China Telecom – Utilising AI to Drive Energy Conservation for Enterprises 

Energy efficiency is central to China Telecom’s carbon targets. In 2025, AI-enabled energy-saving measures cut greenhouse gas emissions by over 16 million tonnes (annual emissions: 13.73 million tonnes) and reduced emissions per unit of telecommunications business volume by 13% year-on-year.

 China Telecom built a unified, cloud–edge intelligent energy-saving platform for cloud-network infrastructure using big data, AI, cloud computing, and large-model technologies. Deployed across 4/5G base stations, communication rooms, data centres, and cloud hosts, it monitors operations and environmental data in real time to shut down or start idle equipment and optimise cooling-system fan speeds for precise energy savings. At the China Telecom Central Intelligent Computing Center, an AI load-response model dynamically adjusted water pump frequency, reducing PUE (Power Usage Effectiveness) by 0.06 to 1.33, saving 712 MWh year-on-year, cutting electricity costs by 498,000 RMB annually, and reducing carbon emissions by 565 tonnes. 

By the end of 2025, the platform managed over 6 million 4/5G base-station sectors and more than 9,000 communication equipment rooms. The daily comprehensive energy-saving efficiency for 5G exceeded 16%, delivering annual savings of over 1 billion kWh and reducing CO₂ emissions by over 500,000 tons. 

Because its network equipment and services are highly complex, China Telecom had to address challenges in connecting specialised equipment, platforms, and business data interfaces, while also enabling unified digital control. To manage complex network services that are difficult and time-intensive to operate and carry high energy-saving risks, the company combined AI strategies with neural-network learning to deliver dynamic, adaptive energy optimisation. 

China Telecom has established a unified AI energy-saving system at scale that is safe, reliable, and provides full energy-consumption visibility. With low deployment and operating costs and high efficiency, it addresses the problem of network-level energy saving and enables large-scale engineering deployment of AI for energy conservation.