能源与人工智能2025

INTERNATIONAL ENERGYAGENCY The IEA examines thefull spectrumof energy issuesincluding oil, gas andcoal supply anddemand, renewableenergy technologies,electricity markets,energy efficiency,access to energy,demand sidemanagement andmuch more. Throughits work, the IEAadvocates policiesthat will enhance thereliability, affordabilityand sustainability ofenergy in its32Member countries,13Associationcountries andbeyond. IEAMembercountries: IEAAssociationcountries: AustraliaAustriaBelgiumCanadaCzech RepublicDenmarkEstoniaFinlandFranceGermanyGreeceHungaryIrelandItalyJapanKoreaLatviaLithuaniaLuxembourgMexicoNetherlandsNew ZealandNorwayPolandPortugalSlovak RepublicSpainSwedenSwitzerlandRepublic of TürkiyeUnited KingdomUnited States ArgentinaBrazilChinaEgyptIndiaIndonesiaKenyaMoroccoSenegalSingaporeSouth AfricaThailandUkraine Please note that thispublication is subject tospecific restrictions that limitits use and distribution. Theterms and conditions areavailable online atwww.iea.org/terms This publication and anymap included herein arewithout prejudice to thestatus of or sovereignty overany territory, to thedelimitation of internationalfrontiers and boundaries andto the name of any territory,city or area. The EuropeanCommission alsoparticipates in thework of the IEA Source: IEA.International Energy AgencyWebsite: www.iea.org In recent years, artificial intelligence (AI) has soared to the top of the political and businessagenda. Once a mostly academic pursuit, it has evolved into an industry with trillions ofdollars at stake. Despite significant uncertainties, it is now very clear: AI is coming. In manysectors, it is already here. This has major consequences for the global energy sector. There is no AI without energy –specifically electricity. At the same time, AI has the potential to transform the sector’s future.However, policy makers and the market have often lacked the tools to fully understand thesewide-ranging impacts. Recognising this gap, the International Energy Agency (IEA) steppedup to address it by leveraging our expertise in data collection and analysis, as well as ourconvening power, to inform and strengthen the global dialogue on these issues. We began a new workstream on the nexus of energy and AI over a year ago, which hasresulted in a series of key activities and outputs, culminating in this special report. InDecember 2024, we held the Global Conference on Energy and AI, the largest internationalgathering on the matter to date, at our headquarters in Paris. It brought together policymakers, the tech sector, the energy industry and international experts to discuss the criticalissues at play. This helped lay groundwork for the AI Action Summit, co-chaired by PresidentEmmanuel Macron of France and Prime Minister Narendra Modi of India, in February 2025– an event to which the IEA made crucial contributions. This special report advances the conversation further. It is the first comprehensive globalanalysis examining all aspects of the links between energy and AI – from pathways to securelyand sustainably meeting energy demand for AI, to how AI itself could transform theproduction, consumption and transport of energy around the world. The analysis exploresthe implications of the rise of AI on energy security, investment, emissions and more –providing a strong factual basis for those thinking through the challenges and opportunitiesahead. This report shows that electricity demand for AI is growing fast globally, even if other sourcesof demand are growing faster. In some parts of the world, the effects of AI on electricitysystems are set to be very significant. With this in mind, we suggest three pillars countriesshould bear in mind as they plan for the future. The first is the importance of finding the right mix of energy sources to deliver theuninterrupted power supply that data centres need to support AI. According to our analysis,there is a role for established technologies such as renewables and natural gas, as well asemergingtechnologies like small modular nuclear reactors(SMRs)and advancedgeothermal. Deciding which options to prioritise may depend on other policy priorities. Yet a sole focus on increasing electricity generation won’t be enough. To deliver the energyfor AI, countries must also think about their infrastructure. That will mean acceleratinginvestment in grids – and working to ensure that data centres, as well as the wider electricitysystem, are as efficient and flexible as possible. Making this a reality will hinge on the final pillar: bolstering dialogue between policy makers,the tech sector and energy industry. This is an area in which the IEA is proud to have taken aleadership role – and will continue to do so. AI could also be an incredibly powerful tool for the energy sector. It is already helping energycompanies optimise their approaches to exploration, production, maintenance and safety –and if AI tools are applied broadly, huge amounts of electricity transmission