Artificial intelligence has intensified concerns about the water consumed by data centers, but the reality is more nuanced than either critics or technology companies often suggest. According to The Atlantic, asking how much water data centers use at the national level misses the more important question of where that water is consumed and how it affects local communities. While AI is driving rapid growth in computing infrastructure, the environmental consequences vary widely depending on geography, cooling methods, and energy sources rather than on the technology itself.
Nationally, data centers account for only a small fraction of U.S. water consumption. In 2023, they used just over 17 billion gallons of water for cooling, a tiny share compared with agriculture. However, these averages can obscure significant local effects. A single hyperscale facility can require millions of gallons of water each day, placing substantial pressure on the municipal water systems of small towns or drought-prone regions. Communities hosting new data centers may therefore experience resource constraints even when national water use remains relatively modest.
The article also explains that reducing water consumption is not always straightforward. Many facilities rely on evaporative cooling because it is energy efficient but consumes more water. Alternatives, such as air-cooled closed-loop systems, dramatically reduce direct water use but require more electricity. If that electricity comes from fossil-fuel-powered plants, the overall water footprint may increase because thermal power generation also consumes large amounts of water. As a result, evaluating sustainability requires considering both direct water use at the data center and indirect water use associated with electricity production.
A major obstacle is the industry’s limited transparency. Many companies disclose little about facility-level water consumption, making it difficult for communities to assess potential impacts before construction begins. The article argues that improved reporting, thoughtful site selection, cleaner electricity sources, and advanced cooling technologies can reduce environmental pressures. Rather than portraying AI data centers as either insignificant or catastrophic water users, it concludes that their true impact depends on local conditions, engineering choices, and responsible planning.