人工智能能源使用的环境成本

Carbon, Water and Land Footprints Acknowledgment UNU-INWEH gratefully acknowledges its host, the Government of Canada, and ongoingfinancial support from Global Affairs Canada. About UNU-INWEH The United Nations University Institute forWater, Environment and Health (UNU-INWEH) is one of 13 institutions comprisingthe United Nations University (UNU),the academic arm of the United Nations. Redistribution This publication may be copied, adapted,reproduced in whole or in part and inany form for educational or non-profitservices without special permission UNU-INWEH specializes in addressingcritical global security and developmentchallenges at the intersection of water,environment, and health. Through research,capacity development, policy engagement,and knowledge dissemination, the institutebridges the gap between scientific evidenceand the practical needs of policymakers andUN member states, with particular attention Disclaimer UNU-INWEH enjoys the generous support ofGlobal Affairs Canada. As a United Nationsentity, UNU-INWEH operates independently.The content of UNU-INWEH publicationsis grounded in scientific evidence and is Environmental Cost of AI's Energy UseCarbon, Water and Land Footprints United Nations University Institute for Water, Environment and Health (UNU-INWEH)Richmond Hill, Ontario, Canada, 2026 © United Nations University Institute forWater, Environment and Health (UNU- © Maps, photos, and illustrations asspecified Authors:Miriam Aczel, Sanaz Chamanara,Mir Matin, Aria Farsi, Tshilidzi Marwala,Kaveh Madani How to cite:Aczel M., Chamanara S., MatinM., Farsi A., Marwala T., Madani K. (2026)Environmental Cost of AI's Energy Use:Carbon, Water and Land Footprints, UnitedNations University Institute for Water,Environment and Health (UNU-INWEH), Contributors:Luying Wang, Jenna Elshurafa Front cover:A row of servers in Google'sDouglas County, Georgia, data center. ISBN:978-92-808-6141-9DOI:10.53328/INR26RMA002Download at:inweh.unu.edu/publications CONTENTS CONTENTSFIGURES 1. AI ACROSS THE GLOBE 1.1 A Technological Explosion1.2 The Expanding Reach of AI Across Sectors1.3 The Global Market for AI1.4 Impacts on the Global Labor Market1.5 Growing Concerns about AI 18 2. AI's Growing Energy Demand 2.1 Energy Intensity of AI Training2.2 AI's Energy Demand and its Environmental Footprints2.3 AI's Training Footprints2.4 Data Centers and AI's Expanding Energy Demand2.5 AI's Contributions to Data Centers' Energy Use 3. AI in Use: Tools and Tasks 3.1 Two Forces Shaping AI’s Operational Footprint3.2 AI Tasks and their Footprints3.3 AI Use at Scale3.4 Conventional Search Versus Enhanced Search3.5 Efficiency Improvement and Rebound Effects 4. The Way Forward 4.1 Guiding Principles for a Responsible AI Ecosystem4.2 From Measurement to Action4.3 Roles and Responsibilities4.4 Putting Principles into Practice FIGURES Figure 1.The evolution of artificial intelligenceFigure 2.Industrial revolutions from mechanization to intelligent systemsFigure 3.Key breakthroughs in deep learningFigure 4.Global electricity supply by sourceFigure 5.Electricity supply mixes in major data center host locationsFigure 6.Global electricity footprintsFigure 7.Footprints of electricity in the world's top 20 data center hubsFigure 8.Global distribution of data center locationsFigure 9.Global electricity footprint intensities across the world’s top 20 datacenter hubs relative to the global averageFigure 10.Energy demand in AI arises from two distinct phases ABBREVIATIONS 4IRFourth Industrial RevolutionAIArtificial IntelligenceCO2eCarbon dioxide equivalentCPUCentral Processing UnitGPTGenerative Pre-trained Transformer GPUGraphics Processing UnitGWGigawatt, 1,000,000,000 WGWhGigawatt-hour, 1,000 MWhHPCHigh-Performance ComputingIoTInternet of ThingskWKilowatt, 1,000 WkWhKilowatt-hour, 1,000 WhLLMLarge Language Model MoEMixture-of-ExpertsT2VText-to-VideoTPUTensor Processing Unit TWhTerawatt-hour, 1,000 GWhUSDUnited States DollarWWatt, powerWhWatt-hour, energy Key Points in Brief AI’s growth The global AI market is expanding rapidly, projected to grow from USD 189billion in 2023 to nearly USD 5 trillion by 2033. This would represent roughly a 25-fold increase in global AI market size over a Global AI expenditure is projected to exceed USD 2.5 trillion in 2026. Generative AI accounted for over 20% of the global AI market in 2026, and is Corporate AI investment exceeded USD 580 billion in 2025, while generative AIalone attracted nearly USD 34 billion in private investment. 78% of organizations reported using AI in their work in 2024, and 40% ofemployers expect to reduce their workforce where AI can automate tasks. About 60% of jobs in advanced economies incorporate AI, versus 26% in low- Nearly half of the world’s data centers are in the United States. Only 16% of countries host AI-specialized cloud compute, and 90% of thatcapacity is concentrated in just 2 countries (the United States and China). Training foo