AI Data Centers: The Hidden Environmental Cost and Innovative Cooling Solutions

Why do AI data centers use so many resources?

With the AI boom, construction of new data centers has skyrocketed, and not without consequence -- some communities that count these facilities as neighbors are now facing water shortages and strained power supplies. While tech's data center footprint has been growing for decades, generative AI has seemingly shifted the impacts of these operations toward the catastrophic. What exactly makes these new data centers such a burden on the environment and existing infrastructure, and is there anything we can do to fix it? The industry believes AI will work its way into every corner of our lives, and so needs to build sufficient capacity to address that anticipated demand. But the hardware used to make AI work is so much more resource-intensive than standard cloud computing facilities that it requires a dramatic shift in how data centers are engineered. Typically the most important part of a computer is its "brain," the Central Processing Unit (CPU). It's designed to compute a wide variety of tasks, tackling them one at a time. Imagine a CPU as a one-lane motorway in which every vehicle, no matter the size, can get from A to B at extraordinary speed. What AI relies on instead are Graphics Processing Units (GPU), which are clusters of smaller, more specialized processors all running in parallel. In the example, a GPU is a thousand-lane motorway with a speed limit of just 30 mph. Both try to get a huge number of figurative vehicles to their destination in a short amount of time, but they take diametrically opposite approaches to solving that problem. Phil Burr is Head of Product at Lumai, a British company looking to replace traditional GPUs with optical processors. "In AI, you repeatedly perform similar operations," he explained, "and you can do that in parallel across the data set." This gives GPUs an advantage over CPUs in large but fundamentally repetitive tasks, like graphics, executing AI models and crypto mining. "You can process a large amount of data very quickly, but it's doing the same amount of processing each time," he said. In the same way that thousand-lane highway would be pretty wasteful, the more powerful GPUs get, the more energy hungry they become. "In the past, as [CPUs evolved] you could get a lot more transistors on a device, but the overall power [consumption] remained about the same," Burr said. They're also equipped with "specialized units that do [specific] work faster so the chip can return to idle sooner." By comparison, "every iteration of a GPU has more and more transistors, but the power jumps up every time because getting gains from those processes is hard." Not only are they physically larger -- which results in higher power demands -- but they "generally activate all of the processing units at once," Burr said. In 2024, the Lawrence Berkeley National Laboratory published a congressionally mandated report into the energy consumption of data centers. The report identified a sharp increase in the amount of electricity data centers consumed as GPUs became more prevalent. Power use from 2014 to 2016 was stable at around 60 TWh, but started climbing in 2018, to 76 TWh, and leaping to 176 TWh by 2023. In just five years, data center energy use more than doubled from 1.9 percent of the US' total, to nearly 4.4 percent -- with that figure projected to reach up to 12 percent by the start of the 2030s. Like a lightbulb filament, as electricity moves through the silicon of computer chips, it encounters resistance, generating heat. Extending that power efficiency metaphor from earlier, CPUs are closer to modern LEDs here, while GPUs, like old incandescent bulbs, lose a huge amount of their power to resistance. The newest generation of AI data centers are filled with rack after rack of them, depending on the owner's needs and budget, each one kicking out what Burr described as "a massive amount of heat." Heat isn't just an unwelcome byproduct: if chips aren't kept cool, they'll experience performance and longevity issues. The American Society of Heating, Refrigerating and Air Conditioning Engineers (ASHRAE) publishes guidelines for data center operators. It advocates server rooms should be kept between 18 to 27 degrees celsius (64.4 to 80.6 degrees Fahrenheit). Given the sheer volume of heat GPUs kick out, maintaining that temperature requires some intensive engineering, and a lot of energy. The majority of data centers use a handful of methods to keep their hardware within the optimal temperature. One of the oldest ways to maximize the efficiency of air conditioning is a technique of hot and cold aisle containment. Essentially, cold air is pushed through the server racks to keep them cool, while the hot air those servers expel is drawn out to be cooled and recirculated. Many data centers, especially in the US, rely on the cooling effect that occurs as water changes from a liquid to a gas. This is done by drawing hot air through a wet medium to facilitate evaporation and blowing the resulting cooled air into the server room, in a method known as direct evaporative cooling. There's also indirect evaporative cooling, which works similarly but adds a heat exchanger -- a device that's used to transfer heat between different mediums. In this type of setup, the heat from the warm air is transferred and cooled separately from the server room to avoid raising the humidity levels indoors. Due in part to their cooling needs, data centers have a tremendous water footprint. The Lawrence Berkeley report found that, in 2014, US-based data centers consumed 21.2 billion liters of water. By 2018, however, that figure had leapt to 66 billion liters, much of which was attributed to what it collectively terms "hyperscale" facilities, which include AI-focused operations. In 2023, traditional US data centers reportedly consumed 10.56 billion liters of water while AI facilities used around 55.4 billion liters. The report's projections believe that by 2028, AI data centers will likely consume as much as 124 billion liters of water. "Collectively, data centers are among the top-ten water consuming industrial or commercial industries in the US," according to a 2021 study published in the journal Environmental Research Letters. About one-fifth of these data centers use water from stressed watersheds, i.e. areas where the demand for water may be greater than the natural supply. Most of the water consumed by data centers evaporates and won't be immediately replenished, while the rest goes to wastewater treatment plants. As a trio of academics explained in an op-ed for The Dallas Morning News, data centers are "effectively removing [drinking water] from the local water cycle." Water used in the cooling process is typically treated with chemicals such as corrosion inhibitors and biocides, which prevent bacterial growth. The resulting wastewater often contains pollutants, so it can't be recycled for human consumption or agriculture. And data centers' water use goes well beyond cooling. A much bigger portion of their water footprint can be attributed to indirect uses, mainly through electricity generated by power plants but also through wastewater utilities. These account for about three-fourths of a data center's water footprint, the study notes. Power plants use water in a number of ways, primarily for cooling and to produce the steam needed to spin their electricity-generating turbines. According to the authors, 1 megawatt-hour of energy consumed by data centers in the US on average requires 7.1 cubic meters of water. "Data centers are indirectly dependent on water from every state in the contiguous US, much of which is sourced from power plants drawing water from subbasins in the eastern and western coastal states," the authors explain. To adequately address the water issue, energy consumption must be reigned in too. One major approach to reduce the massive water footprint of these systems is to use closed-loop liquid cooling. This is already ubiquitous on a smaller scale in high-end PCs, where heat-generating components, such as the CPU and GPU, have large heat exchangers that a liquid is pumped through. The liquid draws away the heat, and then has to be cooled down via another heat exchanger, or a refrigeration unit, before being recirculated. Liquid cooling is becoming more and more common, especially in AI data centers, given the heat that GPUs generate. With the exception of mechanical issues, like leaking, and the water needed to operate the facility more generally, closed-loop systems do not experience water loss and so make more reasonable demands on local water resources. Direct-to-chip liquid cooling drastically cuts a data center's potential water use, and more efficiently removes heat than traditional air-cooling systems. In recent years, companies including Google, NVIDIA and Microsoft have been championing liquid cooling systems as a more sustainable way forward. And researchers are looking into ways to employ this approach on an even more granular level to tackle the heat right at the source. Whereas cold plates (metal slabs with tubing or internal channels for coolant to flow through) are commonly used in liquid cooling systems to transfer heat away from the electronics, Microsoft has been testing a microfluidics-based cooling system in which liquid coolant travels through tiny channels on the back of the chip itself. In the lab, this system performed "up to three times better than cold plates at removing heat," and the company said it "can effectively cool a server running core services for a simulated Teams meeting." A blog post about the findings noted, "microfluidics also reduced the maximum temperature rise of the silicon inside a GPU by 65 percent, though this will vary by the type of chip." Another option is "free" cooling, or making use of the natural environmental conditions at the data center site to cool the operation. Air-based free cooling utilizes the outdoor air in cold locales, while water-based free cooling relies on cold water sources such as seawater. Some facilities couple this with rainwater harvesting for their other water needs, like humidification. Start Campus, a data center project in Portugal, is located on the site of an old coal-fired power station and will use much of its old infrastructure. Rather than simply employ a closed-loop, the high temperatures will require the closed-loop system to interact with an open loop. When the campus is fully operational, its heat will be passed onto around 1.4 million cubic tons of seawater per day. Omer Wilson, CMO at Start Campus, said that by the time the water has returned to its source, its temperature will be the same as the surrounding sea. Start Campus has also pledged that there will be no meaningful water loss from this process. There is another novel cooling method, immersion, in which computing equipment is -- you guessed it -- immersed in a non-conductive liquid suitable to draw heat. Wilson described it as a relatively niche approach, used in some crypto mining applications, but not used by industrial-scale facilities. To keep with both energy and cooling needs, some researchers say the industry must look to renewable resources. "Directly connecting data center facilities to wind and solar energy sources ensures that water and carbon footprints are minimized," wrote the authors of the aforementioned Environmental Research study. Even purchasing renewable energy certificates -- which each represent one megawatt-hour of electricity generated from a renewable source and delivered to the grid -- could help shift the grid toward these sources over time, they added. "Data center workloads can be migrated between data centers to align with the portion of the grid where renewable electricity supplies exceed instantaneous demand." Geothermal resources have begun to look especially promising. According to a recent report by the Rhodium Group, geothermal energy could meet up to 64 percent of data center's projected power demand growth in the US "by the early 2030s." In the Western US, geothermal could meet 100 percent of demand growth in areas such as Phoenix, Dallas-Fort Worth and Las Vegas. For cooling, geothermal heat pumps can be used to "leverage the consistently cool temperatures" found hundreds of feet beneath the surface. Or, in locations where there are shallow aquifers present, data centers can make use of geothermal absorption chillers. These rely on the low-grade heat at shallower depths "to drive a chemical reaction that produces water vapor," the report explains. "This water vapor cools as it is run through a condenser and cools the IT components of a data center using evaporation." Iron Mountain Data Centers operates a geothermally cooled data center in Boyers, Pennsylvania at the site of an old limestone mine. A 35-acre underground reservoir provides a year-round supply of cool water. Geothermal may not be a widespread solution just yet, but it's catching on. In 2024, Meta announced a partnership with Sage Geosystems to supply its data centers with up to 150 megawatts (MW) of geothermal power starting in 2027. While novel cooling methods will undoubtedly help curb some of the AI data centers' excessive resource demands, the first step to meaningful change is transparency, according to Vijay Gadepally, a senior scientist at MIT's Lincoln Laboratory Supercomputing Center. AI companies need to be upfront about the emissions and resource use associated with their operations to give people a clear view of their footprints. Then there is the hardware to consider. Incorporating more intelligent chip design -- i.e. processors with better performance characteristics -- could go a long way toward making data centers more sustainable. "That's a huge area of innovation right now," Gadepally said. And large data centers are often "running underutilized," with a lot of power that isn't being allocated efficiently. Rather than leaning into the push to build more such facilities, the industry should first make better use of existing data centers' capacities. Similarly, many of today's AI models are vastly overpowered for the tasks they're being given. The current approach is "like cutting a hamburger with a chainsaw," Gadepally said. "Does it work? Sure... but it definitely is overkill." This doesn't need to be the case. "We have found in many instances that you can use a smaller but tuned model, to achieve similar performance to a much larger model," Gadepally said, noting that this is especially true for new "agentic" systems. "You're often trying thousands of different parameters, or different combinations of things to discover which is the best one, and by being a little bit more intelligent, we could dismiss or essentially terminate a lot of the workloads or a lot of those combinations that weren't getting you towards the right answer." Each of those unnecessary parameters isn't just a computational dead end, it's another nudge towards rolling blackouts, less potable water and rising utility costs to surrounding communities. As Gadepally said, "We're just building bigger and bigger without thinking about, 'Do we actually need it?'"

Graduate students at UT-Arlington are developing a way to cool data centers without water

ARLINGTON, Texas (AP) -- Sai Abhideep Pundla has been awake since 3 a.m. After a red-eye flight from Las Vegas, where he briefed data center company executives and local government officials about the future of artificial intelligence, he's back in a lab at UT-Arlington, tinkering with a prototype he thinks could solve one of the industry's biggest challenges: how to keep data centers cool without draining finite water supplies. Pundla, a doctoral candidate in engineering, is testing a system that cools the computer servers using a recirculating chemical refrigerant instead of water. It's a timely innovation. Texas is building dozens of massive data centers -- some as large as New York's Central Park -- and experts say they're expected to guzzle millions of gallons of water a year in a state facing an increasingly urgent water crisis. Every snap of a photo, message sent, or Google search requires data, which has to go somewhere. That "somewhere" is a data center. These massive facilities filled with servers that store and process everything we do online keep our digital lives intact. However, keeping all that data takes electricity to power the data centers and cooling systems to keep their equipment from malfunctioning. And both of those require water. While data centers currently consume a small portion of the state's total water supply, according to the state's water plan, some researchers warn that with droughts and population growth, data centers have the potential to help push water supplies to the brink -- especially in Texas' more arid regions. The Houston Advanced Research Center, an independent nonprofit research organization focusing on sustainability solutions, estimates that existing data centers in Texas will consume approximately 25 billion gallons of water, or 0.4% of the state's total water use in 2025. By 2030, this demand could increase up to 2.7% of the total annual water use in Texas. That's the equivalent water consumption of 1.3 million average U.S. households. The state currently has more than 400 data center facilities, with about 70 more on the way. "(It) may not feel like it's a whole lot at the state level," said Margaret Cook, a researcher studying data center water use at HARC. But if massive data centers locate in small communities, she said, they may not be able to handle big jumps in water demand. In the Texas Panhandle, for example, Amarillo residents recently held a community event to oppose five planned data centers and inform others of the potential risks to the Ogallala Aquifer, the region's main water source that is being drained faster than it can be replenished. "Our water sources aren't as reliable as what would be needed for the data center," said Madison Boyle, an event organizer. This growing concern has highlighted the state's limitations in tracking and forecasting water usage by emerging industries. In Texas, companies are required to report historical water consumption, but aren't required to report how much water they expect to consume in the future or where it would come from, making it difficult for communities to plan or better manage their water supplies. Other states are starting to act on data centers' water use. California lawmakers recently passed a bill, which is awaiting the governor's signature, that will require new data centers to report their projected water use before they start operations, while in Minnesota, data center developers are required to consult with the state's environmental agency to make sure their proposed location has an adequate water supply. Some data centers in other states are built near lakes, rivers or in colder climates where there's natural free cooling. In Texas, many data centers are located in areas where water supplies already are under high stress, according to research by Yi Ding, an assistant professor at Purdue University's School of Electrical and Computer Engineering. "You could consume the same amount of water in Texas and Iowa," Ding said. "But the environmental burden is different because Texas is more dry." In the lab in Arlington, Pundla and his colleague Braxton J. Smith flip on power to a server packed with memory modules, processor chips and fans -- the same components that make up massive data centers. A hum kicks in as vapor from a special liquid cools the system. No water involved. Pundla said the industry knows their water and energy consumption is unsustainable. "The thermal management side of data centers is now trying to play catch up" as the industry develops bigger, more powerful computing systems that need more electricity and generate more heat, he said. The urgency to find solutions is growing. Earlier this year, President Donald Trump announced a $500 billion federal push, dubbed Stargate, to build AI data centers and their supporting power infrastructure in Texas. Without smarter cooling, Pundla's engineering professor Dereje Agonafer said, it may come at a steep environmental cost: intensifying water scarcity, forcing homes and farms to compete with industry and disrupting aquatic biodiversity. How data centers use water Data centers use water in two main ways: directly, to cool their equipment, and indirectly, through the electricity they rely on. The current Texas data center market requires 10,234 megawatts of electricity a year, which is equivalent to more than 8 million homes' usage a month, according to the research firm Baxtel that tracks data centers and their development. Data centers are expected to help drive Texas' power demand to nearly double by 2030. Electricity generation also requires water, particularly to cool the natural gas, coal or nuclear power plants that generate most of the state's electricity. That means the more electricity a data center uses, the more water is indirectly consumed. Cook said whether data centers choose to connect to the grid or build their own power supply -- some in Central Texas already are building on-site power plants -- will impact their water needs for cooling. The bigger driver of water use is the need to cool the data center's servers, which run 24/7, generating tremendous heat that can cause servers to fail if they're not properly cooled. Over the years, different methods have evolved to solve the cooling problem. Traditional systems use arrays of big fans to blow cool air across the machines, but that requires a lot of electricity and the fans are notoriously loud. Agonafer, the UT-Arlington engineering professor, said air cooling can account for up to 40% of an average data center's total energy demand. Seeking a more energy-efficient solution, the industry started exploring cooling servers with water. One method is through evaporative cooling: The hot air generated by servers is blown through a water-soaked material. As the water evaporates, it pulls heat from the air, which is then circulated around the servers to keep them from overheating. There's also liquid cooling, a method more AI data centers are adopting. Instead of cooling the air, this method cools the hardware directly. Cold plates are attached to individual chips, and a liquid -- usually a mix of water and a chemical coolant -- circulates through the plates, pulling heat away at the source. There's also closed-loop cooling systems, a technique that will be used at one of the world's largest AI data center now being built outside of Abilene as part of Stargate. In a closed-loop system the cooling fluid -- the Abilene project will use water -- circulates in a sealed circuit. The fluid absorbs heat from the electronic components, gets turned into vapor, then is condensed back into liquid, continuously recirculating through the system. The data center will initially use 8 million gallons of Abilene city water to fill the cooling system. Jeremiah Bihl, an assistant director at Abilene's water utility, said that total is minimal compared to the city's average daily use of 22 million gallons. After that initial fill, the system will need some additional water, but Bihl couldn't say how much or how often. He added that occasionally the system will also need to be completely drained and refilled for maintenance. But he said the center's water use isn't expected to put a deep dent in the city's overall water supply. "We may be the white elephant of all the other AI centers out there, just because of the kind of cooling system that (the data centers) have," he said. Tracking how much water data centers use is complicated Texas's water planning process is built on historical data. Long-term projections of the state's water needs by the Texas Water Development Board rely on past usage trends and historical reported water use. Experts say this model can leave the state's water supply vulnerable to fast-moving growth of new industries like data centers. "If you're having an industry that's having some kind of exponential growth in the state or just growing leaps and bounds, that historical data doesn't give you enough to go on to plan for future use," said Julie Nahrgang, executive director of the Water Environment Association of Texas. For example, generative AI, such as ChatGPT, uses graphics processing units, which are more complex than typical computer processors and generate more heat -- which requires more water for cooling, said Ding, the Purdue professor. The state has encountered similar issues before. When hydraulic fracturing first took off in the state's vast oil fields, state planners were caught off guard by the immense water demand, said Carlos Rubinstein, a water expert and a former commissioner for the Texas Commission on Environmental Quality. He said the industry found alternatives like using brackish groundwater and reusing water. "We managed it OK, but you can only keep managing it OK for so long," Rubenstein said. A similar scenario is unfolding now with data centers. There is no uniform method for reporting or tracking their water usage, and as a result, the state can't accurately predict their future impact on an increasingly precious natural resource, according to experts. "We have a shortage (of) water, but on the other side, we like the Texas economy," Rubenstein said. "It's a tough balance, but we are coming to terms with the fact that water in Texas is the most limiting factor, and we need to find an answer (for meeting both)." In 2024, the state water board tried to get a handle on data centers' water consumption by sending surveys to nearly 70 data centers asking how much ground and surface water they used each month and who was supplying that water to them. Even though a response was required by state law, only one-third of them responded. The consequences for noncompliance are minimal -- a Class C misdemeanor with a maximum fine of $500. In practice, the system relies on voluntary compliance or self-reporting. TWDB then validates the data in those reports. If the state knew what kind of cooling system each facility used as well as its electricity consumption, it could better estimate water consumption. But experts say companies do not always disclose that information for proprietary reasons. In response to these challenges, the Water Development Board said it's participating in industry conferences, reviewing new research, and working to switch data centers to another category of water user to better track them in future water plans. However, experts say data centers will likely not be reflected in state planning documents until 2032. Cook, the HARC researcher, said she's heard from nervous regional planners and groundwater conservation experts. Planning based on historical data "means that all of the planning that they've done is inaccurate," Cook said. "We need to speed up." Doctoral students take on newer cooling technology Inside the UT-Arlington lab, a small team is gathered around a table where the inner workings of a data center cooling machine have been laid out on a white table piece by piece. Unlike the Abilene data center's closed-loop cooling system, this technology uses a chemical mixture instead of water and cools the hardware directly. The refrigerant circulates through a plate mounted directly to the processing unit, the brain of a computer. As the liquid absorbs heat, it evaporates, carrying heat away before being condensed and recirculated, forming a closed loop. This change from liquid to gas and back to liquid is a far more efficient way to remove heat than traditional air or liquid-only cooling systems, said Agonafer, the engineering professor. "It's kind of like a refrigerator," said Smith, the engineering doctoral student. The new technology has cut the energy an average data center uses to cool its equipment from about 40% of total energy use to 5% during lab testing. Though the current demo involves just one server, plans are in place to scale to 40 as the team fine-tunes liquid flow rates, temperature control, and pressure to minimize energy use. It's part of a larger, federally backed effort to transform how data centers are cooled. With a $2.84 million grant from the U.S. Department of Energy's COOLERCHIPS program, UT-Arlington is leading one of 19 projects aimed at slashing the energy and water demands of data center infrastructure. As part of the project, Accelsius, an Austin-based company specializing in data center cooling technology, donated a rack with servers for students to test their system. The project is close to being ready for commercial use, Agonofer, the professor leading the student team, said. Pundla, the doctoral candidate, said the technology is already drawing interest from data center companies. When the project is complete the students plan to launch a start-up company to sell the technology, with the university where it was developed as shareholder. While there are no regulations in Texas requiring companies to adopt such systems, companies selling cooling system technology say data centers have a big incentive to bring down their power demand. Texas law now requires large load consumers including data centers to consult with the state's main grid operator and provide details on how much power they plan to use and whether a project is planned in areas where the grid is already facing supply constraints. During this process data centers join a waitlist. Finding ways to cut energy is more cost-efficient, said Liz Cruz, chief marketing officer at Accelsius. She said the companies she works with prefer to use electricity for computing, not cooling. "Data centers are huge energy and water usage hogs," said Liz Cruz, chief marketing officer at Accelsius. "So it is our responsibility as the builders of (cooling systems) to make the most sustainable decision." ___ This story was originally published by The Texas Tribune and distributed through a partnership with The Associated Press.

Graduate Students at UT-Arlington Are Developing a Way to Cool Data Centers Without Water

ARLINGTON, Texas (AP) -- Sai Abhideep Pundla has been awake since 3 a.m. After a red-eye flight from Las Vegas, where he briefed data center company executives and local government officials about the future of artificial intelligence, he's back in a lab at UT-Arlington, tinkering with a prototype he thinks could solve one of the industry's biggest challenges: how to keep data centers cool without draining finite water supplies. Pundla, a doctoral candidate in engineering, is testing a system that cools the computer servers using a recirculating chemical refrigerant instead of water. It's a timely innovation. Texas is building dozens of massive data centers -- some as large as New York's Central Park -- and experts say they're expected to guzzle millions of gallons of water a year in a state facing an increasingly urgent water crisis. Every snap of a photo, message sent, or Google search requires data, which has to go somewhere. That "somewhere" is a data center. These massive facilities filled with servers that store and process everything we do online keep our digital lives intact. However, keeping all that data takes electricity to power the data centers and cooling systems to keep their equipment from malfunctioning. And both of those require water. While data centers currently consume a small portion of the state's total water supply, according to the state's water plan, some researchers warn that with droughts and population growth, data centers have the potential to help push water supplies to the brink -- especially in Texas' more arid regions. The Houston Advanced Research Center, an independent nonprofit research organization focusing on sustainability solutions, estimates that existing data centers in Texas will consume approximately 25 billion gallons of water, or 0.4% of the state's total water use in 2025. By 2030, this demand could increase up to 2.7% of the total annual water use in Texas. That's the equivalent water consumption of 1.3 million average U.S. households. The state currently has more than 400 data center facilities, with about 70 more on the way. "(It) may not feel like it's a whole lot at the state level," said Margaret Cook, a researcher studying data center water use at HARC. But if massive data centers locate in small communities, she said, they may not be able to handle big jumps in water demand. In the Texas Panhandle, for example, Amarillo residents recently held a community event to oppose five planned data centers and inform others of the potential risks to the Ogallala Aquifer, the region's main water source that is being drained faster than it can be replenished. "Our water sources aren't as reliable as what would be needed for the data center," said Madison Boyle, an event organizer. This growing concern has highlighted the state's limitations in tracking and forecasting water usage by emerging industries. In Texas, companies are required to report historical water consumption, but aren't required to report how much water they expect to consume in the future or where it would come from, making it difficult for communities to plan or better manage their water supplies. Other states are starting to act on data centers' water use. California lawmakers recently passed a bill, which is awaiting the governor's signature, that will require new data centers to report their projected water use before they start operations, while in Minnesota, data center developers are required to consult with the state's environmental agency to make sure their proposed location has an adequate water supply. Some data centers in other states are built near lakes, rivers or in colder climates where there's natural free cooling. In Texas, many data centers are located in areas where water supplies already are under high stress, according to research by Yi Ding, an assistant professor at Purdue University's School of Electrical and Computer Engineering. "You could consume the same amount of water in Texas and Iowa," Ding said. "But the environmental burden is different because Texas is more dry." In the lab in Arlington, Pundla and his colleague Braxton J. Smith flip on power to a server packed with memory modules, processor chips and fans -- the same components that make up massive data centers. A hum kicks in as vapor from a special liquid cools the system. No water involved. Pundla said the industry knows their water and energy consumption is unsustainable. "The thermal management side of data centers is now trying to play catch up" as the industry develops bigger, more powerful computing systems that need more electricity and generate more heat, he said. The urgency to find solutions is growing. Earlier this year, President Donald Trump announced a $500 billion federal push, dubbed Stargate, to build AI data centers and their supporting power infrastructure in Texas. Without smarter cooling, Pundla's engineering professor Dereje Agonafer said, it may come at a steep environmental cost: intensifying water scarcity, forcing homes and farms to compete with industry and disrupting aquatic biodiversity. How data centers use water Data centers use water in two main ways: directly, to cool their equipment, and indirectly, through the electricity they rely on. The current Texas data center market requires 10,234 megawatts of electricity a year, which is equivalent to more than 8 million homes' usage a month, according to the research firm Baxtel that tracks data centers and their development. Data centers are expected to help drive Texas' power demand to nearly double by 2030. Electricity generation also requires water, particularly to cool the natural gas, coal or nuclear power plants that generate most of the state's electricity. That means the more electricity a data center uses, the more water is indirectly consumed. Cook said whether data centers choose to connect to the grid or build their own power supply -- some in Central Texas already are building on-site power plants -- will impact their water needs for cooling. The bigger driver of water use is the need to cool the data center's servers, which run 24/7, generating tremendous heat that can cause servers to fail if they're not properly cooled. Over the years, different methods have evolved to solve the cooling problem. Traditional systems use arrays of big fans to blow cool air across the machines, but that requires a lot of electricity and the fans are notoriously loud. Agonafer, the UT-Arlington engineering professor, said air cooling can account for up to 40% of an average data center's total energy demand. Seeking a more energy-efficient solution, the industry started exploring cooling servers with water. One method is through evaporative cooling: The hot air generated by servers is blown through a water-soaked material. As the water evaporates, it pulls heat from the air, which is then circulated around the servers to keep them from overheating. There's also liquid cooling, a method more AI data centers are adopting. Instead of cooling the air, this method cools the hardware directly. Cold plates are attached to individual chips, and a liquid -- usually a mix of water and a chemical coolant -- circulates through the plates, pulling heat away at the source. There's also closed-loop cooling systems, a technique that will be used at one of the world's largest AI data center now being built outside of Abilene as part of Stargate. In a closed-loop system the cooling fluid -- the Abilene project will use water -- circulates in a sealed circuit. The fluid absorbs heat from the electronic components, gets turned into vapor, then is condensed back into liquid, continuously recirculating through the system. The data center will initially use 8 million gallons of Abilene city water to fill the cooling system. Jeremiah Bihl, an assistant director at Abilene's water utility, said that total is minimal compared to the city's average daily use of 22 million gallons. After that initial fill, the system will need some additional water, but Bihl couldn't say how much or how often. He added that occasionally the system will also need to be completely drained and refilled for maintenance. But he said the center's water use isn't expected to put a deep dent in the city's overall water supply. "We may be the white elephant of all the other AI centers out there, just because of the kind of cooling system that (the data centers) have," he said. Tracking how much water data centers use is complicated Texas's water planning process is built on historical data. Long-term projections of the state's water needs by the Texas Water Development Board rely on past usage trends and historical reported water use. Experts say this model can leave the state's water supply vulnerable to fast-moving growth of new industries like data centers. "If you're having an industry that's having some kind of exponential growth in the state or just growing leaps and bounds, that historical data doesn't give you enough to go on to plan for future use," said Julie Nahrgang, executive director of the Water Environment Association of Texas. For example, generative AI, such as ChatGPT, uses graphics processing units, which are more complex than typical computer processors and generate more heat -- which requires more water for cooling, said Ding, the Purdue professor. The state has encountered similar issues before. When hydraulic fracturing first took off in the state's vast oil fields, state planners were caught off guard by the immense water demand, said Carlos Rubinstein, a water expert and a former commissioner for the Texas Commission on Environmental Quality. He said the industry found alternatives like using brackish groundwater and reusing water. "We managed it OK, but you can only keep managing it OK for so long," Rubenstein said. A similar scenario is unfolding now with data centers. There is no uniform method for reporting or tracking their water usage, and as a result, the state can't accurately predict their future impact on an increasingly precious natural resource, according to experts. "We have a shortage (of) water, but on the other side, we like the Texas economy," Rubenstein said. "It's a tough balance, but we are coming to terms with the fact that water in Texas is the most limiting factor, and we need to find an answer (for meeting both)." In 2024, the state water board tried to get a handle on data centers' water consumption by sending surveys to nearly 70 data centers asking how much ground and surface water they used each month and who was supplying that water to them. Even though a response was required by state law, only one-third of them responded. The consequences for noncompliance are minimal -- a Class C misdemeanor with a maximum fine of $500. In practice, the system relies on voluntary compliance or self-reporting. TWDB then validates the data in those reports. If the state knew what kind of cooling system each facility used as well as its electricity consumption, it could better estimate water consumption. But experts say companies do not always disclose that information for proprietary reasons. In response to these challenges, the Water Development Board said it's participating in industry conferences, reviewing new research, and working to switch data centers to another category of water user to better track them in future water plans. However, experts say data centers will likely not be reflected in state planning documents until 2032. Cook, the HARC researcher, said she's heard from nervous regional planners and groundwater conservation experts. Planning based on historical data "means that all of the planning that they've done is inaccurate," Cook said. "We need to speed up." Doctoral students take on newer cooling technology Inside the UT-Arlington lab, a small team is gathered around a table where the inner workings of a data center cooling machine have been laid out on a white table piece by piece. Unlike the Abilene data center's closed-loop cooling system, this technology uses a chemical mixture instead of water and cools the hardware directly. The refrigerant circulates through a plate mounted directly to the processing unit, the brain of a computer. As the liquid absorbs heat, it evaporates, carrying heat away before being condensed and recirculated, forming a closed loop. This change from liquid to gas and back to liquid is a far more efficient way to remove heat than traditional air or liquid-only cooling systems, said Agonafer, the engineering professor. "It's kind of like a refrigerator," said Smith, the engineering doctoral student. The new technology has cut the energy an average data center uses to cool its equipment from about 40% of total energy use to 5% during lab testing. Though the current demo involves just one server, plans are in place to scale to 40 as the team fine-tunes liquid flow rates, temperature control, and pressure to minimize energy use. It's part of a larger, federally backed effort to transform how data centers are cooled. With a $2.84 million grant from the U.S. Department of Energy's COOLERCHIPS program, UT-Arlington is leading one of 19 projects aimed at slashing the energy and water demands of data center infrastructure. As part of the project, Accelsius, an Austin-based company specializing in data center cooling technology, donated a rack with servers for students to test their system. The project is close to being ready for commercial use, Agonofer, the professor leading the student team, said. Pundla, the doctoral candidate, said the technology is already drawing interest from data center companies. When the project is complete the students plan to launch a start-up company to sell the technology, with the university where it was developed as shareholder. While there are no regulations in Texas requiring companies to adopt such systems, companies selling cooling system technology say data centers have a big incentive to bring down their power demand. Texas law now requires large load consumers including data centers to consult with the state's main grid operator and provide details on how much power they plan to use and whether a project is planned in areas where the grid is already facing supply constraints. During this process data centers join a waitlist. Finding ways to cut energy is more cost-efficient, said Liz Cruz, chief marketing officer at Accelsius. She said the companies she works with prefer to use electricity for computing, not cooling. "Data centers are huge energy and water usage hogs," said Liz Cruz, chief marketing officer at Accelsius. "So it is our responsibility as the builders of (cooling systems) to make the most sustainable decision." ___ This story was originally published by The Texas Tribune and distributed through a partnership with The Associated Press.