How AI Is Solving Problems in China’s Energy Industry
By Luo Guoping, Fan Ruohong and Wang Xintong
In 2020, China Gas Holdings Ltd. turned to artificial intelligence (AI) to solve a persistent industry problem: accurately predicting demand for natural gas.
The state-owned giant, which supplies natural gas to more than 200 million urban residents nationwide, sought to replace forecasts made by humans with AI models that analyze both historical data and real-time variables — including the weather, pricing, and geopolitical and economic shifts, said Han Peng, general manager of China Gas’ digital and AI department.
The shift paid off. AI-driven forecasts helped the company slash how much it spends on stop-gap gas purchases. These “off-contract” purchases are how companies like China Gas meet demand beyond the supplies they secure through pre-negotiated contracts, which typically cover their predicted needs for the coming year. The company was able to reduce these purchases by 50%, saving tens of millions of yuan, according to internal estimates.
China Gas didn’t stop there. It has since introduced virtual assistants to handle customer questions, employed AI to streamline contract reviews, and deployed AI-powered cameras to enforce safety helmet rules. In addition, its predictive AI-powered maintenance systems, which monitor equipment 24/7, have reduced failure rates by 30%.
China Gas exemplifies a broader AI-driven transformation that is starting to reshape China’s energy sector. From oil and gas to power grids and mining, companies are accelerating AI adoption, using the technology to save costs, boost efficiency and enhance safety in production.
The trend illustrates Beijing’s push to harness the cutting-edge technology to boost productivity gains amid stubborn economic headwinds and an intensifying tech rivalry with the U.S.
What’s driving the change
AI is expected to transform the energy sector into an industry powered by data, paving the way for “future energy systems,” according to sources from China Gas and China Southern Power Grid Co. Ltd., one of the country’s two main electricity grid operators.
They credited Hangzhou-based startup DeepSeek’s open-source AI models with significantly lowering the barriers to adopting the technology. They added that companies are using the opportunity to upgrade systems and strengthen competitiveness.
AI adoption gained fresh momentum in January when Deepseek released RI, an open-source large language model that rivals the performance of leading global models at only a fraction of their cost. The release made AI tools more accessible to businesses of all kinds.
At a February meeting, the State-owned Assets Supervision and Administration Commission (SASAC), which oversees state-owned enterprises (SOEs), announced plans to intensify an initiative launched last year that aims for SOEs to play a bigger role in AI development and adoption in China.
A number of state-owned firms, including China Gas and Southern Power Grid, have announced that they have incorporated Deepseek’s models into their operations. By late March, SOEs had applied AI in over 500 operational scenarios, SASAC data showed.
The AI transition is inevitable, Han said. “The only thing we can do is quickly integrate it into our work, actively invest in and embrace this new era,” he told Caixin in a February interview.
Using AI in power and mines
As the world shifts toward green energy, China is integrating more renewable sources like solar and wind into its power grid. However, the variability of these sources poses a challenge.
AI is key to addressing the issue, a Southern Power Grid executive told Caixin, explaining that the technology can balance electricity supply and demand in real time while improving grid reliability through prompt detection of equipment problems.
Southern Power Grid serves 272 million people across five provincial-level regions — Guangdong, Guangxi, Yunnan, Guizhou and Hainan — with renewables making up nearly 40% of its total installed generation capacity.
Since 2020, Southern Power Grid has been using AI tools to assess whether electricity supply meets demand, check grid safety, and develop operational strategies. By 2024, the AI system had enabled planners to work 10 times faster while achieving 98.4% prediction accuracy. The technology also identifies over 90% of typical equipment defects, improving inspection efficiency by a factor of 80.
The mining sector is also adopting advanced technologies to streamline operations.
One such case is CMOC Group Ltd.’s (603993.SH +1.45%) molybdenum mine in Luoyang, Central China’s Henan province. Its complex, uneven ore deposits once forced workers to manually calculate shipments and track ore data on spreadsheets to meet processing needs, said Jia Baoshan, a manager at a CMOC-backed mining firm in Luoyang.
Now, the company uses drone surveys and image recognition tools to analyze ore distribution, with an AI platform helping to reduce certain kinds of errors, Jia said.
Reality check
Sectors like oil and gas, power, mining and materials are “uniquely positioned” to benefit from AI because their reliance on data and analytics for innovation makes them ideal for AI-driven growth and cost cutting, consultancy McKinsey & Co. said in a February analysis.
Still, while there has been rapid adoption of AI in the energy industry, it’s still some way off from causing a true industry disruption, said Lü Yan, a managing partner at consultancy Deloitte China.
The limitations of AI are especially clear in high-risk mining environments, said Wang Danshi, secretary-general of the informationization branch of the China National Coal Association. “For coal mines, the most dangerous processes should certainly be replaced by AI in the future, but these are also the hardest to replace,” he said.
He also warned that for SOEs, rapid adoption of AI might reduce or eliminate certain jobs, potentially hurting the job market in a way that necessitates a prudent approach.
Contact reporter Wang Xintong (xintongwang@caixin.com) and editor Michael Bellart (michaelbellart@caixin.com)
caixinglobal.com is the English-language online news portal of Chinese financial and business news media group Caixin. Global Neighbours is authorized to reprint this article.
Image: Firman Dasmir – stock.adobe.com