate higher latency.

Renewable-rich regions in western China could attract energy-intensive AI training, batch processing and data storage workloads due to their availability of land and growing renewable power generation. Meanwhile, eastern demand centres are expected to remain important for latency-sensitive applications such as AI inference and financial services.

Wood Mackenzie expects eight national computing hubs to remain the dominant centres of China’s computing capacity through 2060. However, grid constraints and continued renewable energy expansion could gradually encourage additional data-centre development in regions with abundant renewable power.

Data Centres Could Become Flexible Grid Resources

The report also highlights the potential for data centres to play a more active role in China’s electricity system through battery storage, advanced power architectures and intelligent workload scheduling.

Non-urgent computing workloads could be shifted to periods when renewable generation is abundant or electricity prices are lower, while latency-sensitive workloads remain closer to major demand centres.

This could enable data centres to act as flexible electricity consumers and potentially absorb surplus renewable generation in regions experiencing periods of oversupply.

Wood Mackenzie expects data-centre demand for green electricity to increase significantly through 2060, while carbon emissions from data centres are forecast to peak by 2035 as the share of renewable power in the electricity mix rises.

Zhao said that unlocking greater synergy between computing and power will require the industry to address data security, service-level agreement requirements, operator participation and the high upfront capital expenditure associated with cross-regional migration and new infrastructure.

The report concludes that China’s AI expansion is increasingly linking computing infrastructure decisions with electricity availability, potentially reshaping the country’s data-centre and power-sector landscape over the coming decades.