By Alefiya Presswala
Communities Push Back Against AI Data Center Expansion
Editor’s Note: This is the first of two Dispatches on AI data centers, with a second article, by Ella Mrofka, to be published in January 2026.
There are more than 5,400 data centers in the United States, accounting for nearly 50 percent of the world’s estimated 11,000–12,000 data centers. As more companies invest in artificial intelligence (AI) and incorporate it into their business models, there has been a push across the country to build even more data centers for AI infrastructure. According to a report authored by the professional services network PricewaterhouseCoopers (PwC), “overall global data center investment was $250 billion in 2023 and is projected to quadruple to $1 trillion by 2027.”
Sam Altman, the CEO of OpenAI, once said: “AI will probably most likely lead to the end of the world, but in the meantime, there’ll be great companies.” While these centers benefit the companies building them, they have detrimental impacts on surrounding communities and the environment. Some corporate media outlets have reported on the issue, but most of the pushback—and much of the coverage—is from independent, grassroots community organizations, nonprofits, and citizen scientists and journalists.
What Is an AI Data Center?
Generally defined, an AI data center is a specialized physical facility that houses the infrastructure required to train and deliver AI services. Because AI systems are so complex and require a lot of energy, they need to have “high-performance computing capabilities,” as well as large amounts of storage, high-speed memory, and a cooling method. In fact, an AI data center requires so much energy that a regular data center (used solely for powering the internet, cloud services, and other digital functions) not designed for AI would collapse under the workload of an AI data center.
According to IBM, there are two main types of AI data centers: hyperscale and colocation. Hyperscale data centers are huge (typically 10,000 square feet or larger) and are built, operated, and owned by large cloud service providers and tech companies (e.g., Google, Amazon, Microsoft, Meta). Hyperscale data centers are specifically designed for large, high-intensity AI workloads, such as training large language models. By contrast, colocation data centers are smaller, shared spaces where multiple businesses can rent space to store their own servers and equipment via contracts that may last a couple of months or even years.
Where Are These Centers?
Companies that want to construct large data centers target areas with inexpensive real estate and weaker local governments, near large bodies of water. Companies looking to build data centers often receive tax breaks from state and local governments looking to attract large tech investors. For example, a 2025 CNBC study found that a Microsoft data center in Illinois received $38 million in specific sales tax exemptions. Another Microsoft data center in Washington state received $333 million in sales tax exemptions between 2015 and 2023.
The nonprofit organization FracTracker Alliance has created a map that shows locations of currently operating data centers in the United States and proposals for new ones. The yellow dots on the map, which represent proposed data centers, appear frequently in Virginia, Georgia, and Pennsylvania, and in other rural areas in Utah and Texas.
Virginia currently houses the most data centers in the United States, specifically in Ashburn, Virginia’s “Data Center Valley.” Texas is becoming a key state for data centers, and cities in Arizona, Illinois, Ohio, Georgia, California, and Oregon are other strong contenders for data center hubs because of lower costs and high connectivity to other places (like California’s connectivity to Asia with undersea cables through the Pacific, or Chicago’s technological connection to both the East and West coasts).
Environmental and Economic Impacts
As noted, all AI data centers require cooling systems to function. Most AI data centers use liquid cooling, meaning they use water (instead of just air) to remove heat from AI chips. This may mean that the water is placed directly on cold plates on the AI chips, or that the chips/servers in the center are placed in a dielectric fluid that absorbs the heat. According to the Environmental and Energy Study Institute:
A medium-sized data center can consume up to roughly 110 million gallons of water per year for cooling purposes, equivalent to the annual water usage of approximately 1,000 households. Larger data centers can each “drink” up to 5 million gallons per day, or about 1.8 billion annually, usage equivalent to a town of 10,000 to 50,000 people. Together, the nation’s 5,426 data centers consume billions of gallons of water annually. One report estimated that U.S. data centers consume 449 million gallons of water per day and 163.7 billion gallons annually (as of 2021). A 2016 report found that fewer than one-third of data center operators track water consumption. Water consumption is expected to continue increasing as data centers grow in number, size, and complexity.
The water that AI data centers use can come from multiple sources, including surface and groundwater, piped sources like municipal water, and/or “purified reclaimed water.”
There are also concerns about noise pollution, caused by hums and buzzes from server fans, HVAC systems, and other equipment used within the data center. Noise pollution disrupts wildlife by interrupting communication, altering animal behaviors, forcing wildlife to leave their habitats, and generally interfering with the balance of a given ecosystem. Additionally, noise pollution can affect water and soil quality and human health, often triggering stress or sleep disruption.
Companies and institutions that want to build data centers claim that these centers will create more job opportunities in the communities where they are building. However, while there may be jobs in data center construction, there are not many permanent positions. A 2017 report from the US Chamber of Commerce found that, on average, a data center employs 1,688 workers during the construction phase, but once it is running, it only provides 157 permanent jobs.
But the larger economic impact of these centers lies in their electricity consumption. In 2025 alone, the tech industry is expected to spend $475 billion on data centers. Currently, data centers account for 4 percent of the country’s electricity demand, but that number is expected to triple within the next three years.
A recent study published by MediaJustice shows that communities in the South are disproportionately affected by rising electricity prices, with rural farming communities and communities of color impacted the most. In South Carolina, for example, “data centers will account for 65 percent to 70 percent of all new energy usage in the state.” Along with the amount of electricity that data centers require, the energy they use frequently causes power outages, highly impacting the rural communities where these centers are often based.
In addition to this, the CNBC study mentioned earlier found that “42 states provide full or partial sales tax exemptions to data centers or have no state sales tax at all. Of those, 37 have passed legislation specifically granting sales tax exemptions for data centers, and 16 of those states have granted nearly $6 billion in exemptions over the past five years.” The analysis calls these state tax breaks a “losing proposition for taxpayers.”
Greg LeRoy, executive director of Good Jobs First, a nonprofit research group that tracks corporate subsidies and advocates for transparency and accountability in economic development, told CNBC, “When tax breaks don’t pay for themselves, only two things can happen: Either public services are reduced in quality, or everybody’s taxes go up in other ways.”
The Case of Lansing, New York
Terawulf, a Maryland-based technology company, has proposed converting an old coal power plant on Cayuga Lake in Lansing, New York, into a large-scale AI data center. The company originally gained revenues from Bitcoin mining operations in New York and Pennsylvania.
Lansing is a rural town in Tompkins County, just north of Ithaca, which is more urban (and home to Cornell University). Lansing’s municipal government and its population are smaller than Ithaca’s, which is likely why Terawulf targeted Lansing. Ithaca is also the base for many environmental nonprofits and a politically active community. Critics of the data center have shown up at Lansing town board meetings, citing environmental concerns, including noise pollution and the proposed use of Cayuga Lake water for cooling. Much of the investigative reporting on the proposed data center has been conducted by local outlets, such as the Ithaca Voice and Ithaca Times, and other independent investigative organizations.
Hunterbrook Media, which calls itself a “new type of newsroom,” publishes investigative studies without requiring readers to pay to subscribe to access them. In August 2024, they published a report on Terawulf that exposed issues with the company’s claims of sustainability and energy efficiency. Grizzly Research, which publishes research on publicly traded companies, authored a report in August 2024 in which they concluded, among other findings, that “TeraWulf is in the business of selling stock while enriching insiders and not in the business of becoming a legitimate crypto miner.” These studies have been cited by Tompkins County residents to argue that Terawulf’s presence in their community will not benefit them.
Residents have also created an Instagram account with the handle @no_datacenter_flx, on which they share important information and updates about the AI data center proposal, including town board meeting times and summaries. Their bio reads: “A grassroots movement to advocate against the AI data center that’s planned to be built in Lansing, NY. Protect our lake, protect our people!”
The people who run this account have most recently been posting a week-long mini-series highlighting the reasons why they oppose an AI data center in Lansing, citing environmental concerns, public health risks, and delays to the US’s transition to clean energy. These posts are extremely detailed and well-researched, showing how community journalists and concerned citizens are at the forefront of the fight against AI data centers. The social media account has been an essential—if ironic, since social media apps rely on data centers to function—tool for many Tompkins County citizens in understanding the issue and joining a movement.
How Other Communities Are Fighting Back
As the number and scale of data centers across the country expands, so too does pushback against them. A report from Data Center Watch found that in the second quarter of 2025 alone, opposition to data centers rose 125 percent. According to the report, “an estimated $98 billion in projects were blocked or delayed, more than the total for all previous quarters since 2023.” The report continues:
Community opposition continues to grow, with 53 active groups across 17 states targeting 30 data center projects in Q2 alone, bringing the total to 188 groups nationwide. During this period, 66% of the tracked protested projects were blocked or delayed. As development expands and media attention intensifies, local groups are learning from one another. Petitions, public hearings, and grassroots organizing are reshaping approval processes—especially in Indiana and Georgia.
In Michigan, residents of Saline Township are protesting a $7 billion data center project called Stargate, which is backed by politicians and major figures in the tech world, including Sam Altman (CEO of OpenAI). This project would be one of the largest data centers in the country. OpenAI and Oracle would use this new Michigan data center to house their infrastructure. Residents claim that the companies were not transparent with their plans to build in their town. More than a hundred people gathered at the state Capitol on December 16, 2025, to protest data center proposals across the state.
The Earth Island Institute writes about resistance in Tucson, Arizona, where a proposal for “Project Blue,” put forth by Beale Infrastructure, would take millions of gallons of drinking water from the desert for cooling purposes. Residents consistently attended town board meetings and wrote numerous emails and op-eds in local newspapers. “I feel like I learned more about Project Blue from the public than the city,” said city councilman Rocque Perez, revealing how concerned citizens take it upon themselves to communicate with each other and fill local news deserts. The Guardian reported that “on 6 August, in an unscheduled vote, council members unanimously decided to discontinue discussions with Beale, each sharing short speeches revealing sharp opposition to Project Blue. Tucsonans packing the council chambers cheered and celebrated; Beale executives, appearing stunned, were booed as they left.” Project Blue no longer.
According to the independent outlet Truthout, more than 230 state and local environmental groups joined together on December 8, 2025, to send a letter to Congress demanding a national moratorium on the construction of new data centers. Senator Bernie Sanders (I-VT) also called for a national moratorium earlier in December. Across the country, local community members are showing up to their town board meetings to actively oppose the construction of data centers and demand more from their governments.
The rapid expansion of AI data centers is often framed as the natural next step in technological growth, but the consequences and impacts of these centers reveal a larger, more complicated story. While these facilities may generate billions of dollars in profit for tech corporations and fuel the advancement of AI systems across the globe, they also impose enormous and detrimental environmental and financial burdens on the communities in which they are built. AI data centers are not being built for the public good when they increase electricity bills, cause power outages, and consume massive amounts of water in vulnerable, rural areas.
But, at the same time, a growing wave of resistance against the construction of these data centers shows that communities are not powerless. From Arizona to New York, residents, environmental organizations, nonprofits, and journalists are demanding that companies be held accountable and be more transparent about construction plans and negative impacts on communities. The continued blockage of $98 billion worth of AI data center projects is proof that taking informed action, showing up to board meetings, organizing, and engaging with one’s community can actually influence government decisions.
As AI continues to shape the global economy and affect our daily lives, decisions about the construction of the centers that house these systems need to be made with the long-term benefit of the public in mind, not behind closed doors where corporations can continue to exploit natural resources and other public goods.
Alefiya Presswala was a fall 2025 intern at Project Censored and has been published in other independent outlets, such as FAIR and the Rochester Beacon. She is currently pursuing a Bachelor of Arts in Journalism at Ithaca College’s Roy H. Park School of Communications and will be graduating in May 2026.