A data center for cryptocurrency mining, cloud services, and AI computing in Stutsman County, North Dakota (source: halbergman/Getty Images).
Rapid expansion in AI and cloud computing is driving a new wave of data center construction in the United States. According to this article on Wired.com, the scale of growth could add up to 44 million tons of CO₂e per year, an extra burden comparable to the annual emissions of some small nations. The analysis argues that strategic siting of these centers can mitigate both water stress and carbon footprint.
Current concentration of data-centers in states such as Virginia and California is driven by tax incentives, infrastructure, and connectivity but those choices often come with environmental trade-offs. Virginia’s goals for 100% clean energy by 2045 may suffer when big server-parks draw on fossil-heavy grids. California, meanwhile, faces severe water scarcity, which threatens life-cycle cooling and local resource supply.
The study identifies several U.S. states, such as Texas, Montana, Nebraska, and South Dakota, as more optimal locations for new data-center build-out. These states strike a better balance: lower water stress, more favorable grid mixes (including higher renewables), and more land availability.
The article also highlights key variables: energy grid carbon intensity, regional water availability (for cooling systems), and future capacity of renewable build-out. The message: siting matters more than simply “build more.” And while advances in cooling and server efficiency will help, they may not be sufficient if massive new build-outs are placed in unfavorable regions. The research underscores that the tech industry’s net-zero pledges are harder to meet than often assumed.
For engineers and content creators covering infrastructure, cloud, and sustainability, this means looking beyond server hardware. The geography, resource context, grid maturity, and local water ecosystem all become part of the design question when planning future data infrastructure.