{"slug": "fear-of-data-centers-outpaces-knowledge-about-them", "title": "Fear of data centers outpaces knowledge about them", "summary": "Utah residents fear the proposed Stratos data center in Box Elder County will increase water use, electricity costs, emissions and heat, but experts say water consumption varies widely and data centers have not typically raised residential electricity rates. The 9-gigawatt project, backed by Kevin O'Leary, is one of over 1,800 U.S. data centers under development, and concerns are fueled by misinformation about water use.", "body_md": "- Utah residents have raised concerns that the proposed Stratos data center in Box Elder County could increase water use, electricity costs, emissions and heat in the region.\n- Estimates about data centers' water use vary widely, and researchers say older claims that each AI prompt consumes a bottle of water are not accurate.\n- Data center water use depends heavily on design, location and cooling methods, with some facilities using far more water than others.\n- Data centers generally have not caused increases in residential electricity rates, with other factors like fuel prices and state policies playing larger roles.\n\nLast week in Springville, Utah, a group of young moms sat in Memorial Park, watching their toddlers play. In the high-70-degree weather, atop the park’s green grass, the conversation turned to data centers.\n\nOne mom had mused how she was going to keep her new infant cool this summer, and Matty Shmitz responded with worry that the approved data center in Box Elder County was going to make the summer’s heat worse.\n\nIn a later conversation with the Deseret News, Shmitz described the data center as a “constant numb pain in the back of my brain.”\n\nThrough TikTok, Instagram and news reports, Shmitz said she has developed serious concerns about the data center.\n\nWill it use an inordinate amount of water? Where will that water come from? Will it drive up the cost of electricity and increase the temperature? And at the root of it, why are we risking Utah’s beautiful land to fuel artificial intelligence, when it seems like artificial intelligence will hurt American society?\n\nThe Stratos Project is one of more than 1,800 U.S.-based data centers in various stages of development. Once built, it will join more than 3,100 data centers already in operation across the United States, the earliest of which were built in the 1990s.\n\nBacked by “Shark Tank” investor Kevin O’Leary, the project as planned is massive compared to other data centers in the U.S. At full buildout, it will have a power capacity of 9 gigawatts — roughly 90 to 225 times larger than the average data center, which uses 40 to 100 megawatts.\n\nThese existing data centers power the digital infrastructure behind the internet.\n\nIf you store images in the cloud, buy anything with a debit card, ask Siri a question, stream a show on Netflix, or load directions on Google Maps, you’re using a data center.\n\nBut data centers’ everyday utility has been lost in a haze of anxiety about new proposals. Shmitz is not alone in her concerns. More than half of Utah’s residents say they oppose the data center in Box Elder County, according to a [Deseret News-Hinckley Institute poll](https://www.deseret.com/politics/2026/05/20/new-poll-reveals-how-voters-feel-about-ai-data-center-in-utah-backed-by-kevin-oleary/) conducted mid-May.\n\nSo it makes sense to ask: Are concerns about data centers well-founded?\n\n### The water consumption confusion\n\nIn 2024, [The Washington Post](https://www.washingtonpost.com/technology/2024/09/18/energy-ai-use-electricity-water-data-centers/) released a report claiming that a 100-word email written by ChatGPT consumes an entire bottle of water or 519 milliliters.\n\nThe article proceeded to scale up that number: one ChatGPT-generated email a week for a year consumes 27 liters of water; one ChatGPT-generated email a week for a year from 10% of the U.S. population (16 million people) requires more than 435 million liters — “equal to the water consumed by all Rhode Island households for 1.5 days.”\n\nReaders were then left to fill in the blanks about usage. In 2025, ChatGPT was [queried about 2.5 billion times](https://techcrunch.com/2025/07/21/chatgpt-users-send-2-5-billion-prompts-a-day/) a day.\n\nBased on the Post’s measurements, 2.5 billion queries (if they were all the size of an email) would require nearly 343 million gallons of water — more than 520 Olympic-sized swimming pools — a day.\n\nWhen Andy Masley, a former physics teacher turned writer, saw this report, it didn’t sit well with him. So he started looking into the article’s methodology, then reached out to the researcher tapped for the calculation, Shaolei Ren, an associate professor of electrical and computer engineering at the University of California, Riverside.\n\nDescribing ChatGPT as a single model that has universal resource needs is incorrect\n\n— Shaolei Ren, associate professor at the University of California, Riverside\n\n“The majority of that bottle of water, even in (Ren’s) own estimates, isn’t actually used in the data center itself. It’s used in offsite power plants to generate the electricity,” Masley told the Deseret News.\n\nAbout half of the water-bottle estimate comes from evaporation off lakes dammed by hydro plants to generate hydropower.\n\n“I’ve been emailing back and forth with (Ren) who made this estimate, and he agrees with me that the actual estimate is way, way smaller,” Masley said.\n\nMasley said more current estimates, from experts like [EcoLogits](https://huggingface.co/spaces/genai-impact/ecologits-calculator), find that individual prompts cost between 1-10 milliliters of water, about 99% smaller than the estimate published by the Washington Post. The amount of water actually used in the data center is smaller still — about 0.2-2 milliliters.\n\nIn a conversation with the Deseret News, Ren said the Washington Post’s report should not be considered an accurate measure of today’s artificial intelligence water demands.\n\n“We cannot just use a number from two years ago to describe today’s system. And we cannot use specific models’ results to generalize other models. I could give you a super efficient AI model that uses almost zero resources, or I could give you a very large model that can — it’s just never correct to say, ‘AI uses this much water,’” Ren said.\n\nEven describing ChatGPT as a single model with universal resource needs is incorrect, he continued. “ChatGPT is not a single model. Nobody knows what exactly they’re doing under the hood. So I don’t know the resource efficiency,” Ren said.\n\nA chatbot’s efficiency depends on a spread of factors, he explained. To make AI more efficient, engineers optimize its hardware, algorithms and how workloads are scheduled. Its efficiency can change dramatically if it’s in thinking mode, reasoning mode or plain text output, as well as a myriad of other ways.\n\nBasing current AI efficiency off a source from several years ago is a recipe for misinformation, Jonathan Koomey, a researcher, formerly at Lawrence Berkeley National Laboratory, who’s been studying data centers for more than 25 years, told the Deseret News.\n\n“One of the things about computing equipment that is hard for people to get their head around is that it changes fast,” he said. “The problem is that three or four years is an eternity when it comes to computing equipment. Things have probably turned over a couple times in terms of the latest tech since then.”\n\nIt’s not as simple as saying, ‘Well, we shouldn’t let data centers use water,’ because if you do that, the cooling will be less efficient\n\n— Jonathan Koomey, an independent researcher, formerly at Lawrence Berkeley National Laboratory\n\n### So how much water does a data center use?\n\nData centers have extreme variability in terms of water use.\n\n“Averages don’t mean anything,” Koomey said. “It’s very specific to data center designs and their locations.”\n\nFor example, Meta’s data center in Eagle Mountain, Utah, used [35.1 million gallons of water](https://sustainability.atmeta.com/wp-content/uploads/2025/10/Meta_2025-Environmental-Data-Index.pdf) in 2024. The facility is 4.5 million square feet and used 1.1 million megawatt hours of energy that year. Meanwhile, Meta’s Prineville, Oregon data center — a 4.6 million square-foot facility — used 1.7 million megawatt hours of energy and consumed about 86.6 million gallons of water.\n\nMost water-cost estimates include both direct and indirect water use. Direct water use happens inside the data center itself, primarily to cool servers, and indirect water use largely comes from the water required to generate electricity to power the facility.\n\nSo a low-water data center would need to be powered by a low-water energy source and use low-water technology within its walls.\n\nKoomey warned that reducing data center water use is not always a net benefit.\n\n“People use water because it’s very energy efficient to cool with,” he explained. “Some data center companies use cooling towers, and that’s a relatively water intense way to do it, but it’s also incredibly energy efficient. And so that saves electricity at the data center.”\n\nIf a developer opts to reduce water by cooling with air, the air-cooling system will likely require a larger amount of water in electricity generation than the water-cooling system would demand.\n\n“It’s not as simple as saying, ‘Well, we shouldn’t let data centers use water,’ because if you do that, the cooling will be less efficient,” Koomey said. “You’ll use more electricity, then you’ll use more water because it turns out the water intensity of electricity generation is actually pretty high.”\n\nWithin a data center, the amount of water required to process a single task is similarly varied. According to peer-reviewed [research](https://www.sciencedirect.com/science/article/abs/pii/S0921344925001892) published last summer, a single task’s water cost varies by a magnitude of 10,000.\n\nNorthern Utah’s Stratos Project has a wide range for how much water it is estimated to use. Project developers say they will likely use between 500 and 1,100 acre-feet of water annually — or 163 million to 358 million gallons of water.\n\nUtah Gov. Spencer Cox [has said](https://www.deseret.com/environment/2026/05/21/utah-emergency-order-drought/) the Stratos Project will use less water than the area currently uses. “There are issues that people should be concerned about for use, but water is not one of them in this particular case,” he said in late May.\n\n### How do data centers compare to other industries in water intensity?\n\nIn a conversation with the Deseret News, Neil Chilson, the head of AI policy at the Abundance Institute, compared data centers’ water intensity with other industries.\n\nMany data centers use “about the same water as a medium sized brewery,” he told the Deseret News.\n\nIn 2021, U.S. data centers collectively used 449 million gallons of water per day, accounting for 0.14% of the country’s total daily water consumption, according to the [University of Georgia](https://fieldreport.caes.uga.edu/publications/TP121/how-data-centers-impact-surface-and-ground-waters/).\n\nFor context, other common goods consume an enormous amount of water. A single walnut takes about [five gallons of water](https://www.almonds.org/why-almonds/growing-good/water-wise) to produce, and a hamburger takes about [660](https://www.latimes.com/food/dailydish/la-dd-gallons-of-water-to-make-a-burger-20140124-story.html).\n\nIn 2018, the average walnut farm yielded about 1.93 tons per acre; a 100-acre farm produces about 193 tons of in-shell walnuts. With the average weight of a single in-shell walnut at [11.7 grams](https://www.chinchiolofarming.com/blogs/the-walnut-chronicle/unveiling-the-weight-of-a-single-in-shell-walnut-a-precise-measure), 193 tons would include 15 million walnuts. So the walnuts produced on a 100-acre farm would consume 75 million gallons of water.\n\nBy comparison, Meta’s Eagle Mountain data center, which is also about 100 acres, consumed about 35 million gallons in 2024 — about half that of a 100-acre walnut farm.\n\nMeanwhile, Americans consume about 50 billion burgers a year. If it takes [660 gallons of water](https://www.latimes.com/food/dailydish/la-dd-gallons-of-water-to-make-a-burger-20140124-story.html) to produce one burger, American burger consumption eats about 33 trillion gallons of water a year.\n\n### Data centers’ thermal load\n\nDr. Robert Davies, a professor of physics at Utah State University, released preliminary analysis on the thermal load, or “waste heat,” of a 9 gigawatt natural gas plant, if built in Hansel Valley where the Stratos Project is proposed.\n\nHe suggested generating the power needed for the data center could [potentially warm](https://www.deseret.com/politics/2026/05/11/utah-governor-cox-defends-box-elder-data-center-as-important-to-national-security-in-ai-arms-race-with-china/) the valley in Box Elder County by several degrees.\n\nDavies did not have a formal uncertainty analysis for his study, and he was unwilling to share his prompt data, results and assumptions with the Deseret News.\n\n“I want to emphasize, this is a relatively simple estimation, not intended to be rigorous,” he told the Deseret News. “Just demonstrating that there’s reason to believe the effect will be significant.”\n\nThe challenge with something like Stratos is that it’s very unusual in that it’s the data center plus the power plant that’s powering the data center.\n\n— Dr. Robert Davies, a professor of physics at Utah State University\n\nIf the Stratos Project is solely powered by an advanced natural gas plant, Davies estimated that it would impose 16 gigawatts of heat on the area. It would be the equivalent of 40,000 Walmart super centers all burning energy at the rate of about 0.4 megawatts, he said.\n\n“The challenge with something like Stratos is that it’s very unusual in that it’s the data center plus the power plant that’s powering the data center. That’s what makes the thermal load so big and so concentrated,” he said.\n\nDavies estimated that an area in the valley, ranging in size from about 116 square miles to 386 square miles, would experience temperature increases by 2℉–5℉ in the day and 8℉–12℉ at night.\n\n“With the numbers I came up with, (dew point) is largely wiped out,” he said. “So you don’t get the overnight dew and frost, which is a large source of water for that ecosystem.”\n\nDavies said he hopes the state conducts thorough ecological analysis on what effects a 16 gigawatt thermal load would have.\n\nHis preliminary thermal load estimate has spread throughout Utah. In her conversation with the Deseret News, Springville resident Matty Shmitz said, “I’ve been told is that (the Stratos Project) would raise the temperature of the state by like five degrees.”\n\nWhile the scale of the Stratos Project is unprecedented, scientists have been able to study a close equivalent, the Urban Heat Island Effect. Cities around the world trap and generate massive amounts of heat, raising areas’ temperatures by several degrees, especially at night. The Stratos Project may essentially create an urban heat island in the middle of a rural valley.\n\nFor example, some areas of New York City are [9.7 degrees hotter](https://www.climatecentral.org/graphic/urban-heat-islands-2024?graphicSet=Unequal+Urban+Heat&location=CONUS&lang=en) than they would be without the city’s massive population and infrastructure.\n\nSeveral days after the Stratos Project was proposed to use solely natural gas, Utah Gov. Cox said project developers were willing to use a variety of clean and renewable power sources.\n\n### The AI boom and electricity demand\n\nIn 2023, data centers made up about [4.4%](https://www.congress.gov/crs-product/R48646) of the country’s annual electricity consumption.\n\nKoomey, who has researched American energy for decades, described the current electricity market as having “moderate demand growth.”\n\n“But it’s not all data centers. It’s not even mostly data centers,” he told the Deseret News. “ChatGPT was announced in 2022. In 2023, electricity was the same as 2018. Zero growth. In 2024, growth was 2.1%. In 2025, growth was 2.5%. That’s total growth for the U.S.”\n\nHe continued, “There has been growth in the last couple of years — a chunk of that is data centers — but it’s not explosive growth. It’s just moderate growth; 2% is normal.\"\n\nKoomey said the hype around data centers reminds him of the dot-com bubble at the emergence of the internet. When the bubble burst in 2000, \"[97%](https://thetimelessinvestor.substack.com/p/they-buried-a-trillion-dollars-underground) of the fiber network was unused,\" he said.\n\n“I don’t think anyone can tell you how much demand there’s going to be even in three years, but having been through these cycles, these hype cycles, several times, I try to caution people not to assume that the conventional wisdom AI boom forecast is necessarily the right one,” Koomey said.\n\n### Will data centers raise electricity costs?\n\nIn the case of the Stratos Project, it is natural to assume that such a monumental increase in electricity demand would lead to increased rates for Utahns.\n\nHowever, during the state’s 2025 legislative session, lawmakers passed [a bill](https://le.utah.gov/~2025/bills/static/SB0132.html) requiring any power development for data center projects to not negatively impact electricity costs for Utahns. All power consumed by the data center will be generated onsite or brought in.\n\nFor those living in states without such laws, empirical evidence suggests that data center construction does not meaningfully lead to increased utility bills.\n\nA University of Southern California professor, Shon Hiatt, recently [presented research](https://www.deseret.com/business/2026/04/15/utah-data-centers-artificial-intelligence-ai-water-energy-electricity-use-economic-impacts-jobs-tax-incentives/) that shows data centers are responsible for only a [0.007%-0.08%](https://sutherlandinstitute.org/wp-content/uploads/2026/05/Building-a-Human-Centered-Digital-Future-Part-1-Data-Centers.pdf) increase in residential power bills.\n\nDifferent research from [Charles River Associates](https://www.crai.com/insights-events/publications/us-retail-electricity-rate-trends-analysis/) found that data centers “have generally not caused retail rate increases,” with the exception of the PJM — a region stretching from the midwest to the mid-Atlantic.\n\nThe [Lawrence Berkeley National Laboratory](https://www.congress.gov/crs-product/R48646) ran a study between 2019 and 2025, which found that data centers did not influence electricity prices in a major way across most areas of the U.S. Instead, the biggest drivers in price hikes were natural gas prices, natural disasters and state energy and environmental policies.", "url": "https://wpnews.pro/news/fear-of-data-centers-outpaces-knowledge-about-them", "canonical_source": "https://www.deseret.com/u-s-world/2026/06/21/what-environmental-costs-do-data-centers-have/", "published_at": "2026-06-22 03:00:01+00:00", "updated_at": "2026-06-22 03:15:27.451950+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-infrastructure", "ai-ethics"], "entities": ["Stratos Project", "Box Elder County", "Kevin O'Leary", "Deseret News", "Hinckley Institute", "The Washington Post", "ChatGPT", "Springville"], "alternates": {"html": "https://wpnews.pro/news/fear-of-data-centers-outpaces-knowledge-about-them", "markdown": "https://wpnews.pro/news/fear-of-data-centers-outpaces-knowledge-about-them.md", "text": "https://wpnews.pro/news/fear-of-data-centers-outpaces-knowledge-about-them.txt", "jsonld": "https://wpnews.pro/news/fear-of-data-centers-outpaces-knowledge-about-them.jsonld"}}