WASHINGTON (TNND) — Data centers are growing at a record pace as the United States races to remain competitive amid the global AI boom. But a new Pew Research Center analysis finds that behind this innovation lies a massive and growing strain on the nation’s energy and water systems. And for ordinary Americans, that could mean higher bills and new long-term environmental challenges.
The desire for AI power is exploding
According to the International Energy Agency (IEA), U.S. data centers will use 183 terawatt-hours of electricity in 2024, or about 4% of the electricity used nationally, according to a Pew review of federal and international data. This is roughly equivalent to the annual electricity consumption of the entire Pakistan.
What is the biggest driver? A hyperscale facility built specifically for AI models.
According to Pew and the IEA, a typical AI-optimized hyperscale center uses the equivalent power of 100,000 homes per year. The new mega-facility could use more than 20 times more power once it comes online.
In major locations, particularly in Northern Virginia, a collection of these centers currently consumes more than a quarter of the state’s total electricity supply, according to a report from the Electric Power Research Institute.
Ripple effects: Soaring banknote prices and grid stress
The higher the demand for electricity, the more power companies will have to upgrade their power grids, and the costs will not go away.
In the PJM electricity market (Illinois to North Carolina), data center growth increased capacity costs by $9.3 billion in 2025-26, according to Pew’s PJM data analysis. result:
+ $18/month in Western Maryland
+ $16/month for housing in Ohio
Looking ahead, Carnegie Mellon University predicts that U.S. electricity prices could rise 8% by 2030 for data centers and crypto mining alone, with even steeper increases in areas with the highest concentrations of data centers.
Water use: the silent problem that no one is ready for
Electricity is not the only resource being pushed to its limits.
Pew cited federal statistics showing that U.S. data centers used 17 billion gallons of water in 2023, much of it used to cool energy-intensive AI chips at hyperscale sites. And by 2028, hyperscale centers alone could consume 16 to 33 billion gallons per year. This is roughly equivalent to the annual usage of a medium-sized city in the United States. Although demand for cooling varies by season and facility design, researchers say the overall trend is clear: The growth of AI means demand for water is increasing rapidly.
Americans are unsure what to think about AI’s environmental impact
Pew’s August 2024 survey suggests that the public is not in consensus.
- 25% believe AI will have a negative impact on the environment
- 20% believe it is helpful
- 25% think it will be a mix.
- 30% don’t know
This uncertainty reflects reality. The United States is witnessing a resource-intensive surge without yet a clear long-term blueprint.
