AI時代のクリーンファーム電源：エネルギー・トライレンマを越えるデータセンター脱炭素への地熱経路

Unofficial AI-generated summary based on the public title and abstract. Not an official translation.

🔬研究者:地熱の信頼性・炭素・コストをトライレンマで整理する比較枠組みの実証例として参照できる。

🏢実務担当者:データセンターの電力調達・冷却設計で、地熱PPAや地熱冷却の脱炭素・コスト効果を検討する材料になる。

🏛政策担当者:AI需要増下での脱炭素電源政策と地熱資源開発支援の根拠として活用できる。

Abstract The evolving development and deployment of Artificial Intelligence (‘AI’) is driving unprecedented 24/7 data center energy demand. A single generative AI query consumes roughly ten times the electricity of a conventional internet search, and one planned US data center has requested 1.4 GW of continuous power. This will exceed the total demand of some small states. These loads reframe the energy trilemma, highlighting the need to simultaneously deliver reliability, low carbon impact, and cost-effective energy, and expose the limitations of intermittent renewables for mission-critical digital infrastructure. Geothermal energy, with capacity factors exceeding 70–90% and very low lifecycle emissions, offers a reliable "clean firm" power and thermal management solutions. This study examines how the United States, UAE, and Saudi Arabia deploy geothermal solutions to decarbonize data centers while advancing all three trilemma pillars. The methodology adopts a three-case comparative analysis. In the United States, we assess Utah's Operation Gigawatt, aiming to double capacity from ∼4 GW, and Ormat Technologies’ long-term Power Purchase Agreements (‘PPAs’), including contracts supplying up to 150 MW of geothermal electricity to hyperscale operators. In the UAE, we evaluated ADNOC's G2COOL geothermal district cooling pilot at Masdar City, using ∼90°C water from ∼2.6 km depth to provide ∼10% of cooling demand with ∼50% lower electricity consumption. In Saudi Arabia, we analyzed KAPSARC resource mapping (Alsuwailem et. al, 2024) and KAUST's 2024 pilot at Thuwal targeting geothermal cooling applications. Each case is systematically assessed against trilemma metrics of reliability, carbon impact, and cost-effectiveness. This study evaluates geothermal applications for power generation and thermal management across three stages of market maturity. The United States case demonstrates reliability through baseload geothermal electricity for hyperscale computing; the UAE delivers carbon impact reduction through operational district cooling in extreme climates; and Saudi Arabia explores cost-effectiveness through early-stage geothermal-exchange systems optimizing cooling-intensive demand. The assessment identifies distinct contributions to reliability, carbon impact, and cost-effectiveness. For reliability, geothermal provides continuous baseload power and long-term PPAs that reduce supply volatility. For carbon impact, geothermal systems emit approximately 90% lower lifecycle CO2 than combined-cycle gas (≈40 against ≈500 gCO2e/kWh) and up to ∼99% lower operational CO2 against the same baseline while materially reducing cooling-related emissions in high-demand regions (IPCC 2011). For cost-effectiveness, for the right application, geothermal cooling and geo-exchange solutions can reduce electricity consumption by up to ∼50%, particularly in markets where cooling accounts for 65–70% of power demand. Across all three jurisdictions, geothermal solutions reconcile accelerating AI-driven electricity demand with system resilience, decarbonization objectives, and cost optimization. By positioning geothermal within the energy trilemma framework, this study demonstrates its strategic role in supporting sustainable large scale digital infrastructure.

• crossref https://doi.org/10.2118/235351-msfirst seen 2026-11-03 05:54:45 · last seen 2026-11-05 06:35:42