The Best US Locations for New Data Centers
▼ Summary
– Tech companies’ massive investments in data centers are significantly driving the US economy, with Meta planning to spend $600 billion and OpenAI $1.4 trillion on infrastructure.
– A new study in Nature Communications analyzes the environmental impacts of US data centers, projecting effects through the decade and questioning tech firms’ net zero promises.
– Data centers’ environmental footprint varies by location, with states using more renewable energy and facing less water scarcity being better suited to minimize carbon emissions and water usage.
– The analysis identifies Texas, Montana, Nebraska, and South Dakota as optimal states for future AI server installations due to their balanced energy and water resources.
– Historically, data center development has concentrated in Virginia and California due to proximity to key areas, fiber connectivity, skilled labor, and tax incentives, with Virginia hosting the most centers in the US.
The strategic placement of new data centers across the United States is becoming a critical economic and environmental issue, driven by massive corporate investments in artificial intelligence infrastructure. Recent analysis highlights that selecting the right location can significantly reduce the environmental footprint of these energy-intensive facilities. As tech giants like Meta and OpenAI commit hundreds of billions to expanding their data center networks, understanding which states offer the best balance of resources is essential for sustainable growth.
A comprehensive study published in Nature Communications examines how future data centers could impact the nation’s electricity grids and water supplies through 2030. Researchers combined data on AI chip demand with state-level information on energy sources and water scarcity to model various scenarios. Their findings suggest that current corporate net-zero pledges may not align with the substantial energy and water demands of new facilities, raising concerns about long-term sustainability.
Professor Fengqi You, an energy systems engineer at Cornell University and co-author of the report, notes that the research began three years ago. He describes the timing as ideal for assessing AI’s influence on climate and water systems. The industry’s expansion has accelerated beyond initial projections, fueled in part by heightened federal attention, creating what he calls “so much momentum right now.”
Location plays a decisive role in a data center’s environmental performance. States with cleaner electricity grids or abundant water resources naturally support more sustainable operations. Facilities in regions dependent on fossil fuels or facing water shortages contribute more to carbon emissions and strain local supplies. Cooling systems alone account for a major share of both energy and water use, making resource availability a top priority.
The analysis identifies several states as optimal for new AI server installations based on these criteria. Texas, Montana, Nebraska, and South Dakota stand out for their favorable mix of renewable energy potential and lower water stress. These regions offer a practical path to minimizing ecological impacts while meeting the computational needs of advancing technologies.
Historically, data center development concentrated in established hubs like Virginia and Northern California. Proximity to government agencies in Washington, D.C., and tech headquarters in Silicon Valley, along with robust fiber networks and skilled labor pools, made these areas attractive. Virginia has further incentivized growth with generous tax breaks, a strategy other states are now adopting to compete for projects.
Industry tracking from Data Center Map shows Virginia leads the nation with over 650 data centers, while California ranks third with more than 320. This existing concentration underscores the shifting focus toward resource-rich states as companies seek to align expansion with environmental responsibility.
(Source: Wired)
