Many data centers rely at least partly on water to absorb heat from air or surfaces, and then cause cooling as it evaporates, just like sweating keeps our bodies cool. The processors inside data centers can reach internal temperatures of up to 176 degrees Fahrenheit, for the same reasons that an overtaxed laptop heats up.

But while there’s little doubt that data centers are consuming significant amounts of energy, water consumption is more complicated. Claims circulating online bring scant clarity, ranging from articles claiming that chatbots can consume 500 ml of water per query to some online commentators declaring AI’s water issue to be nonexistent. And the consumption data from large tech companies themselves are often incomplete and inconsistent.

Experts stress that, overall, data center water consumption pales in comparison to the usage by industries like agriculture and some kinds of manufacturing. Researchers have estimated that data center cooling systems consumed 66 billion liters of water in 2023, less than 1 percent of the nation’s total consumption.

That figure could considerably rise with the rapid buildout of AI data centers, however — which may not be a big problem for water-rich regions but could significantly add to local water stress in drought-strapped places like New Mexico and Arizona.

Fortunately, engineers say there’s a host of things that can be done — and are being done — to reduce the strain on local water resources. “There are many good ideas out there,” says energy systems expert Fengqi You of Cornell University. When you add in efforts to replenish and restore natural water resources, “there’s a good chance that, eventually, it could be net-zero water for the on-site cooling.”

More renewable energy and better siting

Recent advances in AI technology have already reduced water demand for many data centers. Electrical and computer engineer Shaolei Ren of the University of California, Riverside, says that an estimate he made based on GPT-4, an earlier version of the model powering ChatGPT, as recently as 2024 — that drafting a short email would consume 500 milliliters of water — is already outdated because AI models have become more efficient. (This study was the origin of the half-liter-per-query claim.) In 2025, Google estimated that five drops of water are spent on processing a median-length query with its chatbot Gemini.

But even tiny amounts add up, given the rise in AI use. You’s group recently predicted that by 2030, US data centers could consume 731 billion to 1,125 billion liters annually — the latter roughly equivalent to New York City’s annual drinking water supply.

Still, these numbers factor in not just cooling, but also a greater amount of water used to generate the electricity that powers data centers. Much of that comes from burning coal or gas, with water being used to cool steam back into liquid after it’s been used to spin the electricity-generating turbines. In other words, AI’s water consumption can be significantly reduced by shifting to solar and wind, which require little to no water to operate, You says.

You adds that companies should site data centers in areas less prone to drought and with ample renewable energy sources, like regions in Montana, Nebraska, parts of Texas and South Dakota, instead of constructing new data centers in water-stressed places, such as in Arizona, New Mexico and Southern California. Strategic siting along with other measures could reduce AI’s future water footprint by up to 86 percent, he says.

New cooling technologies

Advances in cooling technologies are also helping. The newest data centers that specialize in AI are already more water-efficient than their predecessors.

Pre-AI data centers use fans to carry away the heat from processor-filled racks, then use chilling systems funneling cool water through the building to cool the warm air down again. The heated water is then fed into a cooling tower that removes the heat but also loses some of the water to the atmosphere. This evaporative cooling uses a lot of water, says Eric Masanet of the University of California, Santa Barbara, who researches data center sustainability.

But the energy-hungry processors inside AI data centers generate so much heat that it’s hard to remove it with air alone. That’s why many tech companies — including Amazon, Microsoft and Google — use a more efficient technique called liquid cooling.

In this method, a system of pipes filled with water, sometimes mixed with temperature-regulating chemicals, runs on top of the hardware, cooling the processors directly rather than cooling the entire data center. And because this is a closed-loop system, no water is lost: The heated fluid can be cooled down by exposing the pipes to the outside air, if conditions are cool enough, or an air-conditioner-like system, which uses energy.

But when it’s extremely hot or humid, the centers need to bring online extra, more effective methods that involve water; one of these is to mist the air around the pipes. Mechanical engineer Vaibhav Bahadur at the University of Texas at Austin says that most AI data centers in Texas only use this kind of water-based cooling for the hottest parts of the summer.

Still, he and his colleagues recently estimated that the state’s growing number of data centers — which currently consume less than a percent of the state’s water demand — could be using 3 to 9 percent of that water by 2040. He expects that figure to reduce considerably, though. “Data centers are becoming much more efficient in their water usage,” he says.

Indeed, recent technological developments have made it easier to avoid water-based cooling, even in hot climates, says Josh Parker, head of sustainability for NVIDIA, the company that designs many of the processors that populate AI-specialized data centers. Their latest generation of processors run at such high temperatures that they can be cooled with water at around 113 degrees Fahrenheit which, Parker says, is significantly warmer than traditional approaches. That reduces the need to bring on water-based cooling. Unless local temperatures regularly exceed 113 Fahrenheit, a threshold generally crossed only during severe heat waves, “typically we can just get away with large efficient fans cooling the infrastructure and the ambient air is sufficient to pull the heat away,” Parker says.

Ren cautions that some companies may not want to run their processors that hot because it can interfere with computational performance, while some also house non-AI servers in the same space that cannot take the heat. But other cooling technologies are in development.

Some of these rely on elaborately designed pipe systems that draw heat even more effectively from processors. A number of US companies are exploring immersion cooling, where data center hardware sits in a tank filled with cooling liquid. In China, some companies have taken a similar approach by installing data centers in the ocean, although this makes them hard to access should hardware need replacing or upgrading, Bahadur says.

Advances in cooling technology will be necessary for tech companies to meet sustainability goals they have set for their operations. According to a statement from Amazon, its Web Services “set a goal to be water positive by 2030 — returning more water to communities than our direct operations use — and as of 2024, we’re more than halfway there.”

A spokesperson for OpenAI points to a web page outlining the company’s plans to minimize water use, while Google aims to replenish 120 percent of the freshwater its data centers consume by 2030 through a variety of water stewardship projects.

Some companies, such as Google, are also finding smart ways to use waste heat from data centers to help address other sustainability challenges, such as making building heating less reliant on fossil fuels. Across Europe, some data centers are funneling heat from cooling circuits into district heating systems used to warm buildings.

The future of AI’s water usage hinges on decisions that tech companies, engineers and policymakers will make in the coming years, Masanet says: where data centers are sited, their local energy infrastructure and climate, and how they’re designed. “If you choose one set of choices, your number is going to be off the charts,” he says. “You choose another set of choices, it’s going to be way down here.”