content image

AIの電力戦争:データセンター急増の裏にある送電網の費用は誰が負担するのか?

AI’s Power War: Who Pays for the Grid Behind the Data-Center Boom?

AIの急成長でデータセンターの電力消費が爆発的に増加。送電網の費用負担や気候目標との両立をめぐり、規制当局とビッグテックの攻防が激化している。
分からないところをタップすると
↓日本語訳が表示されます↓

The phrase “power war” is no exaggeration. The International Energy Agency says data centres consumed about 415 terawatt-hours of electricity in 2024, roughly 1.5% of the global total, and could reach around 945 TWh by 2030, with AI as the main engine of growth. In the United States, where 45% of global data-centre electricity use was concentrated in 2024, the IEA expects data centres to account for nearly half of electricity-demand growth through 2030. A U.S. Department of Energy-backed Berkeley Lab study points in the same direction: data centres used about 4.4% of U.S. electricity in 2023 and could rise to 6.7% to 12% by 2028. (iea.org)

The most explosive question, then, is not whether AI needs power, but who pays for the wires, substations, and grid upgrades. Regulators are increasingly trying to stop ordinary households from subsidizing hyperscale campuses. In Ohio, AEP Ohio’s tariff for new or expanded data centers of 25 megawatts or more requires study fees, minimum demand charges, and reimbursement of buildout costs if a project is cancelled or significantly delayed, while the Ohio Consumers’ Counsel says the settlement was designed to shield residential and small-business customers from unfair infrastructure and transmission costs. Virginia’s State Corporation Commission has created a separate GS-5 rate class for hyperscale users; new qualifying large-load customers must generally commit for at least 14 years, and large-load customers must pay at least 85% of relevant transmission and distribution costs each month, even if they use less electricity. At the federal level, FERC opened a PJM review in February 2025 specifically to address co-located AI data centers, grid reliability, and fair costs to consumers. (aepohio.com)

Climate pledges, meanwhile, are not vanishing so much as mutating. Google reported that its data-center electricity demand rose 27% in 2024, yet data-center energy emissions fell 12% as more clean-energy projects came online; even so, its total emissions still rose 11% year on year. Microsoft still says it aims to be carbon negative by 2030, but its 2025 sustainability report admits the journey is “a marathon, not a sprint,” after total emissions rose 23.4% versus its 2020 baseline because of AI and cloud expansion. To close that gap, Big Tech is moving beyond slogans into harder infrastructure: Microsoft’s 20-year deal with Constellation is tied to restarting Three Mile Island Unit 1 in 2028, while Google’s deal with Kairos Power aims to deliver up to 500 MW of advanced nuclear capacity by 2035. The new AI bargain is becoming clear: if society wants frontier models, someone must underwrite the grid—and climate credibility will depend less on rhetoric than on delivering real, round-the-clock clean electricity. (blog.google)

by EigoBoxAI
作成:2026/03/28 03:05
レベル:超上級 (語彙目安:8000語以上)

まだ読んでいないコンテンツ

content image
by EigoBoxAI
作成:2026/03/28 03:03
レベル:上級 (語彙目安:6000〜8000語)
content image
by EigoBoxAI
作成:2026/03/28 03:01
レベル:中上級 (語彙目安:4000〜6000語)
content image
by EigoBoxAI
作成:2026/03/27 21:04
レベル:超入門 (語彙目安:〜300語)
content image
by EigoBoxAI
作成:2026/03/27 21:02
レベル:初級 (語彙目安:300〜1000語)
content image
by EigoBoxAI
作成:2026/03/27 21:01
レベル:中級 (語彙目安:2000〜2500語)
content image
by EigoBoxAI
作成:2026/03/27 15:05
レベル:初中級 (語彙目安:1000〜2000語)
content image
by EigoBoxAI
作成:2026/03/27 15:03
レベル:超上級 (語彙目安:8000語以上)
content image
by EigoBoxAI
作成:2026/03/27 15:01
レベル:上級 (語彙目安:6000〜8000語)
content image
by EigoBoxAI
作成:2026/03/27 09:06
レベル:中上級 (語彙目安:4000〜6000語)
content image
by EigoBoxAI
作成:2026/03/27 09:04
レベル:超入門 (語彙目安:〜300語)