工业革命与 AI 革命的“化石逻辑”：数据中心耗电与百年煤电选址的重合

Data centers in which language models are trained and run consume electricity at industrial-revolution scale. The International Energy Agency’s 2025 Energy and AI report estimates global data-center consumption at around 415 terawatt-hours in 2024, about 1.5 percent of global electricity use, growing at twelve percent a year, and projected to more than double to 945 terawatt-hours by 2030. AI inference, the multifarious use of trained models, is becoming the larger share. And through an ironical pattern in the substructure of history, the largest data-centers of 2025 are sited where the largest coal-fired power stations were a century ago.

For most of the history of life, energy has moved through trophic, or feeding, networks. Producers capture sunlight, consumers eat producers, and predators eat consumers. At every link in the network of predators and their prey energy is lost as heat and total throughput is bounded by the flux of solar radiation…

Get The Continuum, our free Thursday newsletter.