The KUTV Investigation Into Data Centers’ Environmental Impact Reveals That America Is Trading Long-Term Resource Security for Short-Term Tech Growth

The area outside Tremonton, Utah, where the next wave of America’s digital economy is expected to take hold, has an odd quiet. Fences, fields, and the odd truck coming onto the road. A building the size of multiple football fields, humming, drawing power, drinking water, and waiting for an AI future that no one can quite quantify yet, can be found somewhere in the planning documents. It’s the kind of scene that causes you to stop. It turns out that the cloud’s infrastructure is completely outside of the clouds.

The industry is in a difficult position as a result of the recent KUTV investigation into the environmental impact of these facilities. Currently, Utah has about fifty data centers, which is a small number compared to Virginia’s nearly seven hundred, but the approval pipeline is moving more quickly. Commissioners in Box Elder County recently approved yet another significant project. Naturally, the locals are uncomfortable. Furthermore, the professionals they are consulting are not providing neat solutions.

Utah Valley University’s Brandon Amacher, who studies emerging tech policy, told KUTV that a typical data center is a “massive black box dropped into your community.” It’s a thought-provoking phrase. The majority of locals don’t really understand what these buildings do, how much they use, or what they leave behind. Talking to people in towns like Tremonton gives the impression that the conversation is being held above their heads while the trucks are already arriving.

Part of the unease can be explained by the numbers. In 2023, American data centers consumed about 17 billion gallons of water and more than 4% of the nation’s electricity, mostly to cool the energy-hungry chips powering the AI boom. Those numbers are rising. Data center energy demand may double by 2030, according to some forecasts, but predicting anything related to AI at this time feels a lot like predicting the weather with a Magic 8-Ball.

The speculative nature of this moment sets it apart from previous infrastructure booms. Amacher put it bluntly: much of what is being constructed now is a wager on demand that no one has verified will truly materialize. Investors appear to be persuaded. Attracted by tax income and the appearance of being “AI-ready,” state governments are paving the way. Permits are being expedited by federal directives. In the meantime, the effects are absorbed by the local tax base, the local grid, and the local aquifer.

The similarities to previous rural-industrial waves—the coal towns, the fracking booms, and the 2010s warehouse explosions—are difficult to ignore. They all pledged enduring prosperity. Each presented a more intricate message. Despite their shiny server racks, data centers follow a well-known formula: decades of resource draw, a few permanent technical positions, and brief bursts of construction work.

Some locals told KUTV that they no longer recognize AI-generated content on their phones, which is a silent source of anxiety. Some claimed to completely avoid using the technology. Chatbots aren’t the main source of skepticism. It concerns whether anyone has taken the time to consider whether the trade—water, power, land, and certainty—is worth what the other party is promising.

Perhaps the forecasts are accurate. Perhaps all of the megawatts being committed now are justified by the demand for AI. Or perhaps years from now, someone will pass one of these black boxes in the Utah desert and question what we really thought we were purchasing.