Data centers have become a focal point of environmental concerns and political debate, drawing criticism over their electricity and water consumption amid the rapid growth of artificial intelligence (AI). Activists and some Democratic lawmakers have called for restrictions on new data center developments, citing their significant resource use. In response, experts caution that many of these worries may be exaggerated or based on misinterpreted data.

In New York, where a one-year moratorium on approving large data centers was recently enacted, critics argue these facilities consume disproportionate amounts of energy and water, often harming local communities. Demonstrators have employed slogans such as “data centers are energy vampires” to highlight their opposition. However, researchers contend that such rhetoric oversimplifies and overstates the issue.

Jonathan Koomey, an energy researcher specializing in data center electricity consumption, pointed to the overestimation of future data center impact, suggesting some projections are inflated by a factor of three to five. He also emphasized that double-counting of energy use in certain studies has contributed to misconceptions about their scale. According to Lawrence Berkeley National Laboratory data, U.S. data centers used approximately 270 terawatt-hours (TWh) of electricity in 2026, representing less than 10% of national consumption and about 60% of the electricity used for air conditioning. For context, total U.S. electricity use was estimated at 4,200 TWh in 2025.

Concerns about water use have also drawn attention, often citing an exaggerated figure from a Chilean data center reportedly consuming vastly more water than reality supports. Karen Hao’s book “Empire of AI” popularized a statistic that subsequent analysis found inflated by a factor of 1,000. In New York, data centers accounted for only 0.2% of the state’s water use and 3.5% of its electricity consumption in 2025, according to the Taxpayers Protection Alliance.

David Mytton, a sustainable computing researcher at Oxford University, noted that while data centers consume millions of gallons of water daily, this amount is minimal compared with widespread agricultural use, which accounts for around 80% of water consumption in the U.S. He highlighted that even though more data centers are being constructed, they are becoming increasingly efficient, with companies like Amazon aiming to achieve water-positive operations by 2030.

Still, experts acknowledge data centers can pose localized challenges, especially in communities already facing water scarcity. Mytton emphasized the importance of community engagement and tailored solutions to address these issues without halting technological development altogether.

Beyond environmental concerns, Koomey suggested that some anti-data-center activism may be amplified by foreign actors seeking to sow discord within the United States. He acknowledged that while some protests arise from genuine local worries, others form part of broader anti-technology and anti-capitalist movements online.

This debate over data centers appears to be intertwined with broader anxieties about AI’s impact on society and the economy. Experts urge focusing on substantiated issues rather than alarmist narratives. They caution that misinformation about resource use threatens to distract from more pressing conversations about the future of AI and its role in economic transformation.