人工智能的能源悖论：平衡挑战与机遇

Transformation of Industries in the Age of AI Artificial Intelligence’sEnergy Paradox:BalancingChallenges and Opportunities W H I T EP A P E RJ A N U A R Y2 0 2 5 Contents Reading guide3Foreword4Executive summary5Introduction61Electricity consumption of AI71.1The AI life cycle71.2The role of data centres81.3Opportunities to reduce AI system electricity consumption92AI-enabled energy transition112.1Non-exhaustive example opportunities for AI-enabled11electricity reduction2.2Sample use cases123Primary challenges and ecosystem enablers143.1 Infrastructure challenges143.2 Environmental challenges143.3 Overview of ecosystem enablers153.4 Regulatory and policy enablers163.5 Financial incentive enablers163.6Technological innovation enablers173.7Market development enablers174Future outlook of AI energy impact184.1 The deployment and collaboration landscape184.2 AI and energy – 2024 to 2025 outlook22Conclusion23Contributors24Endnotes26 Disclaimer This document is published by the World Economic Forumas a contribution to a project, insight area or interaction. Thefindings, interpretations and conclusions expressed herein area result of a collaborative process facilitated and endorsed bythe World Economic Forum but whose results do not necessarilyrepresent the views of the World Economic Forum, nor theentirety of its Members, Partners or other stakeholders. © 2025 World Economic Forum. All rights reserved. No partof this publication may be reproduced or transmitted in anyform or by any means, including photocopying and recording,or by any information storage and retrieval system. Reading guide The World Economic Forum’s AI Transformation ofIndustries initiative seeks to catalyse responsibleindustry transformation by exploring the strategicimplications, opportunities and challenges ofpromoting artificial intelligence (AI)-driven innovationacross business and operating models. This white paper series explores the transformativerole of AI across industries. It provides insightsthrough both broad analyses and in-depthexplorations of industry-specific and regional deepdives. The series includes: Cross industry Regional specific Impact on industrial ecosystems Impact on regions Blueprint to Action:China’s Path toAI-Powered IndustryTransformation LeveragingGenerative AI for JobAugmentationandWorkforce Productivity Artificial Intelligence’sEnergy Paradox:Balancing Challengesand Opportunities Artificial Intelligenceand Cybersecurity:Balancing Risksand Rewards AI in Action: BeyondExperimentation toTransform Industry Industry or function specific Impact on industries, sectors and functions Advancedmanufacturingand supply chains Media,entertainmentand sport Financialservices Healthcare Transport Frontier Technologiesin IndustrialOperations: TheRise of ArtificialIntelligence Agents Intelligent Transport,Greener Future:AI as a Catalystto DecarbonizeGlobal Logistics Artificial Intelligence inMedia, Entertainmentand Sport The Future ofAI-Enabled Health:Leading the Way Upcomingindustry report:Telecommunications Artificial Intelligencein Financial Services As AI continues to evolve at an unprecedentedpace, each paper in this series captures a uniqueperspective on AI – including a detailed snapshotof the landscape at the time of writing. Recognizingthat ongoing shifts and advancements are alreadyin motion, the aim is to continuously deepen andupdate the understanding of AI’s implicationsand applications through collaboration with thecommunity of World Economic Forum partners and stakeholders engaged in AI strategy andimplementation across organizations. Together, these papers offer a comprehensiveview of AI’s current development and adoption,as well as a view of its future potential impact.Each paper can be read stand-alone or alongsidethe others, with common themes emergingacross industries. Foreword Roberto BoccaHead, Centre for Energyand Materials; Member,Executive Committee,World Economic Forum Jeremy JurgensManaging Director, WorldEconomic Forum Cathy Li Head, AI, Data and Metaverse;Deputy Head, Centrefor the Fourth IndustrialRevolution; Member,Executive Committee,World Economic Forum James MazurekManaging Director, US UtilitiesStrategy Lead, Accenture In today’s economy, artificial intelligence (AI) systemsoffer both challenges and opportunities. As integralcomponents of digital infrastructure, the data centresthat enable AI support a variety of applications,from cloud computing to complex data processing.AI’srapid expansion, however, is accompanied bygrowing electricitydemand, with the largest facilitiesin the world using the same amount of power assmall cities to ensure uninterrupted operation. Datacentres come in varying sizes however, ranging fromlarge, hyperscale facilities with more than 1 gigawatt(GW) of power capacity, to smaller, micro edgedeployments that may draw less than 10 kilowatts(kW) of power.1 achievement of efficiency gains. To achieve this,it’s pivotal to understand in