首页|人工智能发展下算力的电力需求测算方法研究

人工智能发展下算力的电力需求测算方法研究

扫码查看
"人工智能的尽头是能源",算力爆发式增长引发的电力需求快速攀升已成为能源领域研究的热点问题。本文基于算力规模与结构、数据中心能效水平、运算能力等相关因素,构建了数据中心电力需求预测模式。预计到2030年,我国数据中心用电量将达到0。8~1。5万亿千瓦时。建议加强电力与算力协同规划布局、创新可再生能源供给模式、进一步提升数据中心能效水平,开展余热回收示范。
"The ultimate limit of artificial intelligence is energy."The rapid increase in electricity demand triggered by the explosive growth of computing power has become a hot topic in energy research.Based on factors such as computing power scale and structure,energy efficiency levels of data centers,and computational capabilities,this paper constructs a prediction model for the electricity demand of data centers.It is estimated that by 2030,the electricity consumption of data centers in China will reach 0.8 to 1.5 trillion kilowatt-hours.It is recommended to strengthen the collaborative planning and layout of electricity and computing power,innovate renewable energy supply models,further enhance the energy efficiency of data centers,and carry out demonstrations of waste heat recovery.

Artificial IntelligenceData CenterElectricity ConsumptionCalculation Method

杨万涛、田磊、刘凡、林寰

展开 >

中国宏观经济研究院能源研究所,北京 100038

人工智能 数据中心 用电量 测算方法

2024

中国能源
国家发展和改革委员会能源研究所

中国能源

CSTPCDCHSSCD
影响因子:1.059
ISSN:1003-2355
年,卷(期):2024.46(6)