グリッドを超えて：AIデータセンター向けオンサイト電源技術のコスト、炭素、資本要件
Beyond the Grid: Cost, Carbon, and Capital Requirements of On-Site Power Technologies for AI Data Centers (原題)
Eliseo Curcio
🤖 gxceed AI 要約
日本語
本論文は、AIデータセンターのオンサイト電源技術9種を、内生的なグリッド価格と完全サイトベースで評価。ガス複合発電はサイト全体で114 USD/MWhとグリッドの92 USDを上回り、4時間ストレージは年間エネルギーの18%に制限され、グリッドより炭素集約的。水素はコストと炭素の両方で失敗。投資反転により、グリッドパリティはどの予算でも購入不可能で、オンサイト供給はアクセスと深さの製品であると結論。
English
This paper evaluates nine on-site power technologies for AI data centers against a grid with endogenous pricing, using complete-site costs and measured GPU loads. Gas combined cycle costs 114 USD/MWh vs 92 USD grid; four-hour storage is capped at 18% of annual energy and dirtier than grid; hydrogen fails on cost and carbon. An investment inversion shows grid parity is not purchasable at any budget; on-site supply is an access and depth product.
Unofficial AI-generated summary based on the public title and abstract. Not an official translation.
📝 gxceed 編集解説 — Why this matters
日本のGX文脈において
日本のGX文脈では、AIデータセンターの電力需要増加が電力系統に与える影響が注目されており、本論文のオンサイト電源のコスト・炭素分析は、日本のデータセンター立地や電力調達戦略に示唆を与える。また、45V/48Eのような税制優遇の効果を検証しており、日本の類似インセンティブ設計にも参考になる。
In the global GX context
Globally, this paper addresses the critical intersection of AI-driven electricity demand and decarbonization, offering a rigorous framework for evaluating on-site power technologies. It challenges conventional LCOE comparisons and highlights the limits of grid parity, informing transition finance and infrastructure investment decisions for data centers worldwide.
👥 読者別の含意
🔬研究者:Provides a novel complete-site cost framework and investment inversion for on-site power, useful for energy systems modeling.
🏢実務担当者:Informs data center developers and operators on the true costs and carbon implications of on-site generation options.
🏛政策担当者:Highlights the need for grid investment and the limitations of tax incentives like 45V/48E in achieving grid parity.
📄 Abstract（原文）
Interconnection queues, not electricity prices, now govern where data centers can be built, and the standard levelized-cost comparison answers a question no developer faces: it assumes a load profile, freezes the grid price while modeling the demand that moves it, and quotes busbar costs a facility cannot buy. This paper evaluates nine on-site supply technologies against a delivered grid whose price is endogenous to projected data-center demand, on a complete-site basis that retains standby charges, with measured GPU training load, delivered fuel prices, production-pathway carbon, and statutory 45V and 48E incentive mechanics. Nothing beats the wire: gas combined cycle produces at 47 USD/MWh but costs about 114 USD per megawatt-hour of complete site energy against a 92 USD grid; four-hour storage is physically capped near 18 percent of annual energy and, charged at the margin, dirtier than the grid; hydrogen from grid-priced power fails on cost and carbon together. An investment inversion converts these findings into capital terms: conversion-hardware learning buys nothing, because free hardware still exceeds the grid for every low-carbon arm, while global electrolyser deployment on sited sub-20 USD/MWh power brings PEM hydrogen power to about 2.2 times the grid at 300 billion USD and 1.9 times at 1 trillion USD (2.7 and 2.3 for the hydrogen engine), with a carbon reduction of roughly 85 percent (6.8-fold) against grid-power production. Grid parity is not purchasable at any budget. On-site supply is an access and depth product; most current investment targets the wrong term.
🔗 Provenance — このレコードを発見したソース
- openalex https://arxiv.org/abs/2608.08170first seen 2026-09-01 05:06:19
🔔 こうした論文の新着を逃したくない方は キーワードアラート に登録（無料・3キーワードまで）。
gxceed は公開メタデータに基づく研究支援データセットです。要約・翻訳・解説は AI 支援で生成されています。 最終的な解釈・検証は利用者が原典資料に基づいて行うことを前提とします。