Rapid data center expansion fuels debate over water and electricity demands | The Big Deal
The rapid growth of AI is fueling data center expansion, leading to increased water and electricity demands and sparking political battles nationwide. The video highlights different cooling methods, such as closed-loop systems used by H5 in Cleveland, and the significant water consumption of evaporative methods. Nell Green Nylen from UC Berkeley discusses the lack of information for communities to assess the environmental footprint of these facilities.
The rapid expansion of artificial intelligence is driving a boom in data centers, which require substantial amounts of water and electricity. This demand has ignited numerous political disputes across the country regarding the placement of new data centers. Spectrum News Ohio reporter Cory Olyri examined different cooling methods, noting that some, like evaporative cooling, consume vast quantities of water, contributing to the 17 billion gallons directly consumed by data centers in 2023, a figure projected to double by 2028.
Jeremy Lavell, Vice President of Operations at an H5 data center in Cleveland, discussed their closed-loop cooling system, which minimizes water loss but requires more electricity. Ashisa Mohapatra from Serverlift, a sustainability researcher, highlighted the popularity of water-intensive evaporative cooling methods but anticipates a shift towards more efficient solutions due to public backlash against high water usage. The power demands are so significant that companies like Microsoft are securing their own energy sources, with Microsoft reportedly investing $16 billion over 20 years to restart a nuclear reactor at Three Mile Island in Pennsylvania solely for AI power.
Nell Green Nylen, a senior research fellow at the Center for Law, Energy and the Environment at UC Berkeley School of Law, emphasized that data centers have diverse impacts on water and energy usage, as well as physical footprints. She stressed that communities need to thoroughly evaluate these factors, especially in water-stressed regions. Nylen also pointed out a significant challenge: developers often require local governments to sign non-disclosure agreements, leaving communities in the dark about the actual impacts. A Gallup poll indicates about 70% of Americans oppose new data centers in their communities, reflecting concerns beyond just resource consumption, possibly extending to general unease about AI.
Nylen suggested that communities could seek to benefit more from data centers beyond just local jobs and cautioned that many new, larger data centers are being proposed in rural areas. These areas often lack existing water, electrical, and wastewater infrastructure, as well as the capacity to thoroughly evaluate complex data center proposals. She concluded that both developers and communities must improve their understanding of impacts and tailor siting decisions and development conditions to ensure data centers benefit local areas without causing significant harm.