On Monday, SpaceX amended its initial public offering to state that water conditions—including water scarcity, regulations around water, and drought—could constrain data center development.
It isn’t the only tech company trying to assess how water scarcity might impact its business. Water use is emerging as one of the most contentious data center issues. A recent Gallup poll found that seven out of 10 Americans are opposed to data center development, with water scarcity ranking as the top resource concern. Facing increasingly fierce resistance, some tech companies are scrambling to assure the public that they’re facing the issue head-on.
Data centers primarily use water to cool server racks, which throw off massive amounts of heat. One popular technique, known as evaporative cooling, uses fresh water to absorb the heat, which is then pumped to cooling towers where it evaporates outside.
Using more water can save money and reduce emissions for big tech companies by reducing the power needed for cooling that relies on energy-intensive pumps to recirculate water. But it can also come with a large water footprint: Google’s facility in Council Bluffs, Iowa, for instance, which uses evaporative cooling, consumed more than 1 billion gallons in 2024.
Lawrence Berkeley National Laboratory predicted in a 2024 report that hyperscale data centers could consume up to 33 billion gallons of water by 2030 if they relied heavily on evaporative cooling. That’s on par or even less than other thirsty industries, like agriculture or oil and gas—a single fracked well can use 1.5 to 16 million gallons of water—but it poses a risk in regions where water is already scarce. The risk is particularly acute in summer, when data center cooling needs tend to skyrocket at the same time as municipal water use.
“Water is a highly local, highly regional issue,” says Shaolei Ren, a professor of engineering at UC Riverside. “It’s a limited resource, and we have to manage it very carefully.”
Some tech giants, including Microsoft, OpenAI, and Oracle, have made statements in recent months indicating that they are moving away from evaporative cooling entirely in order to save water. That includes OpenAI and Oracle’s massive Stargate expansion in a number of states, including a water-stressed region of Texas.
Google is taking a different approach. On Wednesday, the company rolled out a series of water-related commitments to communities where it has data centers, along with funding announcements for water-related projects in the US.
They include pledges to replenish more freshwater than the company consumes, via investments in local water projects; to scale up the use of reclaimed and recycled water; and to disclose annual water use in data centers. (Other tech companies, including Microsoft, have similar promises around water replenishment and local investment. Google has been working on most of these pledges for a few years.) There’s also a promise to use “a data-driven framework” to decide what data center designs would work best with local watersheds.
Ben Townsend, the global head of infrastructure and sustainability at Google, says that data center design is a lot more complicated than simply swearing off one type of cooling in all cases. The company, he says, has been doing detailed hydrologic assessments of its sites for the past four years to determine what types of cooling would work best.
“Water is scarce in some regions and plentiful in others,” he says. “A one-size-fits-all strategy just doesn’t work.”
In April, Google defended evaporative cooling for areas with what it called “abundant” water in a filing to the European Union as necessary for developing truly sustainable data centers. Google’s arguments line up with new research from Ren and his team, who found that if all data centers in the US were to adopt some kind of evaporative cooling during peak demand, it could free up an additional 10 to 30 gigawatts of power. In areas where grids are stressed but water resources aren’t, using evaporative cooling could provide a meaningful headroom to utilities trying to balance load.
