As China moves to clean up its energy system, AI is beginning to determine how electricity is produced, moved, and used, not in abstract policy terms but in everyday operations.

A factory powered by renewable energy in the northern Chinese city of Chifeng provides a clear example. The facility produces hydrogen and ammonia using only electricity generated by nearby wind and solar farms. Unlike traditional plants, which are connected to the wider power grid, this facility operates on its own closed system. This setup comes with benefits as well as problems. Renewable electricity is clean, but the amount of electricity increases and decreases depending on the weather.

To maintain steady production, the factory relies on an AI-driven control system built by owner Envision. The software continuously adjusts output based on changes in wind and sunlight, rather than following a fixed schedule. According to reports ReutersJian Zhang, Envision’s lead engineer for hydrogen energy, compared the system to a conductor that adjusts power supply and industrial demand in real time.

As wind speed increases, production automatically increases to make the most of the available power. When the situation weakens, power usage will be immediately reduced to avoid strain. Zhang said the system allows the plant to operate with high efficiency despite fluctuations in renewable energy.

Such projects are central to China’s plans for hydrogen and ammonia, fuels seen as important for reducing emissions in sectors such as steel manufacturing and shipping. They also point to a broader strategy. The idea is to use AI to manage the complexity of adding renewable power to the nation’s power grid.

Researchers argue that AI could play a key role in meeting China’s climate goals. Ayana Zheng, an associate professor at Nanjing’s Southeast University who studies low-carbon transitions, said AI can support tasks ranging from emissions tracking to power supply and demand forecasting. At the same time, she warned that AI itself is causing a rapid increase in power consumption, especially through energy-intensive data centers.

China now has more wind and solar power installed than any other country, but absorbing that power efficiently remains a challenge. Cory Combs, associate director at Beijing-based research firm Trivium China, said AI is increasingly seen as a way to make the grid more flexible and responsive.

The idea was formalized in September, when the Chinese government introduced its “AI+Energy” strategy. The plan calls for stronger collaboration between AI systems and the energy sector, including the development of multiple large-scale AI models focused on power grid operations, power generation, and industrial applications. The government aims to roll out dozens of pilot projects and test AI in more than 100 use cases by 2027. Within three years, officials hope China will reach what they describe as a world leader in AI integration in the energy sector.

Combs said the company is focused on highly specialized tools designed for specific tasks, such as managing wind farms, nuclear power plants, and grid balancing, rather than general-purpose AI. Hu Guangzhou, a professor at the China-Europe International Business School in Shanghai, said this approach contrasts with the United States, where much of the investment is in building advanced large language models.

One area where AI could have an immediate impact is demand forecasting. Huang Lurui, assistant professor at Xi’an Jiaotong Liverpool University, said the electricity grid must always match supply and demand to avoid power outages. Accurately predicting renewable energy output and power usage allows operators to plan ahead, store energy in batteries when needed, and reduce reliance on coal-fired backup plants.

Experiments are already underway in some cities. The city of Shanghai has launched a city-wide virtual power plant that connects dozens of operators, including data centers, building systems and electric vehicle chargers, into a single, coordinated network. During testing last August, the system reduced peak demand by more than 160 megawatts. This is roughly equivalent to the output of a small coal-fired power plant.

Combs said such systems are important because modern power generation is increasingly distributed and intermittent. “We need something very robust that is predictable and can account for new information quickly,” he said.

Beyond the power grid, China is also considering applying AI to its domestic carbon market, which covers more than 3,000 companies in emissions-intensive industries such as power, steel, cement, and aluminum. Together, these sectors produce more than 60% of the country’s carbon emissions. Chen Jibin, a senior manager at Berlin-based think tank Adelphi, said AI could help regulators verify emissions data, refine free quota allocations and give companies clearer insight into production costs.

Still, risks are increasing as well as opportunities. Research shows that by 2030, China’s AI data centers could consume more than 1,000 terawatt-hours of electricity per year, which is about the same as Japan’s current annual usage. Lifecycle emissions from the AI ​​sector are expected to increase rapidly and peak well past China’s 2030 emissions target.

Xiong Qiang, a postdoctoral fellow at Renmin University of China who worked on one of the studies, said the results reflect the reality that coal still makes up the majority of China’s electricity mix. He warned that the rapid expansion of AI could complicate national climate goals if energy sources do not shift quickly enough.

In response, regulators began tightening regulations. The 2024 action plan calls for data centers to improve energy efficiency and increase their use of renewable electricity by 10% each year. Other efforts are encouraging new facilities to be built in western regions where wind and solar resources are more abundant.

East Coast carriers are also testing new ideas. An undersea data center will be opened near Shanghai that will use seawater for cooling to reduce energy and water usage. Developer Hailanyun said the facility will draw most of its electricity from an offshore wind farm and could be replicated if the project proves viable.

Despite the increasing energy demand from AI, Xiong argued that if applied carefully, the overall impact of AI on emissions can still be positive. He said AI is being used to optimize heavy industry, power systems and carbon markets, and could continue to be an integral part of China’s emissions reduction efforts, even as it creates new pressures that policymakers must manage.

(Photo courtesy of Matthew Henry)

See also: Can China’s chip stacking strategy really challenge Nvidia’s AI dominance?

Want to learn more about AI and big data from industry leaders? Check out the AI ​​& Big Data Expo in Amsterdam, California, and London. This comprehensive event is part of TechEx and co-located with other major technology events. Click here for more information.

AI News is brought to you by TechForge Media. Learn about other upcoming enterprise technology events and webinars.