A new study by a University of California, Riverside research team in collaboration with Caltech reports that the rapid growth of artificial intelligence and cloud computing is straining community water systems, which may need billions of dollars in upgrades to meet peak water demands from data centers. The findings were released on Mar. 12.
The issue is significant because as more data processing centers are built to support AI and cloud services, their cooling needs can create sudden spikes in local water use—especially during hot weather. Without improvements in efficiency, researchers estimate that by 2030, these facilities could require an additional 697 million to 1.45 billion gallons of peak daily water capacity, roughly equivalent to New York City’s typical daily supply.
Shaolei Ren, associate professor at UC Riverside’s Bourns College of Engineering and lead author of the study, said the cost for necessary infrastructure could range from $10 billion to $58 billion depending on how quickly data centers expand. “Even if you have money, the water source is another challenge,” Ren said. “In many cases, the water is naturally replenished by snowpack and reservoirs. But reservoirs and snowpack are limited. You may have money to build treatment plants and pipes, but money can’t buy more snowpack.”
The report notes that while annual figures for data center water use often appear modest, actual demand can spike six to ten times higher during heat waves—and sometimes even exceed thirty times average usage for certain planned facilities. This forces municipal utilities to invest in infrastructure capable of handling rare but extreme peaks.
To address these challenges, the researchers recommend that developers report peak rather than average annual water use; partner with communities on funding upgrades; add enough capacity or efficiencies to offset their own impact; and coordinate with utilities by adjusting cooling methods based on grid or system stress levels.
Ren concluded: “People recognize power as a constraint for data center growth,” he said, “but most of them haven’t realized water is a hidden and even more binding constraint in many communities.”
