The Hidden Cost of AI: How Data Centers Are Straining Water, Power, and Communities

by Ella Mrofka / February 6th, 2026

Editor’s Note: This is the second of two Dispatches on AI data centers; read the previous article, by Alefiya Presswala, here.

The residents of Saline Township, Michigan, are upset. Soon, construction will begin on a 2.2 million-square-foot data center located just outside the small rural town, which will ultimately serve two Big Tech companies: Oracle and OpenAI. Area residents oppose the facility because they fear it will strain the region’s power grid and increase their electricity bills, disrupt the area’s farmland, and consume much of the area’s water (as the plant will ultimately use more water than any other single user in the Great Lakes region). Some residents are taking their concerns to court, while others believe these new data centers can lead by example if they protect the environment during building and usage.

Many other cities around the United States are also fighting the construction of new data centers, and for similar reasons: concern that they will lead to water shortages and soaring energy bills. Community members are signing petitions, and, according to Data Center Watch, $98 billion worth of data center investments have been blocked as communities organize to fight for more and stricter regulations on these massive facilities and their demand for water and energy.

Water Usage

Big tech companies such as Amazon, Google, and Microsoft are continuing to expand their data centers due to the rapid rise of AI. An April 2025 study by SourceMaterial and The Guardian determined that these companies plan to expand the number of data centers by 78 percent.

For AI to answer a prompt, it performs many calculations, all of which generate heat. Data centers use vast amounts of water to cool servers and other equipment. The Washington Post reports that large data centers can use up to five million gallons of water a day, equivalent to the daily water needs of 10,000 to 50,000 people.

According to a new study summarized in a December 17 article in The Verge, AI could consume between 312.5 and 764.6 billion liters of water in 2025. To put that figure in perspective, the country’s largest metropolis, New York City, uses approximately 1 billion gallons, or 4.5 billion liters, of water per day. So, in 2024, existing AI data centers sucked up as much water as a city of 9 million people used in 5 months.

In 2023, Microsoft acknowledged that 42 percent of its water came from areas experiencing water stress, while Amazon did not report a figure. It is common to build data centers in dry regions to decrease the risk of corrosion damage to their servers. This affects water-scarce areas like Arizona and Virginia, where many US data centers are being built. Newton County,  Georgia, is on track to face a water deficit by 2030 after Meta broke ground on a $750 million data center, according to the New York Times.

Northern Virginia is home to the most data centers in the world, even as the state continues to experience drought and increased water scarcity. Roughly 70 percent of global internet traffic runs through data centers in Virginia. The Virginia Data Center Reform Coalition fights for transparency from these big tech companies to connect data centers and water issues. Without full transparency, these companies can continue to target water-stressed areas without public understanding of how natural resources are being depleted.

For instance, Mike Doble, a communications advisor for the Virginia Data Center Reform Coalition, explains that demand for electricity in Virginia is expected to double or triple “over the next 15 years. … an unprecedented amount of increase.”

Microsoft claims that it will build a zero-water data center by 2027 and will be water offsetting alongside Google by 2030, but this does not change the fact that hundreds of already operational data centers are all highly water-reliant.

Energy Usage

As AI growth accelerates, energy usage will also skyrocket. According to a 2024 report by the Lawrence Berkeley National Laboratory, data center energy use has tripled over the past decade and is projected to double or triple again by 2028. Data centers are projected to account for nine percent of global electricity consumption by 2030.

Scientists and researchers argue that AI companies also fail to fully disclose how much energy their current models use or the projected demands of future projects.

In 2024, natural gas accounted for forty percent of the energy used by US data centers, according to the International Energy Agency (IEA). Renewables, nuclear power, and coal also supplied electricity to data centers. Natural gas is projected to remain the largest source of energy for data centers through 2030. Focusing on increasing renewable energy use in data centers should be a top priority.

Increasing energy demands by data centers affect electricity bills for Americans. According to a Bloomberg News analysis, electricity costs in areas near data centers are as much as 267 percent higher than they were five years ago. People located near data centers are not the only ones feeling the effects. A July 2025 report by researchers at Carnegie Mellon University found that the average US electricity bill could increase by eight percent by 2030 due to data centers and cryptocurrency mining.

The “simple” solution, according to the New York​​ Times, is to ask big tech companies to draw less power from the grid when it is most strained, relying on batteries to power operations during those periods. However, this is not a realistic solution because there is no law requiring big tech companies to do this.

Resistance and Solutions

A coalition of 230 environmental groups is campaigning for a halt to the construction of new data centers, due to environmental concerns, The Guardian reported in December 2025. “The rapid, largely unregulated rise of datacenters to fuel the AI and crypto frenzy is disrupting communities across the country and threatening Americans’ economic, environmental, climate and water security,” according to a letter sent to Congress by members of the coalition, which includes national and international groups such as Food & Water Watch, Greenpeace, and Oil Change International, as well as regional and state-based organizations.

Given AI’s harmful environmental impacts, steps should be taken to slow and prevent the construction of new data centers until they are built in ways that do not contribute to further environmental destruction.

In order for AI and data centers to become more sustainable, companies would need to be transparent about how much energy and water they use. Without accurate data from all major tech companies, the effectiveness of potential solutions is limited.

Doble, the spokesperson with the Virginia Data Center Reform Coalition, recommended that “the place to make your mark is at the local and state level, knowing who’s making the decisions, because right now, big businesses are making the decisions.” To achieve data center transparency, state lawmakers must enact policy.

In 2023, Oregon lawmakers introduced a bill requiring data centers to run on entirely clean energy by 2040. Amazon lobbied against the bill, leading to its demise.

Using solar and wind energy in data centers instead of relying on unsustainable sources such as natural gas, coal, and oil can change future projections. The operation and maintenance of solar energy systems are up to 60 percent more cost-efficient than conventional technology. Solar and wind energy are becoming more efficient and should be used to power AI data centres.

Using closed-loop cooling systems allows for both rainwater and recycled wastewater to be used multiple times. This method reduces freshwater use by 70 percent.

It is hard to rely fully on renewable energy sources, given the outdated electrical grid being used today. Experts believe the grid needs to be updated, as 70 percent of power lines are more than twenty-five years old.

With AI continually on the rise, there is no time to waste. As a society, we need to have conversations about potential solutions and implement them soon, so that we don’t leave future generations in further crisis.

• First published at Project Censored.

FacebookTwitterRedditEmail

Ella Mrofka is a sophomore journalism major at North Central College in Illinois. In summer 2025, she completed an internship with Project Censored. Read other articles by Ella.

Trump’s Big Tech Pledge Won’t Do: Advocates Make Case for Nationwide Moratorium on Data Centers

“Nothing short of a halt to the data center rollout will suffice... to ensure that people and the environment are fully protected.”

An aerial view shows cooling vent fans on the roof next to generators on the lower level of a Digital Realty data center in Ashburn, Virginia on November 12, 2025.
(Photo by Andrew Caballero-Reynolds/AFP via Getty Images)

Brad Reed
Mar 06, 2026
COMMON DREAMS

Several Big Tech CEOs met with President Donald Trump on Wednesday and pledged to fund their own energy infrastructure needed to power their artificial intelligence data centers that have caused US utility bills to spike over the last year.

That same day, Food and Water Watch slammed the pledge as “wholly inadequate” and released what it described as a “first-of-its-kind report” outlining the massive environmental and human costs imposed by the AI data center explosion.



Praising Denver and Other Cities for Leading the Way, Sanders Renews Call for National Data Center Moratorium



Trump’s AI Data Center ‘Ratepayer Protection Pledge’ Derided as Unenforceable, ‘Theatrical Stunt’

Among other things, the report states that data centers’ vast energy needs are throwing a “lifeline to the fossil fuel industry,” while undermining the many gains made from the revolution in clean power technology.

“AI expansion is largely fueled by dirty energy sources,” the report notes. “In the US, over 40% of energy for data centers comes from natural gas, 24% from solar and wind combined, 20% from nuclear, and 15% from coal.”

The report also pours cold water on Trump’s plan to have Big Tech build its own energy infrastructure to power its data centers.

“Power plants can’t come online fast enough to fuel this growth,” the report explains. “Data centers in New York state are seeking more than 9,000 megawatts (MW) of new demand—about 1.5 times the power consumption of every household in the state in 2024. Georgia Power predicts that energy sales will almost double by the early 2030s, largely driven by data centers. This steep demand increase can raise residential electricity costs—regardless of whether the new data centers pull from the grid or not.”

Electricity isn’t the only resource consumed in vast quantities by AI data centers, and the report also shines a light on the enormous amounts of water required to keep the facilities from overheating.

“The amount of water consumed by data centers more than tripled from 2014 to 2023,” the report explains. “By 2028, US data centers could use as many as 720 billion gallons of water each year just to cool AI servers. This is equal to over 1 million Olympic-size swimming pools—or enough water to meet the indoor needs of 18.5 million American households.”

Food and Water Watch says that the report’s findings point to only one solution: A moratorium on AI data center construction along the lines of what US Sen. Bernie Sanders (I-Vt.) proposed last year.

“The well-documented harms of AI data centers cannot be resolved with piecemeal regulations or vague promises from AI enthusiasts of a utopian future,” the report concludes. “Nothing short of a halt to the data center rollout will suffice until a comprehensive regulatory framework is developed to ensure that people and the environment are fully protected.”

Meghan Pazik, senior policy advocate with Public Citizen’s Climate Program, also criticized Trump’s AI data center pledge on Thursday and argued that the president’s plan “isn’t doing anything binding to cut energy bills.”

“Data centers increase residential energy bills by upwards of 250% and many communities are left in the dark on these projects from the start,” said Pazik. “Asking corporations to sign meaningless ‘agreements’ fits Trump’s tired pattern of seeking fake concessions from corporations that translate to zero action or relief.”

Trump’s AI data center pledge comes at a time when US voters are facing increasing economic pressure across multiple fronts. In addition to data centers’ impacts on utility bills, Americans are also facing increased costs from Trump’s global tariffs on imported products and a spike in gas prices caused by the president’s war against Iran.

While Trump has claimed to be prioritizing cutting costs with the data center pledge, he was dismissive of Americans’ concerns about paying more for gas this week, telling Reuters in an interview that “if [gas prices] rise, they rise.”

Biden signs executive order that seeks to grow artificial intelligence infrastructure in the U.S.

By — Sarah Parvini, 

Associated Press

Jan 14, 2025 

LOS ANGELES (AP) — President Joe Biden on Tuesday signed an ambitious executive order on artificial intelligence that seeks to ensure the infrastructure needed for advanced AI operations, such as large-scale data centers and new clean power facilities, can be built quickly and at scale in the United States.

The executive order directs federal agencies to accelerate large-scale AI infrastructure development at government sites, while imposing requirements and safeguards on the developers building on those locations. It also directs certain agencies to make federal sites available for AI data centers and new clean power facilities. Those agencies will help facilitate the infrastructure’s interconnection to the electric grid and help speed up the permitting process.

WATCH: How artificial intelligence impacted our lives in 2024 and what’s next

In a statement, Biden said AI will have “profound implications for national security and enormous potential to improve Americans’ lives if harnessed responsibly, from helping cure disease to keeping communities safe by mitigating the effects of climate change.”

“However, we cannot take our lead for granted,” the Democratic president said. “We will not let America be out-built when it comes to the technology that will define the future, nor should we sacrifice critical environmental standards and our shared efforts to protect clean air and clean water.”

Under the new rules, the departments of Defense and Energy will each identify at least three sites where the private sector can build AI data centers. The agencies will run “competitive solicitations” from private companies to build AI data centers on those federal sites, senior administration officials said.

Developers building on those sites will be required, among other things, to pay for the construction of those facilities and to bring sufficient clean power generation to match the full capacity needs of their data centers. Although the U.S. government will be leasing land to a company, that company would own the materials it creates there, officials said.

With less than a week before President-elect Donald Trump takes office, a big question is whether the incoming administration will keep or rescind the new order. Much of the order’s focus is on reducing the bottlenecks of getting energy-hungry data centers connected to new sources of electricity.

“It has to be a priority because otherwise you’re going to have blackouts, you’re going to have citizens or businesses being affected by this,” said computer scientist Sacha Luccioni, climate lead at the AI company Hugging Face. “Making it easier to facilitate interconnection of infrastructure to the electric grid is kind of a no brainer that would be useful for the next administration, no matter what their priorities are in terms of sustainability or climate.”

Biden said the efforts are designed to accelerate the clean energy transition in a way that is “responsible and respectful to local communities” and does not add costs to the average American. Developers selected to build on government sites will be required to pay all costs of building and operating AI infrastructure so that development does not raise electricity prices for consumers, the administration said.

The orders also direct construction of AI data centers on federal sites to be done with public labor agreements. Some of the sites are reserved for small and medium-sized AI companies, according to government officials.

Government agencies will also complete a study on the effects of all AI data centers on electricity prices, and the Energy Department will provide technical assistance to state public utility commissions regarding electricity tariff designs that can support connecting new large customers with clean energy.

As part of the order, the Interior Department will identify lands it manages that are suitable for clean energy development and can support data centers on government sites, administration officials said.

“The volumes of computing power, electricity needed to train and operate frontier models are increasing rapidly and set to surge even more,” said Tarun Chhabra, deputy assistant to the president and coordinator for technology and national security. “By around 2028, we expect that leading AI developers will be seeking to operate data centers with as much as five gigawatts of capacity for training AI models.”

Deploying AI systems at scale also requires a broader network of data centers across different parts of the country, he said.

“From a national security standpoint, it’s really critical to find a pathway for building the data centers and power infrastructure to support frontier AI operations here in the United States,” he said, adding that building data centers in the U.S. will prevent “adversaries from accessing these powerful systems to the detriment of our military and our national security.”

WATCH: How Russia is using artificial intelligence to interfere in elections

That type of investment will also prevent the U.S. from growing dependent on other countries to access AI tools, Chhabra said.

The executive order comes on the heels of the Biden administration’s proposed new restrictions on exports of artificial intelligence chips, an attempt to balance national security concerns about the technology with the economic interests of producers and other countries. That proposal raised concerns of chip industry executives as well as officials from the European Union over export restrictions that would affect 120 countries.

Missing from the order is how to manage the water consumption of AI data centers. There is a growing concern in states with multiple data centers over how to balance the economic development they bring with their impact on water resources as they use vast amounts of water for cooling, said J. Alan Roberson, executive director of the Association of State Drinking Water Administrators.

“Across the country, everyone is trying to get a better idea of the impact of data centers on water use now and in the future,” he said.

The executive order could have instructed federal agencies to collect information about how much water data centers use to help state and local officials making zoning decisions about whether to allow them, but it did not, he added.

AP writers Matt O’Brien and Jennifer McDermott in Providence, Rhode Island contributed to this report.

Left: FILE PHOTO: An AI (Artificial Intelligence) sign is seen at the World Artificial Intelligence Conference (WAIC) in Shanghai, China July 6, 2023. Photo by Aly Song/ Reuters/ File Photo

RelatedAs artificial intelligence rapidly advances, experts debate level of threat to humanity

By Paul Solman, Ryan Connelly Holmes
Hollywood’s video game performers head to picket line over artificial intelligence

By Sarah Parvini, Associated Press
Scientists harness power of artificial intelligence to battle wildfires

By Miles O'Brien
Go Deeperai
artificial intelligence
executive order
joe biden
By —

Sarah Parvini, Associated Press

Praising Denver and Other Cities for Leading the Way, Sanders Renews Call for National Data Center Moratorium

“We need serious public debate and democratic oversight over this enormously consequential issue,” said the senator. “The time for action is now.”


Prime Data Center is seen in Vernon, California on December 25, 2025.
(Photo by Myung J. Chun/Los Angeles Times via Getty Images)

Julia Conley
Feb 24, 2026
COMMON DREAMS


Although Denver Mayor Mike Johnston is a vocal supporter of artificial intelligence and has pushed to adopt AI-driven products to power the city’s infrastructure, he joined City Council members on Monday in announcing a moratorium on the construction of massive AI data centers—the latest sign, said US Sen. Bernie Sanders, that the push to stop corporations from building the energy-guzzling, pollution-causing facilities is not “radical, fringe, and Luddite” as some claim.

Johnston, a Democrat, and other local officials across the country who are pushing to block the construction of data centers “are right,” said Sanders (I-Vt.). “Data centers will have a profound impact on land and water use, and will drive up electricity costs.”



‘What a Surprise’: Sanders Undeterred by Bezos-Owned Washington Post’s Dismissal of AI Data Center Pause

As grassroots community groups and experts have warned, AI will also “likely have a catastrophic impact on the lives of working-class Americans, eliminating tens of millions of blue- and white-collar jobs in every sector of our economy,” said the senator, who proposed a nationwide moratorium on AI data centers in December.

He renewed that call after Johnston and the Denver City Council announced the city would halt any plans for new data centers for at least several months and would require projects that are already permitted or under construction to follow new guidelines once they’re finalized by local officials.

“We need a federal moratorium on AI data centers,” said Sanders.




Johnston said in a statement that he believes “data centers power the technology we depend upon and strengthen our economy,” but stressed that “as this industry evolves, so must our policies.”

“This pause allows us to put clear and consistent guardrails in place while protecting our most precious resources and preserving our quality of life,” said the mayor.

The city plans to review regulations for data centers that would target “responsible land, energy, and water use as well as zoning and affordability for ratepayers.”

Soaring electricity bills across the country have been linked to the build-out of data centers, which have cropped up as President Donald Trump has pushed to preempt state and local regulations on AI. As CNBC reported last year, residential utility bills rose 6% in August nationwide, but much higher price hikes were reported in states with high concentrations of data centers, like Virginia (13%) and Illinois (16%).

Sanders’ office issued a report last October showing that AI, automation, and robotics could replace nearly 100 million jobs over the next decade, including 40% of registered nurses, 47% of truck drivers, 64% of accountants, and 89% of fast food workers.

And a study published in Nature Sustainability last year found that data centers could consume as much water as 10 million Americans and emit as much carbon dioxide as 10 million cars.

At a forum last week at Stanford University, Sanders joined Rep. Ro Khanna (D-Calif.) in warning that the expansion of AI data centers is meant to increase the wealth of billionaire tech moguls with no regard for how working Americans are affected.

“The question that we should be asking day after day… is who is pushing this revolution, who benefits from it, and who gets hurt?” Sanderss said.

In Denver, the moratorium was announced ahead of a planned community meeting scheduled for Tuesday evening at Geotech Environmental, where neighbors are planning to speak out against the 170,000-square foot DE3 data center being built in the Globeville-Elyria-Swansea (GES) area by the Denver-based company CoreSite.

The burden that will be placed on locals if the project is completed “is not accidental,” reads a petition by the local grassroots community organization GES Coalition. “It is the outcome of colonial dispossession and extraction, then decades of zoning, redlining, highway construction, and industrial siting that concentrated pollution next to working-class homes alongside the legacy of the Vasquez Boulevard/I-70 Superfund site, a 4.5-square-mile smelting contamination footprint affecting multiple neighborhoods.”

Meanwhile, state legislators have introduced at least two bills regarding AI data center development. One, House Bill 1030, would offer sales and use tax exemptions for data center builders—and would slash state general fund revenue while also triggering a $106 million reduction in tax credits for low-income households.

Another, Senate Bill 102, would require data centers to use renewable energy sources and ensure their energy use does not raise rates for consumers.

Grassroots efforts to block the construction of data centers have taken off in places including Saline, Michigan; Port Washington, Wisconsin; and Tucson, Arizona, where community members successfully blocked plans for a new center owned by Amazon.

State lawmakers in Maine, South Dakota, and Oklahoma are also considering moratoriums or limits on new data centers.

“We need serious public debate and democratic oversight over this enormously consequential issue,” said Sanders. “The time for action is now.”

JPMorgan: Global AI Data Centers Will Consume Huge Amounts of Fresh Water

By ZeroHedge - May 23, 2024

While energy consumption of data centers steal the headlines, the water-intensive nature of their operations are overlooked.

Bluefield research: water consumption by global data centers (including on-site cooling and off-site power generation) has grown 6% annually from 2017 to 2022.

Immense water demand from data centers in areas where water resources are scarce could spark "increased competition can strain water availability, even causing data center closures."



Wall Street banks are in a frenzy over "The Next AI Trade," piling into the 'Powering up America' investment themes, whether that's power grid companies, commodities, such as copper, gold, silver, and uranium, and artificial intelligence chipmakers, to accommodate the explosion of generative artificial intelligence data centers anticipated nationwide through the end of the decade and beyond.

JPMorgan's Asia Pacific Equity Research desk is the latest bank to jump on AI trade in a note titled "Deep Dive into Power, Cooling, Electric Grid and ESG implications."



Focusing on AI data center power consumption is too repetitive at this point, considering we've laid it all out on a silver platter for premium ZH subs in the "The Next AI Trade" and "The Next AI Trade Just Hit An All-Time High."

As well as this real-world example... Goldman Finds Commercial Power Demand In Virginia Explodes Higher As 'Next AI Trade' Soars

Even Blackstone Chief Executive Officer Steve Schwarzman and BlackRock Chairman and Chief Executive Larry Fink have jumped onto the power grid and AI investment theme as there is plenty of upside in the years ahead - unless AI demand doesn't shit the bed.

Back to JPM's note, authored by analyst William Yang and his team, which near the end explained, "While data centers have been scrutinized for heavy electricity use, the water intensive nature of their operations has been comparatively overlooked."

Citing data from Bluefield Research, Yang said total water consumption by global data centers (including on-site cooling and off-site power generation) has grown 6% annually from 2017 to 2022. He said by 2030, water consumption could jump to 450 million gallons per day. To put this in perspective, that's 681 Olympic-sized pools of fresh water that will be needed each day to cool global data centers in about 4.5 years.



"By 2027, the same authors suggest that global AI demand may be accountable for 4.2 – 6.6 billion cubic meters of water withdrawal, more than the total annual water withdrawal of half of the United Kingdom when taking account of the combined scope 1 and scope 2 operational water withdrawal," Yang pointed out.

He said the immense water demand from data centers in areas where water resources are scarce could spark "increased competition can strain water availability, even causing data center closures."

Here are the various ways to cool data centers via water:



Much of the water usage at data centers is "because millions of gallons of water each day are evaporated in cooling systems designed to off-load server heat," the analysts said.

We'd love to know where the critics of crypto miners are now, as AI data centers are set to consume massive amounts of power and water.

Are any NGOs or Greta going to protest AI data centers? We doubt.

By Zerohedge.com

Consumers Are Footing the Bill for AI’s Insatiable Appetite for Energy

By Haley Zaremba - Jul 30, 2025

• The rapid growth of data centers, particularly due to AI, is significantly increasing energy demand and jeopardizing clean energy initiatives by extending the life of fossil fuel plants and promoting new ones.
• 
  
• The issue is compounded by "phantom data centers," which inflate projected energy demand and give utilities leverage to expand fossil fuel infrastructure.
• 
  
• This surge in energy demand and the resulting infrastructure projects are projected to lead to higher energy bills for consumers, especially in the Southeast United States.
• 
  

As data centers place more and more demand on global power grids, policy and economic priorities are shifting from creating more clean energy to creating more energy, period. Projected clean energy additions are simply not enough to meet the runaway demand of the global tech sector, meaning that climate goals could be at risk. 

The proliferation of artificial intelligence is causing massive increases in energy demand from data centers, and the areas that host them are struggling to keep up. A 2024 study from scientists at Cornell University found that generative AI systems like ChatGPT use up to 33 times more energy than computers running task-specific software. As a result, it is estimated that each AI-powered internet query consumes about ten times more energy than traditional internet searches. But these numbers are just our best guess – we don’t really know how much energy AI is sucking up, because the companies who are piloting AI platforms aren’t sharing those numbers. 

But we know that the overall picture is pretty grim. Last year, Google stated that the company’s carbon emissions had skyrocketed by a whopping 48 percent over the last five years. “AI-powered services involve considerably more computer power - and so electricity - than standard online activity, prompting a series of warnings about the technology's environmental impact,” the BBC reported last summer. While Google hasn’t publicly revised its goal of becoming carbon neutral by 2030, the tech firm has admitted that "as we further integrate AI into our products, reducing emissions may be challenging." 

Already, the uptick in energy demand from data centers is causing new plans for gas- and coal-powered plants as well as extending the life of existing fossil fuel operations across the United States. Utility Drive reports that “at least 17 fossil fuel generators originally scheduled for closure [are] now delaying retirement” due to data center demand, and that “utilities in Virginia, Georgia, North Carolina and South Carolina have proposed building 20,000 MW of new gas power plants by 2040” for the same reasons. 

The issue is particularly acute in the Southeast. Major utilities in Virginia, North Carolina, South Carolina and Georgia project that they will collectively add 32,600 MW of electrical load over the next 15 years. The Institute for Energy Economics and Financial Analysis reports that in Virginia, South Carolina and Georgia, “data centers are responsible for 65% to more than 85% of projected load growth.”

However, it could be the case that this projected demand growth is overblown, and that states will add extra gas power capacity – and therefore extra greenhouse gas emissions – unnecessarily. Because the competition for energy sources is so fierce between data centers, the project managers of new centers are likely to reach out to many different power providers at once with speculative connection requests, creating redundancies and a compounding issue of “phantom data centers.” This inflates demand and makes accurate projecting extremely difficult. 

A study published last year by Lawerence Berkley National Lab calculated exactly how big the phantom data center issue might be, and they found that projected energy demand could be as much as 255 terawatt-hours of energy higher than real energy demand. That’s enough energy to provide power to more than 24 million households.

However, it’s not in utilities’ interest to simplify interconnection processes and ferret out phantom data centers. In fact, the panic over rising energy needs from data centers is giving them great leverage to expand their businesses and push through huge fossil-fuel powered energy projects. Plus, while building new plants and extending the lives of old plants is costly, those costs will be borne by the ratepayers. 

Consumers across the U.S. – and especially in the data-center-laden Southeast – can expect their energy bills to rise in response. "We are witnessing a massive transfer of wealth from residential utility customers to large corporations—data centers and large utilities and their corporate parents, which profit from building additional energy infrastructure," Maryland People's Counsel David Lapp recently told Business Insider. "Utility regulation is failing to protect residential customers, contributing to an energy affordability crisis.”

By Haley Zaremba for Oilprice.com

Utilities, AI, and the Quiet Raid on Consumers

By Leonard Hyman & William Tilles - Jul 29, 2025

• Utilities are quietly signing electricity deals with AI firms that shift major capital costs to regular customers.
• 
  
• Utilities are placing AI-related infrastructure in their regulated rate base, socializing risks and costs while offering AI firms preferential, often secretive, pricing.
• 
  

• Policymakers would be wise to force AI firms to build their own power infrastructure, shielding consumers from excessive costs.

• 
  

Ok, we are cynical. The current electric utility policy environment is not exactly what you would call a level playing field, fairly balancing corporate and public interests. Quite the contrary. Right now, we have highly profitable (and politically influential) corporations facing underpowered civil servants in diminished regulatory agencies. State regulators are in a position to grant data centers and possibly other enormous users of electricity the opportunity to milk huge subsidies from unsuspecting consumers. How? By putting these vast new power-generating resources in the utility’s rate base, thereby socializing these enormous incremental costs, facilitated by pro-business politicians.

Harvard researcher (Daniel Oberhaus, “How AI Could Be Raising Your Energy Bill, ” Harvard Magazine, July/August 2025) cites more evidence that utilities are making AI power deals whose terms are not public that burden the rest of their customers. One utility plans to build several large power plants to serve a long term contract with an AI site, put the plants in their regulated rate base, but so far, no details on apportionment of costs have been released. Another large utility has been accused of offering a cut-rate deal to an AI firm with the full expectation that the rest of its customers would make up the profit differential. Stated simply, residential utility customers would be subsidizing corporate or AI  electricity usage. In addition, several AI centers announced deals in which they would take the output of existing deregulated stations. Those are perfectly above-board transactions, but not neutral to consumers who have to finance new power stations at current, relatively high prices to replace the output taken by AI. Just a guess, but we think that if AI adds 10% to system sales, it will add 30% to the fixed and capital costs of the utility, so if the AI firm gets a discount, who pays the higher costs? You guessed it.                                                    

Next, let’s consider a few wild thoughts.

For instance, might we be in an AI bubble, on par with the dot.com bubble and the power generator bubble, and way back, the bowling alley bubble? These economic or speculative bubbles burst because actual demand did not materialize to keep up with supply based on overly optimistic forecasts. As a result, entrepreneurs overbuilt or somebody came up with a better or cheaper product. “Of course”, you say, ”this time is different.” But someone always says that.

Let’s consider possibilities. First, the AI providers may have wildly overestimated electricity demand, which will leave a lot of underutilized AI data centers looking for ways to dodge those big power bills. Second, the Chinese will develop cheap, low powered AI related computer chips that make the US version look uneconomical. For example with electric vehicles, Tesla came first but Chinese auto makers have caught up fast and now offer superior products. As Andrew Carnegie supposedly and ungrammatically said, “Pioneering don’t pay.” Third, when quantum computing makes an earlier-than-expected appearance, will all those AI data centers become white elephants?

Mind you, we don’t object to building AI centers. Entrepreneurs should take chances and reap the rewards. We just object to making consumers subsidize this essentially speculative activity via their electric bills. Given the intricacies of the electric network, and cross subsidies in pricing, the divergence in cost between new and old plant, and the political pressure to give large corporations what they want via hidden extra charges in the customer's electric bill, we suggest that there is only one way to protect consumers from what AI will do to the grid. That is by requiring that AI and similar power guzzlers build their own generation and networks. Put all these new expensive generating assets behind the meter so to speak. That means build, not snatch existing stations that consumers will have to replace. That way, data centers will pay full freight and consumers will only have to deal with all the other costs increasingly burdening the utility network.

By Leonard Hyman and William Tilles for Oilprice.com

Fracking Industry Executives Are Salivating Over the AI Data Center Boom


Data centers are driving a surge in the construction of power generation facilities involving fracked gas.

By Derek Seidman ,

Truthout

January 3, 2026

Aerial view of a data center being constructed.Gerville / Getty Images

Support justice-driven, accurate and transparent news — make a quick donation to Truthout today!

“The demand for power and for AI is like nothing I’ve ever seen.”

These words were uttered during an October earnings call, not by a wide-eyed tech executive, but by Jeff Miller, the CEO of Halliburton, one of the world’s biggest oilfield services corporations.

Like droves of other companies tied to the fracked gas industry, Halliburton is pivoting toward servicing the data center boom with new loads of methane-emitting, gas-fired power generation to feed the artificial intelligence (AI) bubble being stoked by Silicon Valley billionaires and allied corporate elites.

Halliburton is merely one of many fossil fuel companies that are striking deals to power the insatiable electricity needs of data centers. More specifically, the natural gas industry — frackers, gasfield services, pipeline companies, power suppliers — are positioning themselves as bold saviors ready to step in and meet AI’s bottomless energy demands.

In doing this, fracked gas companies are fully backed by the Trump administration, with its ideological dedication to fossil fuels and its cozy relationship with billionaire oil and gas donors. For its part, Big Tech is going along, largely sidelining its purported commitments to renewable energy.

Data Centers Devour Electricity. Private Equity Is Buying Utilities to Cash In.
The seizure of public utilities for the sake of profit may lead to a disaster for consumers — and the planet. By Derek Seidman , Truthout November 11, 2025

As the global climate emergency intensifies, none of this bodes well. Despite being sold as a “clean” fossil fuel, fracked gas emits loads of methane, one of the most potent greenhouse gases, and local communities — often rural, low-income, and predominantly Black — bear the brunt of the combined nexus of data center behemoths and fossil fuel power generation being constructed in their backyards.

“The natural gas industry is directly aligning with the data center industry.”

“The natural gas industry is directly aligning with the data center industry,” Tyson Slocum, the director of Public Citizen’s Energy Program, told Truthout. “From a climate perspective, and from a local environmental perspective, data centers represent a significant impediment to action on climate change.”

“Icing on the Cake”

Natural gas currently provides over 40 percent of the electricity for data centers, making gas-fired power stations their largest power source, according to the International Energy Agency.

The massive demand for electricity from data centers is driving a surge in the construction of power generation infrastructure fueled by natural gas.

One McKinsey analyst recently noted that over 100 gigawatts (GW) of new gas-fired projects are being planned. “To put that number in perspective,” he said, “over the last five years, the U.S. added only about 35 GW of gas,” meaning this is “almost triple what it was.”

Utilities announced a slew of new gas-fired power projects in 2025 — a factor that helps explain private equity’s rush to acquire utilities. Utility powering of data centers is expected to skyrocket by 22 percent this year.

Industry leaders are swooning. “It’s been 40 or 50 years, or so, since we’ve seen demand grow the way it’s growing and is expected to grow,” said one executive of energy giant NRG.

“AI is obviously a big part,” he added.

The core driver of the fracked gas industry over the past decade has been the booming production of liquified natural gas (LNG) for export, which has vastly accelerated since the Obama administration. Today, the U.S. leads the world in natural gas production and exports.

While LNG exports are unrivaled as “the demand driver and the profit center for the domestic natural gas industry,” says Slocum, data centers “provide a significant additional profit cushion.”

“Data centers are far and away the largest variable that is increasing electricity demand,” he said. “They’re sort of the icing on the cake.”

“Help Nourish That Appetite”

With 3 million miles of gas pipelines tightly networked across the country, Slocum says the fracked gas industry has positioned itself as the prime supplier for powering data centers.

This is seen through a flurry of data center deals struck by corporations across the fracked gas supply chain — independent drillers, oil and gas majors, pipeline companies, and oilfield services companies.

EQT, a top U.S. natural producer based in Western Pennsylvania, the heart of the vast Marcellus Shale formation, struck a deal to supply two huge data center hubs in Appalachia — the Shippingport and Homer City projects — with a combined 1.5 billion cubic feet per day of gas supplies.

“Just to put this in perspective, that’s enough natural gas to power two of New York City,” gloated EQT CEO Toby Rice, who also added that Homer City and Shippingport are “just the first steps of multiple steps in multiple projects.”

EQT is hardly alone. Fracking giants like EOG and Antero Resources are striking deals to position themselves to supply data centers. Comstock Resources, owned by Dallas Cowboys owner Jerry Jones, is partnering with utility giant NextEra to “keep the lights on at a plethora of data centers” in Texas, says Natural Gas Intelligence.

Chevron, the second-largest U.S. oil and gas company, is getting in the game. “AI data centers require massive amounts of energy to function,” the company gushed in a February press release.

But have no fear, they promised. “Chevron is tapping into natural gas to help nourish that appetite,” the oil giant declared, announcing a new deal to build gas-fired power plants for data centers.

“Unleashing American Energy”


While the massive energy demands of artificial intelligence are pulling the fracking industry toward data centers, the Trump administration’s policies are also playing a critical role.

“National policy under Trump is prioritizing fossil fuels for data center development” and “particularly natural gas,” Slocum told Truthout.

“It’s explicit in Trump’s July executive order on artificial intelligence, where he defines the criteria that data centers need to meet to qualify for expedited approval, and it lists every energy source except wind and solar,” Slocum added.

That July 2025 executive order, titled “Accelerating Federal Permitting of Data Center Infrastructure,” clearly emphasizes the role of fossil fuels in powering data centers. It defines “Covered Components” — the “materials, products, and infrastructure that are required to build Data Center Projects” — using language that foregrounds “energy infrastructure” like “natural gas pipelines or laterals” and “natural gas turbines” and “coal power equipment” with no mention of words like “wind” or “solar.”

As Slocum has written, “a data center proposed to be powered by wind and solar will not qualify for expedited treatment, whereas a fossil fuel powered facility would,” adding that fossil fuel-powered data centers could also qualify for direct federal subsidies.

This codification of national policy around fossil fuels as the core supplier of data centers aligns with Donald Trump’s campaign promise to “Drill Baby Drill” and his adamantly pro-fossil fuel January 2025 executive order on “Unleashing American Energy.”

Since taking office in January 2025, the Trump administration has gone to war against renewable energy, freezing permits around solar and wind projects and denigrating windmills.


Trump and the Fossil Fuel CEOs



Trump’s ideological commitment to feeding data centers with fossil fuel-powered electricity is wedded to his cozy relationship with the fracking billionaires who helped bankroll his 2024 reelection.

Trump saw oil and gas barons as a key constituency during his 2024 reelection campaign, and industry billionaires like Harold Hamm and Kelcy Warren showered Trump with millions in donations.

As Truthout has previously noted, Trump filled out his cabinet with fossil fuel executives and boosters, including Chris Wright, the fracking-liquid-guzzling former CEO of Liberty Energy, who Trump made energy secretary.

Since then, the fossil fuel industry has benefited handsomely from Trump’s policies of environmental deregulation and tax subsidies — and a bonanza of new business tied to data centers.

Warren is the co-founder and executive chairman of pipeline giant Energy Transfer and a longtime Trump ally. Energy Transfer is awash in new business supplying data centers from Texas to Arizona and is partnering with AI giants like Oracle.

Hamm is the founder and former CEO of fracking giant Continental Resources, and is a major Trump donor who is also backing Trump’s White House ballroom. He may be Trump’s biggest fossil fuel industry ally.

“Together,” writes The New York Times, Trump and Hamm “have remade federal policy to benefit oil and gas companies, including Mr. Hamm’s Continental, and put off the transition to greener alternatives like solar power and batteries.”

The Hamm Institute for American Energy, a research and policy hub funded by Hamm, who also serves on its board, has supported natural gas for powering AI and hosted an April event on “powering AI” that featured key Trump cabinet members leading the administration’s energy policies.

Liberty Energy — again, formerly led by Trump Energy Secretary Chris Wright — is also striking deals to power data centers in Pennsylvania.

“The only thing that’ll prevent us from leading in AI is the failure to build this electric generating capacity that needs to happen,” Wright recently raved to the Council on Foreign Relations, adding that he was “using emergency powers to stop [the] closure of coal plants” and “expediting the permitting of building of new plants.”


Big Tech’s Conformity



For its part, Big Tech has quickly backtracked on its net zero commitments as it embraces gas-fired power generation for data centers.

For example, Entergy, a power generation behemoth that relies mostly on fossil fuels, is currently constructing three massive gas-fired plants to power a $10 billion Meta data center in Louisiana.

Less than a decade ago, Meta CEO Mark Zuckerberg waxed about addressing the climate crisis. Fast forward to today, and headlines like “Meta goes all in on gas to power a mega data center” are being published, while Zuckerberg and Meta are being lambasted by lawmakers, energy analysts, and community groups.

This fits a larger pattern, which Truthout has reported on, of Silicon Valley CEOs backtracking on purported climate aims and criticism of Trump to remain the president’s good graces and profit off his policies.

“Trump demands conformity, and Big Tech is providing that conformity by turning its back on its traditional commitments to increase and rely upon renewable energy,” said Slocum.

Of course, all this will exacerbate the climate crisis. The production, transport, and burning of natural gas releases huge amounts of methane, which is 86 times more powerful than carbon dioxide at trapping heat in Earth’s atmosphere.


Under Trump, data centers are also giving new life to dying and dirty coal plants.

Under Trump, data centers are also giving new life to dying and dirty coal plants, and many data centers are backed up by super-polluting diesel generators that some fear could come to be used more frequently.

As journalist Adam Mahoney has reported, in states like South Carolina and Texas, Black households disproportionately bear the brunt of the data center boom and AI’s growing fossil fuel emissions trail, especially with on-site power generation.

Many communities are resisting both data centers and their dirty emissions — from Memphis, Tennessee, to Bessemer, Alabama, to Santa Teresa, Arizona.

To be sure, Big Tech firms are also partnering on some projects with renewable energy sources to service data centers. While U.S. data centers are overwhelmingly powered by fossil fuel sources, the International Energy Agency notes that the proportion of solar, wind, and other renewables involved will increase going into the 2030s and beyond.

But this raises another question: Do we really want the renewable energy transition to be dominated by — and wasted on meeting — the profit-centered priorities of Big Tech and Wall Street firms and their fixation with the unproven, supposed wonders of AI


This article is licensed under Creative Commons (CC BY-NC-ND 4.0), and you are free to share and republish under the terms of the license.

Derek Seidman is a writer, researcher and historian living in Buffalo, New York. He is a regular contributor for Truthout and a contributing writer for LittleSis.