Data centers have recently become a focal point of anti-AI sentiment: celebrities, politicians and protesters are calling for slower growth. The debate is particularly intense in the United States, where public opposition has risen sharply.
Why this matters
Rising public resistance could threaten the expansion of infrastructure needed for AI development and operation. Current estimates put at least 4,000 data centers in operation nationwide, with roughly 3,000 more under construction or planned.
What sparked the protests?
Criticism centers on data centers' water and energy use, as well as local environmental and community impacts. One poll indicates that 70% of Americans oppose the construction of new data centers. As a symbolic protest, former NFL player Jason Kelce announced a partnership between his Garage beer brand and Liquid Death asking Americans to send actual urine to data centers to protest their water usage.
Political implications
The issue has reached electoral politics: Republicans warn that deteriorating public perceptions of data centers could have political consequences and weaken previously steady support from the party and investors. Disputes over data center projects have surfaced in political races in Michigan, Wisconsin, Texas and Ohio.
How data centers work and why they're important for AI
Data centers are large facilities that house rows of high-performance computers called servers. They power video streaming, searches, email and other cloud services. Modern centers supporting AI use specialized chips—such as graphics processing units (GPUs)—that perform the computations needed to train and run models. Engineers guide the process, but servers do the bulk of the work.
Servers generate substantial heat, so facilities require cooling systems and often large amounts of water to prevent overheating.
Consumption and numbers
Globally, data centers accounted for about 1.5% of worldwide electricity consumption in 2024; the International Energy Agency (IEA) estimates that share could rise to 3% by 2030. A Berkeley Lab report updated in June 2025 estimates that U.S. data centers could consume between 9.5% and 15.3% of U.S. electricity by 2030. Regarding water, some data centers can use up to 5 million gallons per day for cooling alone; in 2023, U.S. data centers consumed roughly 17 billion gallons of water, primarily for cooling. Berkeley Lab also estimates an indirect water footprint of about 211 billion gallons associated with producing the electricity consumed by data centers, noting regional variation.
At the same time, newer closed-loop cooling designs can reuse water and significantly cut fresh-water demand. Compared to other industries, data centers use less water nationally than sectors such as cattle production, power generation, residential consumption, golf courses and steel manufacturing.
Energy management and environmental concerns
Before the AI boom, cloud providers often sited facilities near renewable energy sources. The AI buildout has pushed some developers toward fossil-fuel sources—such as gas turbines—to meet large, continuous power demands. Critics warn this could increase long-term fossil fuel demand and local air pollution from backup generators.
Industry response and what to watch next
Big technology firms remain committed to building data centers. Nvidia, which designs GPUs, is investing billions into infrastructure financing. Tech companies are also trying to persuade local communities to approve and support projects.
Going forward, key questions include how the backlash evolves and whether it becomes a major factor in upcoming midterm elections and future policy decisions.
