A new study published on arXiv details the significant water consumption of U.S. hyperscale data centers, distinguishing between direct cooling and electricity generation pathways. The research indicates that these facilities consume approximately 300 GL of water annually, with three-quarters of this amount attributed to electricity production. The study highlights that direct cooling water use is concentrated in stressed basins in the western and south-central U.S., while electricity-related water consumption is concentrated in a few eastern grid regions with heavy fossil fuel reliance. The findings suggest that decisions regarding cooling design and water sourcing have local impacts, whereas electricity planning and procurement have regional impacts. AI
IMPACT Highlights the substantial water footprint of AI infrastructure, potentially influencing future data center siting and energy procurement strategies.
RANK_REASON Academic paper detailing environmental impact of AI infrastructure. [lever_c_demoted from research: ic=1 ai=0.7]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
