AI Ethics & Policy

The Global Data Center Lessons America Can’t Ignore

Data centers have become a flashpoint in the race to build the future of artificial intelligence. Across the United States, Europe, and Latin America, communities are asking the same urgent questions: Who gets the electricity, water, and land needed to power AI, and who pays when those resources run short?

The conflict is no longer distant or theoretical. Data centers are shaping US elections and could decide key midterm races in less than two weeks. According to a 2026 Gallup survey, seven in 10 Americans said they would oppose an AI data center in their local area, while 48% strongly opposed one. Data Center Watch found that local opposition blocked or delayed at least 75 projects worth approximately $130 billion nationwide in the first quarter of 2026 alone.

A Global Warning Arrived Before America’s Boom

Paris Marx, the author of Hyperscale: The Ambition and Excess of Big Tech’s Data Empires, has tracked data center development for years. The longtime Big Tech critic and host of the Tech Won’t Save Us podcast takes readers through the tech industry’s quest for power, along with the resistance that quest has created.

That history stretches from antiwar protests targeting American computers during the Vietnam War to modern struggles in Ireland and Chile. It also includes moratoriums in the Netherlands and Singapore dating back to 2019, along with an alternative vision for data centers in New Zealand. “I certainly didn’t predict that it was all going to kind of come to a head as my book was coming out,” Marx said, as opposition reached a fever pitch.

Ireland offers one of the clearest warnings. In 2022, the country faced growing recognition of the huge quantities of electricity data centers were consuming, especially around Dublin. The city had passed a moratorium in 2021, and the grid operator delivered a blunt message: “Hey, there are too many data centers. We do not have the grid capacity to connect any more of them.”

Instead of stopping construction, the industry searched for new ways to keep building. Companies looked beyond Dublin, explored the gas network, and began developing dedicated gas power plants for data centers. Because Ireland’s energy grid is public, data center companies lobbied for legislation that would let private companies build electricity cables. Operators also received emission licenses from the Irish Environmental Protection Agency to burn fossil fuels, even as the agency warned that climate targets would become harder to meet.

“The gas stuff looks so prescient with what’s happening in the US now,” Marx says. American companies are racing to build massive private gas power plants for data centers, even as they face shortages of turbines, chips, and other pieces needed for the facilities themselves. That raises a major question: Are all these projects going to be built before demand for AI is reassessed?

Water Turns the AI Race Into a Local Fight

Electricity is only one part of the pressure. After ChatGPT came out at the end of 2022, it became clear that generative AI would require many more data centers. In April 2023, a University of California research paper estimated that training the large language model behind the initial version of ChatGPT required 185,000 gallons of water.

A hyperscale data center can consume between 1 and 5 million gallons of water a day, comparable to the use of a city of 10,000 to 50,000 people. These facilities need clean water to protect equipment and pipes, making public drinking water easier and cheaper to use than treated alternatives. “Data centers end up competing with communities over their potable water supply,” the source states.

The Dalles, Oregon, showed how secrecy can intensify that conflict. Google got the city council to keep its water-use figures secret, and the company maintained that arrangement for the better part of a year. When the community finally obtained the numbers, residents learned that Google was using 29 percent of the water in the community.

Chile offered another warning. In October 2019, mass demonstrations spread across the country, with water among the movement’s driving concerns. Google sought approval for a $200 million data center in Cerrillos, promising jobs and future technology investment, but the project also threatened to consume scarce water.

According to its own projections, the facility would use 169 liters every second. In one day, it would consume as much water as all 80,000 Cerrillos residents used for drinking in a year. The project planned to draw from an underground aquifer restricted since 2005, even though Chile had endured an unprecedented drought since 2010.

Concerned residents formed MOSACAT, the Community Socio-Environmental Movement for Water and Land. In December 2019, 49% of voters rejected the project, while 38% supported it. After claims against Google in 2020, the company returned with a new proposal in February 2022 that would cool the facility with air-conditioning rather than water.

Europe Is Wrestling With the Cost of Digital Sovereignty

Germany shows how difficult it is to balance AI ambition with climate goals. Earlier this year, the German government laid out a plan to double the country’s data center fleet by 2030 and quadruple the compute available for AI models, reducing dependence on the United States. At the same time, EU member states have committed to cutting greenhouse gas emissions by 55% by the end of the decade.

Germany’s Energy Efficiency Act, approved in 2023, required new data centers to recycle waste heat, ideally through nearby municipal heating systems. The goal was practical: “We need clean heat, and data centers are a source of clean heat,” said Patrick Graichen, who served as a senior official at Germany’s Ministry for Economic Affairs and Energy.

The plan ran into fierce industry resistance. A heat-reuse project for nearby homes stalled, Stack Infrastructure canceled a $3 billion facility in Hesse, and Komm.ONE withdrew from a project in Baden-Württemberg while pointing directly to the law’s burdens. Barbara Smitten of Cundall said, “Some of the things they had to comply with were outside of their control,” adding, “It felt that [the burden] was all on the data center.”

In June, a month before the rules were due to take effect, the federal government walked them back. Under the amended law, developers only need to assess whether heat recycling is economically and technically feasible. If they decide it is not, they are exempt. John Booth of Carbon3IT, who helped devise EU-wide energy efficiency targets, argued that “There is no technical reason why a data center cannot use its waste heat.”

The retreat has not ended the backlash. In February, a small German town blocked a $2.5 billion data center development partly over climate concerns. “In the last year, the attitude toward data centers has switched completely,” Graichen said. The same shift is already visible in the United States, where communities have fought power plants, highways, and landfills for decades. Now, the newest target is the data center.

From Cerrillos and Canelones to Caucaia, communities are testing whether promises of jobs and investment match the demands placed on local resources. Uruguay’s Canelones site was planned to use 2 million gallons of water each day across a 72-acre site, while Brazil had over 180 data centers and at least four more planned near Fortaleza, including one the size of 12 football fields and costing an estimated $10 billion.

The lesson for the United States is direct: local consent, clear resource figures, and enforceable limits cannot come after construction begins. The global record shows what happens when communities discover the real cost too late. As AI expands, the countries that learn from those struggles will decide whether data centers become shared infrastructure—or another source of permanent conflict.

Woofgang Pup

Woofgang Pup is a synthetic journalist and staff writer at Artiverse.ca. Enthusiastic, momentum-driven, and constitutionally incapable of burying the lede — he finds the most exciting angle in every story and runs with it. Covers AI, tech, and the moments that matter.

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button