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.
The post The Hidden Cost of AI: How Data Centers Are Straining Water, Power, and Communities appeared first on Dissident Voice.
This content originally appeared on Dissident Voice and was authored by Ella Mrofka.