Fueled by major investments, tax incentives, and access to reliable power and high-capacity fiber networks, Georgia is emerging as a hub of data center development in the United States. Despite the promising economic benefits, the rapid growth of these facilities has sparked debate over water use — especially during periods of peak demand.
That lack of attention to peak water use, rather than total annual consumption, is one of the biggest blind spots in current policy discussions, according to Shaolei Ren — a professor at the University of California, Riverside who studies AI’s environmental footprint.
“Most of the focus right now is on total annual water use,” Ren said. “But some data centers have very spiky usage during summer. That creates a lot of challenges for infrastructure planning.”
AI’s water impact is not only about how much water is used, but also when it is used. Hot summer months, during peak demand, can place sudden pressure on local water infrastructure. The peak usage data is rarely disclosed, and the water footprint of AI-driven data centers is frequently underreported.
Unlike electricity consumption, which is tracked and discussed more publicly, it is harder to measure water use, which is felt most acutely at the local level. Residents and local governments are frequently responsible for funding infrastructure upgrades as a result.
Although data centers are contributing to local economies, they consume significant amounts of water and electricity. In some areas, residents worry that the increased water withdrawals could strain supplies or potentially cause private wells to dry up during periods of high-demand.
How Data Center Design Affects Water Usage
Data center designs play a significant role in the determination of water consumption. Facilities that rely on evaporative cooling systems might use less energy, but consume more water, particularly in hotter climates. These design choices directly affect the strain placed on local water infrastructure.
Peak water demand matters because water systems must be built to handle the highest single-day usage — even if that level of demand occurs only a few times a year. Expanding water infrastructure to accommodate increased demand may cost hundreds of millions over several decades.
In Newton County, Georgia, for example, peak water consumption from existing data centers costs local water systems roughly $10 million annually .
Why Georgia Attracts Data Centers
Georgia appeals to data center developers due to a combination of pro-business regulatory climate, abundant land, and access to reliable energy and high-capacity fiber networks. These assets make the state an ideal location for the infrastructure needed to support AI technologies.
Local governments are welcoming data centers, because of the jobs they create and tax revenue generated. However, these benefits come with compromises that are not always fully visible to surrounding communities.
“Most people aren’t even aware of the physical existence of AI computing,” Ren said. “It uses energy, water, emits carbon — these are real costs to society. Local impacts, especially on water infrastructure, are often overlooked.”
As Georgia continues to attract more data centers, Ren argues that local officials need to evaluate water capacity alongside electricity availability to understand the true costs and benefits of this development.
