Computing-Power-Electricity Synergy: SenseTime's SenseCore Offers a New Paradigm for “Carbon Reduction and Intelligence Enhancement” in AI Infrastructure

On October 30, a seminar titled "Policy Interpretation of Document No. 1192 and Seminar on Implementation Paths of Direct Green Power Connection and Computing-Power-Electricity Synergy" was held in Lingang, Shanghai. The seminar was hosted by the China Communications Industry Association Data Center Committee (CIDC) and co-organized by SenseTime, Shanghai Damao Technology, and China IDC Quan.

The event brought together representatives from government authorities, research institutes, energy enterprises, data center operators, and leading AI technology companies. They engaged in in-depth discussions on topics such as computing-power-electricity synergy, direct green power connections, and the high-quality development of computing power infrastructure. The discussions provided new ideas for policy implementation and industry collaboration and innovation.

The event was chaired by Jin Heping, Deputy Chief Engineer of China Three Gorges Corporation and Chairman of the CIDC. Participating organizations included the China Electric Power Planning & Engineering Institute, the China Three Gorges Shanghai Electric Power Planning Institute, the Shanghai Power Exchange Center, 21Vianet, UCloud and GDS Holdings.

Lin Hai, General Manager of Intelligent Computing Centers under SenseTime’s SenseCore Business Group, delivered a keynote speech, highlighting the innovative practices and explorations of the SenseTime Lingang AI Data Centre (AIDC) in achieving computing-power-electricity synergy.

From AI 1.0 to 2.0: Building a Trinity AI Infrastructure System

Lin Hai pointed out that in the 11 years since its establishment, SenseTime has evolved from the computer vision era of AI 1.0 to the generative intelligence era of AI 2.0. It has formed a strategic framework comprising "SenseCore (computing power) + Large Models + Industry Applications", forming a trinity that connects the entire process from computing power supply to model training and subsequent industry implementation. Since SenseTime built and put its AIDC into operation in Lingang in 2020, it has managed over 25,000P of computing power nationwide. The SenseTime Lingang AIDC was certified as the country’s first Grade 5A Intelligent Computing Center by the Ministry of Industry and Information Technology (MIIT) last year. This year, it has achieved an energy storage configuration of 36 MWh for 10,000P-scale intelligent computing power, setting an important benchmark for AI infrastructure and energy storage applications in China.

Driving Green Intelligent Computing Development through Computing-Power-Electricity Synergy Innovation

Focusing on "Computing-Power-Electricity Synergy", Lin Hai identified the key challenges facing current AI infrastructure, including high power costs, scarce energy consumption indicators, and limited capacity. He emphasized that the deep integration of computing power and electricity will be crucial for the green transformation of the AI industry.

In terms of technical pathways, SenseTime has built a computing-power-electricity synergy platform at the Lingang AIDC. This platform has established a full-chain collaborative system integrating "data-algorithm-strategy-execution". Utilizing self-developed energy large models and an algorithm architecture spanning from "server" to "energy block", it realizes closed-loop intelligent scheduling that encompasses "perception-prediction-decision-execution". The goal is to create a new paradigm for intelligent computing resource operations featuring networked scheduling, inclusive accessibility, enhanced sustainability and security. Currently, the accuracy rate of energy demand prediction based on the energy block model is nearly 90%, and the accuracy rate of scheduling decisions is about 95%. During high-load periods, the model can intelligently trigger adjustments in energy storage systems, achieving a dynamic balance of "electricity following computing power". This significantly improves the computing density and energy utilization efficiency of the park.

Promoting Industrial Co-construction and Open Cooperation

Lin Hai stated that moving forward, SenseTime will continue to advance the standardization and commercialization of the computing-power-electricity synergy platform, promote the combination model of "computing power + energy storage + platform" in multiple parks across the country, and create replicable templates for green intelligent computing infrastructure. Additionally, SenseTime will expand its platform capabilities, work with various industry stakeholders to explore new scenarios, such as direct green power connections, demand response, and market-oriented transactions, and accelerate the intelligent, low-carbon transformation of AI computing infrastructure.

According to the recently released 2025 China AI Infrastructure Market Report by leading research institutions Frost & Sullivan and the Head Leopard Research Institute, SenseTime has been recognized for its outstanding technological innovation capabilities and extensive industry practices, earning a place in the "Leaders Quadrant" and ranking among the top tier of China's AI infrastructure providers.

As a leading domestic AI infrastructure leader, SenseCore will continue to promote the implementation of green computing innovation and intelligent scheduling systems, support the high-quality, low-carbon, and intelligent development of national computing infrastructure, and work with industry partners to jointly build a sustainable AI infrastructure base.