美联储官员在近期鹰派转向中强调，AI热潮带来的需求激增是通胀压力的重要来源。本报告重新评估AI需求对消费者价格通胀的影响，区分真实效应与统计误测。报告分析四条传导渠道：一是软件及配件，AI推升内存等配件价格并带动具AI功能软件涨价，但预计其对核心PCE通胀约20bp的贡献中大部分源于误测（权重高估及未对软件质量调整）；二是消费电子，电子投入成本上升推高电脑等价格，预计到明年末共贡献约15bp；三是电力价格，数据中心用电需求推高电价，对企业能源成本贡献5-10bp核心通胀，对家庭电费另贡献5-10bp整体通胀；四是工业金属，数据中心建设用金属价格上涨贡献约5bp，建筑工人工资外溢证据有限。总体看，AI价格压力当前约推升核心PCE通胀45bp，其中约一半为真实影响，预计2026年底见顶于50bp，2027年底回落至10bp；对整体PCE通胀的影响同期约高5-10bp。
Fed officials have highlighted surging demand from the AI boom as a key source of inflationary pressure during their recent hawkish pivot. In this Analyst, we reassess the impact of AI demand on consumer price inflation, distinguishing genuine effects from the effect of mismeasurement. We consider four channels of impact. First, AI has raised software and accessories inflation sharply, but we estimate that most of the roughly 20bp contribution to core PCE inflation from this category reflects mismeasurement that exaggerates the impact of surging accessories prices. Second, AI-related cost pressures for electronic inputs have led to price increases for computers and other consumer electronics, contributing roughly 15bp in total to core PCE inflation until late next year. Third, demand for power from data centers has boosted electricity prices, contributing 5-10bp to core inflation and an additional 5-10bp to headline inflation. Fourth, prices for industrial metals used for data center c