"As of late 2025, the United States was home to 5,427 AI-related facilities and counting, or roughly 45% of the world's data centers. They're scattered throughout the country, taking up millions of square feet, often covering hundreds or even thousands of acres, dwarfing the typical comparison to football fields. Since modern AI runs on tens of thousands of heavy and power-hungry graphics processing units (GPUs), the energy pull from these centers can require hundreds of megawatts, or even a gigawatt, of power."
"As a result, the report noted, electricity costs in communities near large data center developments have risen by as much as 267% compared with 2020 levels. "The U.S. energy narrative in 2026 is defined by a paradox," wrote the team, led by strategist Monica Guerra. "America is producing record amounts of oil and exporting record volumes of natural gas, while electricity costs are rising in key regions, pressuring affordability and catapulting energy pricing as a defining political issue ahead of the midterm elections.""
Oil and natural gas production reached record highs while electricity costs are rising in key regions, pressuring affordability. Americans view soaring power bills as a new inflation indicator ahead of November. Rapid proliferation of data centers, often in rural areas, is being linked to local electricity price increases. Data centers and generative AI are forecast to account for roughly 75% of incremental load growth through 2030. By late 2025 the United States hosted about 5,427 AI-related facilities, roughly 45% of the world's data centers, occupying millions of square feet and sometimes thousands of acres. Modern AI can require hundreds of megawatts to a gigawatt, contributing to local cost spikes up to 267% since 2020 and elevating energy affordability as a decisive midterm political issue.
Read at Fortune
Unable to calculate read time
Collection
[
|
...
]