Lee Zeldin, the Administrator of the Environmental Protection Agency (EPA) weighed in as mass protests against AI data centers erupted across the country over the weekend, amplifying fears, many of them plainly false, about water consumption, electricity demand, and even claims that data centers are overtaking farmland.
Zeldin said the Trump administration is committed to ensuring these facilities are “built correctly,” while urging leaders nationwide to lean into AI infrastructure as the United States ramps up for an intensifying technological race with China.
EPA Administrator @LeeMZeldin reacts to @CityJournal's report that found data centers don't raise utility costs:
— Manhattan Institute (@ManhattanInst) July 20, 2026
"The president has been pushing this Ratepayer Protection Pledge. He meets with these hyperscalers, he wants these projects to be built correctly, and the reality… pic.twitter.com/SFYDS03eRa
"The president's been pushing his ratepayer protection pledge. He meets with these hyperscalers. He wants these projects to be built correctly. And the reality is, is that most Americans want the United States to be the artificial intelligence capital of the world," Zeldin said. "They don't want us to lose this race to places like China. That next level of looking at an individual project needs to understand the reality that there's a wide range of how these projects could get built."
For a local community trying to decide whether or not a project is good or not, you can ask yourself three simple things. Is this project going to make my bills go up or down? Is this project going to be a net taker from the grid or a net provider onto the grid? And three is, what is it going to do with the water? Is it a closed-loop data system? Is it going to be 100 percent recyclable?
"And there's been all this controversy and focus on the worst projects and the worst fears," he continued. "And there hasn't been enough focus on the data center projects that are out there in America that are checking all of these boxes. And if we want to win this race to make America the AI capital of the world, we need to be leaning into this. Leaders need to lead. And what you saw from the governor of New York, that is not leadership and her state's going to suffer because of it."
So far, the loudest fears surrounding AI data centers have all been actively contradicted by available data. Analysis from the Manhattan Institute indicates that, in many regions, data centers can actually help reduce electricity costs, while price increases are far more closely tied to state-level energy policies, particularly aggressive and often inefficient green energy mandates. The same pattern holds for water usage, another frequent point of concern. In reality, AI data centers across the country consume less water than golf courses or residential lawn irrigation, and a single facility often uses less water than a square mile of farmland.
What tends to get overlooked is the economic upside. Data centers are powerful drivers of growth, expanding the tax base and easing the burden on residents. In Loudoun County, Virginia, the global epicenter of AI data centers, these facilities now generate between 40 and 50 percent of local tax revenue, dramatically lowering property taxes for homeowners and funding public services at levels that would otherwise be unattainable.
Yet despite these benefits, a wave of opposition, often fueled by misinformation, continues to gain traction. Some have even raised concerns that foreign adversaries are amplifying these narratives to slow U.S. progress in a strategically critical industry. If that opposition succeeds, the consequences will extend far beyond any single project. The United States risks ceding ground in the AI race to China while states voluntarily wall off one of the most significant engines of economic growth in the modern era.
New York has already offered a preview of that path, with Governor Kathy Hochul’s decision to impose a one-year moratorium on new AI data center construction, a move that may prove less like caution and more like self-sabotage as the next technological era takes shape.