Caixin Weekly | The AI+ Era in the Energy Sector is Underway (AI Translation)

财新周刊｜能源业“AI+”进行时

The article provides a comprehensive overview of the rapid integration and challenges of artificial intelligence (AI) in China’s energy and resources sectors as of early 2025, fueled by breakthroughs such as DeepSeek’s generative AI and the national "AI+" initiative for state-owned enterprises (SOEs) [para. 1][para. 2]. Senior executives from leading companies like Zijin Mining and China Gas recognize the transformative potential of AI, citing its intelligence, open-source nature, and low cost as key to meeting industry-specific needs [para. 1]. Nationally, at least 66 major AI models have been deployed by dozens of SOEs across core sectors such as oil, gas, electricity, coal, and metals, with more than 500 application scenarios identified [para. 1]. Consulting giant McKinsey identifies the sector’s data-rich environment as advantageous for AI-driven innovation, expected to enhance growth and reduce costs [para. 1].

AI deployment is accelerating in industry, notably in safety, efficiency, and cost control [para. 1][para. 2][para. 3]. In energy, traditional AI—such as for image recognition in inspections and predictive maintenance—is already mature, while generative AI applications are emerging in both general and industry-specific scenarios [para. 3]. Companies such as China Southern Power Grid and China Gas expect AI to shift operations from experience-based to data-driven systems, lowering thresholds for adoption and opening the path to a more competitive "energy system of the future" [para. 3].

AI’s impact in energy is most pronounced in load prediction, procurement optimization, intelligent customer service, defect identification, and predictive equipment maintenance [para. 15][para. 17][para. 18][para. 19]. For example, China Gas uses AI to predict gas load with greater accuracy, automate contract approvals (with a 90% approval rate), and reduce manual review costs by 70% through image recognition [para. 17]. Predictive maintenance with AI has lowered equipment failure rates by 30% [para. 17]. Southern Power Grid fully automated its load forecasting, boosting work efficiency tenfold and achieving an average prediction accuracy of 98.4% [para. 19].

In mining, AI is revolutionizing ore matching, equipment scheduling, and resource exploration. For instance, Zijin Mining has reduced ore blending deviation from 11–15% to under 3% through digital platforms [para. 22]. Global giant BHP employs AI across exploration, mining, and supply chain optimization—enabling the discovery of deeper copper deposits, such as at Oak Dam, nearly doubling exploration depth capabilities [para. 23][para. 24].

Despite progress, barriers persist: enterprises remain cautious on large-scale and core business AI adoption due to high reliability and safety demands, the "black box" nature of large models, data privacy/security, cost, and the need for proprietary "small models" tailored to vertical industries [para. 29][para. 30][para. 32][para. 33]. AI tends to find success in well-defined scenarios where data is reliable and standardized, with large models still expensive and requiring further maturity before widespread, affordable commercialization—projected in about five years [para. 34][para. 36].

Human capital is another critical bottleneck. Industry-wide shortages of interdisciplinary AI talent and business-technology integration hinder deep adoption, and high-quality, well-governed data is still lacking [para. 68][para. 69]. Companies are investing in dedicated teams and bridging business and technology expertise to advance adoption [para. 69][para. 70].

Looking ahead, AI has the potential to produce disruptive innovations in physically inaccessible processes like exploration, and in process-intensive industries like chemicals, potentially shortening R&D cycles for new materials [para. 80]. However, substantial policy support, standardized data management, and a well-developed talent pipeline are vital to harness AI’s full promise in China’s energy and resources sectors [para. 74][para. 75][para. 76][para. 77].