Everybody Hates Data Centers
Anarchists, union activists, Indigenous organizers, and disgruntled Trumpists find themselves side by side in the fight.
Between 4,000 and 5,000 data centers are actively humming in the U.S. right now, draining energy and, in the case of some of the hyperscale ones, consuming as much as 5 million gallons of water per day. Even this does not satisfy the demand cultivated by the tech industry, however: At least 3,000 more data centers are under construction or planned, prompting a diverse grassroots mobilization against their construction.
Indigenous people are resisting continued attempts to exploit their land, air, and water. Rural white folks, some of whom voted for Donald Trump, are now going door to door, outraged about rising electricity costs and water shortages.
Opposition to data centers has put labor unions in motion, with the Graduate Employees’ Organization at the University of Michigan using the slogan “AI is not inevitable,” seeing artificial intelligence as a dystopian force made possible through data centers. These efforts represent the front line in the struggle against attempts by high-tech billionaires to create a dismal world in which they hold all the power and the rest of us serve their interests.
“There’s this big techno-feudal battle happening right now. We’re watching an Empire crumble, right?” Krystal Two Bulls, an Oglala Lakota/Northern Cheyenne anti-data center organizer with the activist group Honor the Earth, told me. “Technology is the last frontier, and whoever has the most advanced generative AI has the power at this moment.”
Organizers from Tennessee, Wisconsin, and Arizona share strategies for resisting data centers in other communities. By Derek Seidman , Truthout April 19, 2026
Native Americans are some of those leading the organizing against data centers, from Virginia and upstate New York, though Montana, the Dakotas, Arizona, and Oregon. According to Honor the Earth, a national Indigenous sovereignty organization, there are currently at least 106 data centers being proposed on or near Native land.
Data centers are the material constructs enabling the perpetuation of large language models (LLMs), commonly referred to as a form of “intelligence.” LLMs aggregate knowledge and then generate information based upon processing and sorting huge inputs of words, but despite advances in their ability to differentiate between believable scenarios and nonsense, in most ways the algorithm still operates linguistically rather than conceptually. It’s not smart, it’s just massive.
Developed within the framework of capitalism, artificial intelligence threatens human civilization in several ways. In addition to its use by corporations to rapidly replace human labor — both blue-collar assembly line and white-collar administrative positions — there are genuine fears by many involved in its development that AI might someday not only take human jobs, but also lead to the mass destabilization of human societies, increasingly repressive surveillance regimes, or in the most alarmist imaginings, even an existential risk to human survival. Future killing machines like something from the Terminator films aside, AI is already being used to choose targets — and this process has resulted in the murder of Iranian school children and people in Gaza.
As Krystal Two Bulls argues:
Generative AI is being used to surveil citizens, violating our right to privacy, and then that is being turned around and used for military purposes. We can see the connection directly to violence. As Native peoples we honor the Earth, which is why it is so important to resist the construction of these data centers and the entirety of the AI infrastructure.
Looking at the global context in which AI is being developed, these construction projects drive us further from an adequate response to the climate crisis. We have known for several decades that continuing to burn fossil fuels will destroy the unique climate that supports human flourishing on Earth. It is imperative that we stop extracting and burning them, yet with Trump’s return to power and Republicans in both chambers of Congress, what little progress was made in the U.S. is being reversed. This is happening at the very moment we are approaching various climate tipping points, accelerating climate collapse.
“It’s about utilities and humming noises and water usage, definitely, but even more so it is about the way Big Tech has continually extracted with no punishment.”
For the last 20 years, the Tonawanda Seneca Nation in western New York has resisted industrial development at the site of the Science, Technology and Advanced Manufacturing Park (STAMP). The 1,250-acre sprawling site is located next to Big Woods, a natural area of cultural and practical significance for the Seneca Nation, as well as a federal and state-recognized wildlife refuge. The nearby white settlement effort to site a data center here led the Seneca Nation’s chiefs’ council to ask Grandell “Bird” Hallett Logan — the community language resource coordinator at the reservation — to be the spokesperson for the Seneca Nation. Logan told me that the organizing began with Native people from his community, but in time, a group developed called the Allies of the Tonawanda Seneca Nation, bringing together people from the Seneca Nation with those from the non-Native community elsewhere in the state.
According to the Sierra Club, a nondisclosure agreement signed by the Genesee County Development Center prevents the community from even knowing which tech giant the data center would serve. Before work can begin on the site, the planning board of the local town of Alabama, New York, has to approve the project, and it must receive a state environmental review, opening up several avenues to stop it. In addition, the data center’s opponents are also considering a lawsuit, which could delay or prevent its construction entirely.
Logan explained how they are organizing: “Much of our messaging is about how our customs and usage are at risk due to environmental harm. We worry that this noise pollution from a data center would cause the animals nearby to flee.” Discussing the support that the Allies of the Tonawanda Seneca Nation have received from the surrounding non-Native community, Logan said, “Some people are supportive of us because they see how their county, their state, is acting in ways that continue to harm our customs, and they’d prefer that their government not take part in such actions.”
Los Alamos National Lab Data Center
Los Alamos National Laboratory, founded in 1943 for the Manhattan Project and instrumental in the development of the first atomic bombs, is planning to build a $1.25 billion data center on acreage that the University of Michigan, which is based in Ann Arbor, purchased nearby in Ypsilanti Township. Workers at the university, some of whom previously did Palestine solidarity work and organized around how AI is affecting teaching, have now taken up the struggle to stop the data center.
Nathan Kim, a member of the University of Michigan Graduate Employees’ Union who previously helped run workshops like an AI “Workers’ Inquiry,” told me that he found the opposition against the data center in the surrounding community inspiring, bringing campus union workers into coalition with people living near the campus. This “shifted how I thought about politics — it wasn’t the case that everything was hopeless, but instead that people would continue to fight back, and always would,” Kim said.
People across the U.S. are in motion against data center construction. As Kim observes, “The reality of this issue being mobilizing in new and interesting ways meant that many people were angry, but few people saw themselves as leaders from the outset.”
This is common when social or environmental issues impact local communities. One can think of the people affected by contaminated drinking water in Flint, Michigan, or those impacted by toxic waste in the Love Canal neighborhood of Niagara Falls, New York. People experience anger and anxiety, often followed by frustration over their interactions with local officials. With regard to data centers, Kim said: “People seem to have a built-up resentment and hatred of Big Tech that data centers have become an outlet for. It’s about utilities and humming noises and water usage, definitely, but even more so it is about the way Big Tech has continually extracted with no punishment.”
We can use the space of the emerging, imperfect coalitions against the data centers to create openings to redirect anger away from scapegoats and toward the systems of power.
What is noteworthy about how people are organizing against data centers in this part of Michigan is the utilization of the type of horizontal structures that came out of the anti-authoritarian organizing of the anti-globalization movement at the turn of the century and the subsequent Occupy movement.
As Samantha Stewart, who lives in Ypsilanti Township and is involved in the struggle against the data center, says about the campaign, “It’s designed with a spokescouncil style, so there are a million working groups (literally I know of 17 that meet regularly) all doing what makes sense to them to stop the data center.”
In a traditional spokescouncil model, smaller affinity groups and working groups send delegates to a spokescouncil to make decisions. “There are big monthly meetings where everyone comes together, they always have food and child care. No one is in charge; each working group does what it wants to stop the data center,” Stewart explained, adding that what holds this free-wheeling assembly together are “some small working agreements, just that we don’t condemn each other publicly … and that we don’t collaborate with police in the prosecution of each other.”
Organizing against data centers in Michigan has brought together people with many types of backgrounds. “Ypsilanti is more working-class with a very active anarchist movement … But the fight against data centers in Michigan as a whole is very diverse,” Kim points out. For example, he said he went to an organizing “meetup once in Augusta housed in a barn that was a U of M football watch party room and had ‘Faith, Family, Freedom’ dangling over the door.” Kim says, in his organizing work, that “encounters are somewhat frequent” with Trump supporters. But, as Stewart points out, “The campaign has been so welcoming!”
Stewart added:
People hate the data centers for all sorts of reasons! The serious negative health consequences are really concerning to a lot of people. People are worried about increased bills and the destruction of our beloved park. Lots of people are against escalating war, and really worry about [associating with labs known for] developing nuclear weapons, or that we will be a military target based on this facility.
Stewart said that when she has “gone out knocking on doors, there have been a ton of folks with Trump yard signs who absolutely hate the data center. The Republicans struggle because they mostly don’t have organizing experience and aren’t sure how to get things done together.”
Trying to stop the data center has brought these disparate groups together. “We are a multiracial and very trans organizing group, so we don’t gel super well with them, but they do their own stuff and we do coordinate,” Stewart said.
Addressing the diverse nature of data center opposition in Michigan, Kim said: “I see the left’s job today to understand the kernel of truth at the heart of each Trump voter’s beliefs,” and then to “identify how that can be met with a transformative socialist vision where we take care of each other, can have our basic necessities met, and stand for peace.”
Trump Coalition Fissures
The tech and fossil fuel industries are two of the biggest sections of the ruling class backing Trump and the current transformation of the U.S. into a more openly authoritarian state. But fissures are starting to open up within sections of the grassroots MAGA movement. First, Steve Bannon and other MAGA activists expressed their racist, anti-immigrant opposition to Big Tech’s desire to grant visas to highly skilled workers. MAGA members have also expressed dissent for the U.S.-Israeli war against Iran, as well as the broader U.S. policy of providing Israel with financial and military support, including throughout its genocide of Palestinians and its war on Lebanon. Now many people who voted for Trump several times are opposing the construction in their communities of the tech industry’s sacred data centers.
“I want to build trust and bravery across a broad group of people so that we are ready for the next thing we need to fight.”
As Indigenous people, union organizers, and those on the left come into contact with these disgruntled Trumpists, the opportunity exists to develop mutual understandings. Perhaps these encounters will widen the worldview of those who follow Trump and who are hostile to other working people, including those who have recently arrived in the U.S. looking for a better life.
The process of widening a person’s worldview entails hard discussions. If we approach these with a strong understanding of the material grievances that draw people toward Trumpism, we can use the space of the emerging, imperfect coalitions against the data centers to create openings to redirect anger away from scapegoats and toward the systems of power driving ecological destruction, exploitation, and war. This presents an opportunity to cultivate a unifying understanding of the role of class oppression and how to resist capitalist exploitation.
The struggle against data centers is but one front in the fight against rising authoritarianism and ecological collapse. It is an issue that resonates across political divides, opening opportunities, and potentially sparking a larger movement. Stewart has the necessary long-range perspective, saying: “I want to build trust and bravery across a broad group of people so that we are ready for the next thing we need to fight.”
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.
Paul Messersmith-Glavin
Paul Messersmith-Glavin is a longtime organizer whose work has appeared in Waging Nonviolence, New Politics, Upping the Anti, and the Oregonian, as well as Dan Berger and Emily K. Hobson’s Remaking Radicalism: A Grassroots Documentary Reader of the United States, 1973-2001. He works as a mentor with the Institute for Social Ecology and is part of the Perspectives journal collective whose book, Visions and Interventions: 30 Years of Perspectives on Anarchist Theory (AK Press) comes out this July.
Be less polite: How to cut your AI impact as UN report reveals data centre energy use rivals nations
"That extra ‘please’ you put there can make a huge difference,” says one of the report's authors.
The environmental footprint of data centres already rivals some of the world's largest countries, according to a United Nations University report released on 3 June.
Their water use, energy use and pollution is predicted to double in just four years as use of artificial intelligence grows.
Much of the growth of data centres is being driven by AI. About 20 per cent of data centres’ energy is currently due to AI, but that should grow to 40 per cent by 2030, the report said.
AI users can reduce the climate impact of their queries by less polite and more concise in their queries, one of the report's authors advises.
The majority of people – 70 per cent – are polite to AI when interacting with it, according to a survey carried out by British publisher Future in 2024. Of the respondents, 55 per cent said they do this because "it's just the nice thing to do", while 12 per cent said it was because "when the robot uprising happens, I don't want to come first".
Electricity use equal to that of Argentina
Last year, global data centres used 448 trillion watt-hours of electricity, more than all but 10 countries of the world, said the report. That electricity use produced about 189 million tonnes of carbon dioxide, about the same amount as Argentina, and producing that much energy consumed about 4.5 trillion litres of water, according to the report on the environmental consequences of AI's energy use.
By 2030, data centres will account for nearly three per cent of the world's projected electricity use, with 935 trillion watt-hours. If data centres were a country, the country would be projected to rank sixth-highest in power use in 2030. That would produce nearly 399 million tonnes of carbon dioxide, the report said. The study focused on energy use and didn’t examine the massive amount of water used to cool data centres.
“If you look at these numbers, we're seeing scales comparable to nations,” says study co-author Kaveh Madani, a water scientist and director of the United Nations University Institute for Water, Environment and Health in Canada. “The demand is enormous.”
First global look at ecological impact of data centres
The report is significant because of the credibility and authority of the UN, not just because of any one set of eye-popping numbers, says Fengqi You, a Cornell University energy engineering professor who directs the college’s AI sustainability issues.
“Its value is that a UN institution is putting carbon, water, land, life-cycle impacts and environmental justice into one frame” for an issue that is often shrouded in secrecy and partial disclosures, says You, who was not part of the report.
“The general public should be concerned, but not panicked,” he adds.
Jean Su, director of the Energy Justice Program at the Center for Biological Diversity, said the report is important because it is the first UN, or even global, report “that shines a light on the environmental harms of AI”.
National Artificial Intelligence Association President Caleb Max emphasises how his industry is becoming more efficient and how it benefits the public: “AI is rapidly becoming part of our everyday lives and adding benefits that improve safety, [help people] live longer, work more efficiently, enhance food production, and reduce poverty. The evidence is growing daily that the energy return on investment of AI development is transformative for our world and therefore more than worth it.”
Josh Levi, president the Data Center Coalition, says the industry takes its environmental impact seriously.
“We remain committed to working with policymakers, local communities, and industry partners to ensure that as data centres grow, they do so responsibly, transparently, and in ways that reflect the best available practices,” he said in a statement.
The report came just after Californian city Monterey Park became the first in the US to vote for a permanent ban on data centres on Tuesday (2 June).
How much energy your query uses and how to trim it
Madani, also the winner of the most recent of the Stockholm Water Prize, says the numbers show the environmental cost of AI, which may seem cleaner at first glance than other mechanical devices, such as cars and furnaces, that have visible pollution.
"AI is not just a virtual thing. We’re talking about something that has physics, something that has real impacts. There is infrastructure there. There is energy that is being used," Madani says. "A lot of hardware is behind all these operations that to us seem very, very clean because we don’t see smoke out of our devices. On our cellphone, there is no visible smoke or out of our computer or something. But somewhere else someone is suffering."
People can reduce AI’s massive energy appetite by being less polite and more concise in their queries, Madani says. The report found that cutting word use in requests by 30 per cent can reduce energy used by AI by 25 per cent. That would save about the same amount of electricity as what about 700,000 people in Africa use in a year, the report said.
“If you’re too polite, then that extra ‘please’ you put there can make a huge difference,” Madani says. “You’ve got to be very precise and be short.”
A typical ChatGPT-style query is about 200 times more energy-intensive than the type of basic text classification used in an email spam filter, for example. AI-generated images or video require much more energy.
And the more complicated the AI, the more energy it takes to train or learn. The report said GPT-3 used about 1.3 billion watt-hours to train, but the next version used 50 to 70 billion watt-hours.
But it's not training that really feasts on power, says study co-author Miriam Aczel, a United Nations University environmental policy researcher. About 90 per cent of the power use of AI comes from operational requests, she says. GPT alone accounts for 2.5 billion prompts a day, she says.
Efficiency still means more power use
Even though tech advocates can argue that their machines are becoming more efficient, there's a common paradox that finds when things get more efficient, they are used more often and total energy use soars even if individual uses are more efficient, Madani says.
While some companies tout the use of renewable energy for data centres, Madani says that means the supply of clean electricity is depleted and thus dirtier energy is used elsewhere.
One of the problems in conducting this study is that many companies and places are not transparent about what data centres and AI are consuming or even where and how big they are, Aczel and Madani say.
“We cannot manage what companies do not disclose,” Cornell's You says.
File photo of a data center facility near Boardman, Oregon. (Photo by Jordan Gale/Oregon Capital Chronicle)
Oregon Capital Chronicle
By Alex Baumhardt
(Oregon Capital Chronicle) — Electric utilities in Washington and Oregon are turning to gas to meet rapid and growing energy demand from data centers, according to recent reports.
Two analyses from the Hood River-based conservation organization Columbia Riverkeeperand the Seattle-based think tank Sightline Institute show that a growing number of Northwest utility companies are spending on new gas-powered energy infrastructure or buying gas-powered energy from other states to power new demand from data centers.
In some counties, public utility districts are permitting gas-powered generators to provide data centers with backup energy, rather than waiting for them to get more power from the grid, and some data center companies are hooking up their own on-site gas generators. For their part, data center companies said they are investing in communities when they show up and working with utilities to find the cleanest energy possible.
The effect is that both Oregon and Washington are at risk of missing established emission reduction targets meant to help curb the impacts of global warming, researchers found. And a growing number of utilities are using booming data center demand to justify skirting climate rules in both states that mostly ban the build-out of new gas infrastructure, citing the need for regional energy reliability.
“In the absence of enough renewable energy supply, we’re seeing utilities turn more to gas in this situation,” said Audrey Leonard, staff attorney at Columbia Riverkeeper and one of the authors of the group’s report. “That is new, because up until the last few years we were making progress towards our clean energy targets in Washington and Oregon. We were really diversifying our clean energy mix, and it was always going to be a challenge — we definitely had work to do — but the way I characterize this is that data centers are turning that challenge into a crisis.”
Representatives for some of the biggest data center owners in the region said they are doing their part to connect to existing clean energy sources and invest in new renewable projects in both states, and that their presence is a boon for communities, not a burden.
“Amazon is committed to being a responsible neighbor in Oregon, where we’ve invested more than $60 billion since 2010 through infrastructure and jobs,” said Margaret Callahan, an Amazon spokesperson. Callahan said the company’s data centers in the region are 10% more energy efficient than the industry average and that the company has invested in massive wind and solar projects across Oregon.
Morgan Babinec, a Microsoft spokesperson, emailed the Capital Chronicle a link to a company presentation espousing the company’s data center benefits in Washington. It notes that Microsoft has “committed to achieving 100% renewable energy coverage globally by 2025,” and that “our data centers in Washington are transitioning the backup generators to use a renewable biofuel that reduces net carbon emissions.”
Missing targets
Under a 2020 executive order from former Oregon Gov. Kate Brown, state leaders and agencies have passed laws and policies meant to reduce Oregon’s greenhouse gas emissions to 45% below 1990 levels by 2035 and 80% below 1990 levels by 2050.
Until recently, the state was on track to meet those targets, according to 2023 Oregon Department of Energy modeling. But in October, department officials reported that given the massive rise in energy demand for data centers, along with Trump administration rollbacks of federal clean energy policies and fuel economy standards for cars, the state wouldn’t hit its 2035 goals until 2037.
To meet both Oregon and Washington’s climate goals — which include electrifying almost the entire transportation sector in both states by 2050 — the states also need to replace at least 65 million megawatt hours of existing coal and gas power generation with power generated by renewable sources such as wind and solar, the Columbia Riverkeeper researchers wrote.
Instead, federal energy officials have used data centers as a justification for keeping Washington’s largest coal-burning power plant running late last year, despite state laws requiring it to be shut down. And Puget Sound Energy, Washington’s largest utility, has contracted for six new gas turbines to be built at a new gas power plant at an undisclosed site in Washington, according to Sightline Institute.
Natural gas is almost entirely methane, a potent greenhouse gas that, when burned, emits carbon dioxide. Carbon dioxide and methane are the main heat-trapping gases causing global warming.
Gas-powered workarounds to clean energy mandates
Of the more than 100 electric utilities in Oregon and Washington, two private investor-owned utilities — PacifiCorp and Portland General Electric — and four public utility districts and cooperatives have absorbed nearly all the region’s new data center loads in the past decade, according to a 2025 Sightline Institute analysis.
Five of those utilities rely today on buying far more “unspecified power” from wholesale electricity markets to meet customer demand than they did a decade ago. Unspecified power is almost always natural gas or coal, and the Washington State Department of Ecology estimates emissions from unspecified sources are roughly equal to natural gas emissions.
In Oregon, unspecified power purchases have driven up what were relatively low emissions from consumer-owned utilities since 2019, according to Oregon Department of Environmental Quality data. Most of the increase is from the Umatilla Electric Cooperative buying more unspecified power to meet demand from Amazon data centers in recent years, the department found.
Callahan, the Amazon representative, said the company recently made a deal with the Umatilla Cooperative to choose the energy supply used for its data centers rather than leaving it up to the utility to buy the cheapest option on the wholesale market.
Still, the company seeks more backup and reliable power when it urgently needs it, and the Oregon environmental quality department recently fined an Amazon data center in Hermiston for violating its air quality permit by running an emergency diesel generator for 50 hours more than allowed.
Although Oregon and Washington have mostly prohibited building new gas power plantsor importing more gas-powered electricity, both offer some workarounds. In each state, utilities can seek exemptions if they claim overall reliability is in jeopardy or the costs are too high to procure clean energy. And in Oregon, the climate protection rules only apply to investor-owned utilities.
Utility commissioners in central Washington’s Grant County, where a public utility district provides power for at least 27 data centers, recently cited a lack of transmission capacity to approve plans for the company VoltaGrid to build a new 12-megawatt methane gas power plant to supply a hyperscale data center campus owned by the multinational tech company Vantage.
It will include 14 mobile gas engines that are enclosed in a semi-tractor trailer, according to the Columbia Riverkeeper report. To fuel the plant, VoltaGrid plans show the company would truck gas from the city of Moses Lake to Quincy, requiring 16 trips daily between the cities.
Grant County Public Utility District commissioners are also weighing new gas-fired power plants near Moses Lake or Quincy, and possibly investing in a natural gas plant in Idaho, a state without emission reduction targets, according to Columbia Riverkeeper’s account of a January 2026 commission workshop.
At that meeting, commissioners considered multiple proposals, including a new 40- to 120-megawatt gas-fired plant in the county, and 10 to 20 megawatts of natural gas fuel cell generators.
Another way
Sightline Institute researchers noted that much of the rush to build out gas infrastructure for data centers is to ensure enough energy to power peak demand, such as when residential heaters and air conditioners are running during extreme weather, on top of the overall base-load needed to serve customers consistently throughout the year.
Laura Feinstein, author of the Sightline report on new gas power being sought for data centers, recommended lawmakers in Oregon and Washington take a similar approach as officials in Texas, requiring that data centers power down when demand from most other customers is high and the grid is stressed.
A 2025 study from industry consultancy Energy and Environmental Economics, also known as E3, has been used in a growing number of cases to justify allowing more gas onto the grid in the Northwest for data centers, according to Feinstein, because the consultants said 9 gigawatts of energy would be needed by 2030. That would mean in the next four years roughly doubling the total amount of energy currently powering Oregon today.
But a separate study by Sylvan Energy Analytics, a firm founded by former E3 consultants, found that power would be unnecessary if Oregon and Washington required data centers to power down during demand spikes.
Leonard from Columbia Riverkeeper similarly cautioned against building out more fossil-fuel energy to meet data center demand, given unpredictability over how much energy the centers will actually use in the next few decades. She said the data center industry often inflates how much energy it will need.
“Because the energy demand of data centers is something that varies widely, data centers should not be used to justify new gas infrastructure,” she said.
Can AI Save More Energy Than It Consumes?
- Biglaw firm Duane Morris argues the energy sector's greatest AI-related risk is not surging power demand but failing to adopt AI tools fast enough to remain competitive.
- MIT research challenges industry claims that AI efficiency gains will offset its enormous energy consumption, while new data centers continue to be approved at record pace.
- AI shows genuine promise in clean energy applications -- from nuclear fusion modeling to EV battery recovery -- but the AI investment boom is simultaneously diverting capital away from next-gen energy research.
The artificial intelligence boom has created unprecedented pressure and anxiety in the energy industry. The public and private sector alike are expending enormous amounts of effort trying to quantify the amount of electricity that will be needed to power data centers in the near future, and get ahead of the skyrocketing energy demands headed for our already outdated and beleaguered electric grids. But the answer to the energy monster that AI is unleashing could very well lie in the application of AI tools.
A new article published by Biglaw firm Duane Morris argues that the most prescient AI-related risk for the energy industry is not the one posed by the demands of the sector itself, but the risk of falling behind in AI integration and application. The firm argues that the energy sector has an obligation to consider the ways in which large language models can be an asset, concluding that "AI should not be viewed only through the lens of risk avoidance."
"The risks of AI remain real and must be governed thoughtfully," the Energy Intelligence article goes on to say. "But in a sector responsible for critical infrastructure, the greater long-term risk may not be using AI too aggressively -- it may be failing to use it enough."
Indeed, proponents of AI adoption argue that although training and operating large language models eats up an enormous amount of energy (not to mention other finite resources such as water), AI will be instrumental in making a wide array of industries significantly more energy-efficient. In fact, through these widespread efficiencies, some experts say that AI has the potential to save more energy than it consumes overall.
However, critics say that these claims are overblown and the result of wishful thinking rather than rigorous modelling. A 2025 report from MIT challenges such claims, pointing out that touted efficiency gains have not yet come to fruition, and may not be forthcoming. And while numbers on AI's efficiency gains -- and even the amount of energy that AI is currently using -- are still lacking, new data centers are being greenlit at lightning speed.
"AI's integration into almost everything from customer service calls to algorithmic 'bosses' to warfare is fueling enormous demand," the Washington Post wrote in an article published last summer. "Despite dramatic efficiency improvements, pouring those gains back into bigger, hungrier models powered by fossil fuels will create the energy monster we imagine."
Moreover, it is just this fear of 'being left behind' that's fuelling the AI boom, arguably even more than actual demand. There is question as to whether rapid AI integration into everything from our energy grids to our electric toothbrushes -- no, really -- is going to create a more sophisticated and energy-efficient world, or whether it's just a resource-intensive bid to stay relevant in a rapidly changing global economy.
Wherever you stand on the issue of AI integration, it's increasingly clear that AI has some extremely promising applications in next-gen clean energy technologies. Researchers are using large language models to conduct "needle in a haystack" type inquiries to find the best methods and materials to advance nuclear fusion modelling, for example. In the renewable energy sector, AI is being used to improve forecasting of energy supply and demand for greater grid stability. And AI could even soon be used to give new life to dead EV batteries.
The massive energy needs of AI are also pushing increasing and intensified research efforts into cutting edge clean energy technologies such as nuclear fusion, advanced geothermal, and space-based solar power. But Big Tech is running on natural gas while it powers research into these clean energy ambitions. And, overall, research into next-gen energy is suffering from the AI gold rush as investors redirect their attention.
AI's role in the energy sector is anything but simple. And it's true that avoiding AI integration entirely won't solve the problem. But if the energy sector is going to eschew risk aversion and lean into the AI boom as Duane Morris suggests, it needs to have a strong policy foundation and a much smarter AI strategy going forward.
By Haley Zaremba for Oilprice.com
