China advances clean energy transition

The report also detailed multiple technological breakthroughs.

In the realm of long-duration energy storage, large-scale vanadium flow battery stations have entered operation, and the efficiency of compressed air energy storage systems has surpassed 70 percent.

New business models such as virtual power plants and distributed energy have mushroomed. The number of electric vehicle charging infrastructure units reached 20.09 million in 2025, an increase of almost 50 percent year-on-year.

Beyond its borders, China's unparalleled manufacturing scale has been a boon to global decarbonization, coupled with expansive direct investments. In traditional and new energy cooperation alike, China has maintained a dual-wheel drive approach.

Figures released by the EPPEI reveal that over the past decade, Chinese overseas investments in thermal power exceeded 60 million kW of capacity with a total investment of over $30 billion.

From 2022 to 2025, China participated in more than 500 overseas new energy projects across 140 countries and regions, boasting an overseas solar photovoltaic capacity of over 100 million kW and wind power capacity of over 30 million kW, the EPPEI said.

Crucially, the country's cumulative solar PV module exports have exceeded $180 billion, covering more than 200 countries and regions with an overseas installed capacity of over 1.2 billion kW.

Chinese solar technology giant Longi recently signed a landmark agreement with Romanian developer Enexus for 50 megawatt-hours of liquid-cooled battery energy storage systems.

The deal, marking one of Longi's largest storage deployments in Europe, builds upon more than 200 megawatts of the company's solar modules deployed across Enexus projects since 2025. It underscores a strategic shift for Chinese enterprises: moving from traditional module supply to providing comprehensive, utility-scale technology collaborations.

"The next phase of the global energy transition demands a total departure from fragmented, multivendor systems," said Dennis She, vice-president of Longi and president of Longi Energy Storage.

As Romania pushes to add approximately 5.8 gigawatts of ground-mounted solar capacity by 2030, these integrated systems will allow local grids to shift surplus daytime electricity to peak demand periods. By mitigating solar variability and reducing power curtailment, the Chinese-supplied hybrid infrastructure directly bolsters the resilience of Romania's evolving power network.

Meanwhile, the energy landscape is undergoing another seismic shift, driven by the explosive growth of artificial intelligence.

The report noted that AI and the energy sector are increasingly converging, with energy systems becoming a critical testbed for AI applications, while at the same time, the massive power demands of AI computing are reshaping energy infrastructure planning.

"The focus of AI development is shifting from foundational model breakthroughs to large-scale applications and actual value creation," said Li Zhenjie, deputy chief engineer of the energy science and technology innovation research institute at the EPPEI.

"With its long production chain, rich specialized scenarios and deep data accumulation, alongside high requirements for safety, reliability and real-time decision-making, the energy industry is becoming a vital arena for AI."

The NEA released the first batch of 51 high-value "AI + Energy" scenarios in May, covering eight major areas including power grids, new energy and emerging business models. Energy enterprises are now rapidly deploying specialized large language models to optimize operations.

For instance, AI is being utilized in the hydro-wind-solar integrated management of the Yalong River basin. By coupling Huawei's Pangu weather model with local meteorological and hydrological data, operators have significantly improved power generation forecasting accuracy, overcoming traditional difficulties in modeling complex weather patterns over high-altitude regions.

In thermal power, AI is assisting in dynamic combustion optimization and flexible peaking adjustments, allowing legacy coal plants to respond in seconds to balance the grid when intermittent solar and wind power fluctuates.

In new business models, AI acts as a crucial key to breaking bottlenecks in resource aggregation and process optimization.

Through precise algorithms, AI enables the smart control of virtual power plants, intelligent optimization of green hydrogen production processes, smart carbon reduction in industrial parks and the collaborative scheduling of new-type energy storage systems, said Li.

Yet, as AI optimizes the grid, it also demands vast amounts of electricity, he said.

China's intelligent computing infrastructure is in a rapid expansion phase. By the end of June, the nation's intelligent computing scale marked a 177 percent year-on-year surge, with 52 mega-clusters comprising tens of thousands of graphics processing units already built.

This computational boom has translated directly into new power loads. In 2025, China's computing centers consumed 170 billion kWh of electricity, accounting for 1.6 percent of the total societal power consumption. By 2030, nationwide electricity consumption for computing power is projected to skyrocket to 800 billion kWh, claiming roughly 6 percent of national electricity usage, Li said.

"Future global competition in AI will not merely be a contest of computing scale or algorithmic efficiency; it will increasingly depend on the ability to acquire large-scale power in a low-carbon, highly economical and sustainable manner," he said.

According to Li, this immense, stable green energy capacity provides a critical advantage for China's expanding computing centers, allowing them to meet surging electricity demands without derailing national carbon reduction goals.

"Energy is no longer just the bedrock for AI, and AI is no longer merely a tool for the energy sector," he said. "The two are moving from mutual support toward deep coupling, interacting through real-time data, dispatch mechanisms and operational workflows to forge a unified, intelligent and green infrastructure for the future."

zhengxin@chinadaily.com.cn