AI at Half the Price: Harnessing Off-Grid Hydropower and Direct Cooling for Scalable Intelligence
Keywords:
AI, Data Center Infrastructure, Off-Grid, Decentralized Computing, Free Cooling, Inference Cost ReductionAbstract
The rapid growth of artificial intelligence has significantly increased semiconductor investment, exposing energy and cooling infrastructure as the primary bottleneck for computational progress [1]. Using the Republic of Georgia as an illustrative case study, this paper proposes a framework for developing a decentralized network of AI data centers co-located with small and medium-sized hydroelectric power plants [11]. This model, which is equally applicable to other hydro-rich geographies such as Canada, Alaska, the Pacific Northwest, Kazakhstan, Nepal, Tibet, Scandinavia, the Alps and the Carpathians in Europe, targets the primary operational costs in AI: electricity consumption and thermal management. By eliminating grid-related transmission losses (10%-20%) and utilizing river water for direct cooling (further 25% to 50% electricity consumption reduction), our approach can reduce total electricity usage by up to 60% in optimal scenarios. This strategy offers a practical and economically sustainable method to support high-performance AI workloads globally.
Downloads
References
de Vries, "The growing energy footprint of artificial intelligence," Joule, vol. 7, no. 10, pp. 2191-2194, Oct. 2023, doi: 10.1016/j.joule.2023.09.004.
Z. Wang et al., "Power for AI data centers: Energy demand, grid impacts, challenges and perspectives," Energies, vol. 19, no. 3, Art. no. 722, Jan. 2026. Available: https://www.mdpi.com/1996-1073/19/3/722.
Goldman Sachs Global Investment Research, "U.S. data center power demand projected to double by 2027: Powering the AI era," Goldman Sachs, May 2026. Available: https://www.goldmansachs.com/insights/articles/us-data-center-power-demand-projected-to-double-by-2027.
Electric Power Research Institute, Powering Intelligence: Analyzing Artificial Intelligence and Data Center Energy Consumption. Palo Alto, CA, USA: EPRI, May 2024.
International Energy Agency, "AI is set to drive surging electricity demand from data centres while offering the potential to transform how the energy sector works," Apr. 2024. Available: https://www.iea.org/news/ai-is-set-to-drive-surging-electricity-demand-from-data-centres-while-offering-the-potential-to-transform-how-the-energy-sector-works.
M. Zhang, "How much does it cost to build a data center?" Dgtl Infra, Nov. 2023. Available: https://dgtlinfra.com/how-much-does-it-cost-to-build-a-data-center/
WorldData.info, "Energy consumption in Georgia," 2024. Available: https://www.worlddata.info/asia/georgia/energy-consumption.php.
Cambridge Centre for Alternative Finance, "Cambridge Bitcoin Electricity Consumption Index (CBECI)," Judge Business School, University of Cambridge, 2023. Available: https://ccaf.io/cbeci/index.
G. Mukhigulishvili, "Renewable energy sources," UNDP Georgia, pp. 1-120, 2022. Available: https://www.weg.ge/sites/default/files/undp-georgia-eu4climate-green-deal-assessment-2022-geo.pdf.
Marcy, "EIA study examines the role of high-voltage power lines in integrating renewables," U.S. Energy Information Administration, 2018.
Galt & Taggart, "Electricity market in Georgia," Sector Research Rep., pp. 1-25, 2025. Available: https://ramad.bog.ge/galt/EFjS4MMq-Energy-conference-presentation_GT_ENG_Site.pdf.
Georgian National Energy and Water Supply Regulatory Commission, "Tariffs for hydro power plants," Official Resolution on Marginal Tariffs of Electricity Generated by Hydro Power Plants, 2024. Available: https://gnerc.org/en/tariffs/tariff-el-energy/production/hidroeleqtrosadgurebis-tarifebi/21554.
U.S. Department of Transportation, "Intelligent Transportation Systems (ITS) costs of fiber optic cable deployment," Federal Highway Administration, 2022.
P. Yin, Y. Guo, M. Zhang, et al., "Performance analysis of lake water cooling coupled with a waste heat recovery system in the data center," Sustainability, vol. 16, no. 15, Art. no. 6542, Jul. 2024, doi: 10.3390/su16156542.
S. Xu, H. Zhang, and Z. Wang, "Thermal management and energy consumption in air, liquid, and free cooling systems for data centers: A review," Energies, vol. 16, no. 3, Art. no. 1279, Jan. 2023. Available: https://www.mdpi.com/1996-1073/16/3/1279.
H. E. Burford, L. Wiedemann, W. S. Joyce, and R. E. McCabe, "Deep water source cooling: An untapped resource," pp. 10-15, 2012.
World Bank, "Electric power transmission and distribution losses (% of output)," World Development Indicators, 2023. Available: https://data.worldbank.org/indicator/EG.ELC.LOSS.ZS.
Occupational Safety and Health Administration, "Occupational noise exposure," Standard 1910.95, U.S. Department of Labor. Available: https://www.osha.gov/noise.
