AI Reshapes Energy and Mining Sectors (AI Translation)

【周刊提前读】人工智能重构能源矿业 企业应用仍在早期

At the start of 2025, generative artificial intelligence (AI) technologies such as DeepSeek attracted significant attention among leaders in China’s energy and resources sector. Major companies, including Zijin Mining and China Gas, highlighted AI's transformative potential at annual work meetings, noting that open-source, cost-effective generative models could satisfy industry-specific needs. Executives voiced the need for deeper and broader investment into AI applications, foreseeing industry-wide disruption [para. 1].

Central government involvement accelerated with the State-owned Assets Supervision and Administration Commission (SASAC), launching AI+ initiatives. By early 2025, over 66 large AI models were deployed across central SOEs in oil, chemicals, grids, coal, and metals, with AI utilized in more than 500 scenarios in energy and industrial manufacturing. Consultancy McKinsey identified energy, oil & gas, and mining as sectors inherently favoring AI due to heavy reliance on data-driven innovation. Experts posited that AI could become a new core driver of industrial transformation, integrating deeply into operations from production and safety to logistics and planning [para. 2][para. 3].

The initial applications centered on optimizing resource supply, power generation, and consumption. Energy consulting leaders see AI as focusing on four main pillars: growth, transformation, efficiency, and cost. Still, the path to full disruption may be gradual. AI in energy is split between mature traditional applications—such as image recognition for predictive maintenance—and emerging generative AI scenarios, including knowledge bases, multilingual customer service, and safety-critical verticals. Future visions include shifting from experience-driven to data-intelligent approaches, enabled by models like DeepSeek lowering adoption barriers [para. 4][para. 5][para. 6].

As China’s grid becomes more decentralized and renewables increase their share (with Southern Grid serving 272 million across five provinces and renewables approaching 40% of total installed capacity), AI is critical for real-time grid balancing and equipment monitoring. Against this backdrop, energy sector digitalization and AI adoption have accelerated. However, leaders remain sober, acknowledging challenges with model reliability, explainability, and risks to employment as automation advances [para. 7][para. 8][para. 9].

Domestic case studies abound: China Gas began large-scale AI deployment in 2020, initially for gas load forecasting, now extending into safety, customer service, and contract management—achieving a 90% approval rate in AI-based workflow and reducing manual review costs by 70%. Southern Grid uses AI for load forecasting (achieving 98.4% accuracy) and fault detection. Mining giants like BHP employ AI for resource exploration, operation optimization, and supply chain improvements, exemplified by the Oak Dam copper discovery attributed to AI-aided analysis and Machine Learning [para. 11][para. 12][para. 13][para. 14][para. 15][para. 16].

Nevertheless, the energy sector’s adoption pace is checked by high expectations for return on investment (ROI), data security concerns, technical complexity, and a preference for private, domain-specific models. Cost and reliable, high-quality data remain significant hurdles. Companies are building proprietary AI infrastructure; for instance, Shandong Energy Group invested around 150 million RMB in a fully privatized AI training center to ensure data security [para. 17][para. 18][para. 19][para. 20].

Scalable AI applications flourish where tasks are well-defined, data is abundant, and industry barriers are modest. In coal mining, AI-enhanced visual inspection for safety checks, such as rock burst prevention, has improved work efficiency by 85%. However, challenges remain in highly specialized fields (e.g., chemicals), where diversity of processes and tight tolerances sharply limit off-the-shelf AI solutions [para. 21][para. 22].

Talent and organizational capacity are viewed as crucial bottlenecks, with a dire need for cross-disciplinary expertise spanning operational knowledge and AI fluency. Leaders stress the importance of building robust internal structures and developing AI deployment “roadmaps,” as only 48% of surveyed executives felt prepared for technological disruption, while 82% saw deep AI understanding as essential for future leadership [para. 23][para. 24][para. 25].

In sum, though still in the early stages, AI is percolating through China’s energy and resources sectors—primarily as a catalyst for efficiency and incremental innovation. The most disruptive potential lies in physical-world optimization (resource exploration, manufacturing) and rapid material discovery. Policy, data quality, security, and workforce development are essential for scaling to industry-wide transformation [para. 26][para. 27].