Tag
The article discusses the variability in water use for AI data centers and questions the adequacy of a single global metric, suggesting that operators report freshwater use, recycled-water use, local water stress, and consider energy use for meaningful reduction comparisons.
California has signed new laws requiring data centers to disclose their electricity and water usage, aiming to increase transparency and manage environmental impacts.
US data centers tripled their annual water consumption to 17 billion gallons in 2023, driven by generative AI growth, with indirect water use from electricity generation comprising over 80% of the total footprint.
Nvidia announces a warm-water cooling system that can reduce on-site data center water use significantly, but critics note that water consumption for electricity generation and chip manufacturing remains unaddressed, meaning the total water footprint is only partially reduced.
Jeff Bezos argues that AI data centers should be prioritized for water and energy resources over human consumption to enable the development of superintelligence, sparking controversy over AI's environmental impact.
This article from Utah Clean Energy estimates the proposed Stratos data center in Utah would consume 9 GW of electricity from on-site gas plants, doubling the state's peak demand, and produce significant emissions and water consumption.
A report highlights that a data center consumed 30 million gallons of water without initially making payments, drawing attention to resource usage and billing practices in technology infrastructure.