State Grid Corporation of China AI Orchestration Platform Coordinates 2.03 GW of Dispatchable Capacity in Shanghai

State Grid Corporation of China's city-scale AI orchestration platform for Shanghai's power grid coordinates 2.03 GW of dispatchable capacity across 49 virtual power plant operators and more than 4,000 users, according to the World Economic Forum's MINDS case study. The case study reports more than RMB 8 billion saved through avoided generation and transmission investments, 1.236 billion kWh of clean energy consumed annually with 510,000 tonnes of CO2 reduced per year, 99.999% voltage compliance and 99.9983% grid reliability. It also reports improvements of 12.5% in forecasting accuracy, 72% in regulation response and 51% in settlement efficiency, and over 15,000 active users across grid operations and virtual power plant teams. Shanghai's grid serves more than 1.7 million corporate users and 10.6 million residential users. Rising electricity demand, carbon reduction targets and a growing share of variable renewable energy add to operational complexity. Manual coordination is too slow for second-level decisions in response to renewable fluctuations, while expanding thermal generation or storage capacity entails high costs and additional carbon impact. The platform integrates forecasting, trading, regulation and settlement into a single operational system. It is built on the Guangming Large Model, a proprietary multimodal large language model trained on more than 200 terabytes of power system data and a domain-specific token library of 3.5 million energy-related terms. It continuously predicts and optimizes power flows using real-time data from edge gateways, enterprise systems and grid sensors. The system combines spatiotemporal modelling, Model Predictive Control and NSGA-II optimisation to support adaptive control with second-level response times. It runs on an on-premises architecture with edge-cloud coordination, allowing real-time orchestration of thousands of distributed energy devices such as air conditioners, electric vehicle chargers and industrial equipment. Multiple intelligent agents manage forecasting, trading, regulation and settlement within a unified workflow, supported by human-in-the-loop oversight through operational dashboards. The architecture has enabled applications such as reactive power demand response in Shanghai's Lin-gang industrial zone, helping stabilize voltage levels without requiring additional infrastructure investment.