Global EnergyMap

About Global EnergyMap

Global EnergyMap addresses a critical challenge: industrial activities tied to energy combustion are a major driver of global carbon emissions, and while research focus has shifted from regional to asset-scale analysis, existing asset-level inventories remain incomplete—with most entries offering only vague location data, hindering accurate monitoring and low-carbon retrofitting at the asset level. To tackle this, Global EnergyMap leverages an automated AI Agent system integrating Vision-Language Models (VLMs) and Large Language Models (LLMs). Through a top-down systematic remote sensing approach, this AI Agent locates and verifies the precise geographic coordinates of 11,761 energy-intensive industrial plants worldwide, including 4,551 coal-fired power plants, 4,068 integrated cement plants, 1,206 oil refineries, and 1,936 steel plants, using remote sensing imagery. By harnessing multi-source remote sensing data and public information, Global EnergyMap delivers, for the first time globally, a comprehensive analysis of new construction, monthly operations, decommissioning, and low-carbon retrofitting activities of these facilities from 2013 to 2024. It introduces a novel AI-driven research paradigm to resolve this scientific challenge, with outcomes that will accelerate the precise achievement of Sustainable Development Goals (SDGs) at the asset level. Developed by the Department of Earth System Science at Tsinghua University, Global EnergyMap embodies cutting-edge interdisciplinary research at the intersection of remote sensing, artificial intelligence, and climate science.