The rapid expansion of data centers to support artificial intelligence workloads is reshaping energy demand and pushing local infrastructure to its limits. Researchers at Georgia Tech highlight that training large AI models requires immense computing resources, often requiring tens of thousands of graphics processing units to run for days. That demand is driving a sharp increase in data center construction, particularly in regions with reliable power and favorable tax incentives. These facilities could account for as much as 12% of U.S. electricity use by 2028, tells Tech Xplore.
As these data centers proliferate, communities are confronting rising electricity costs, debates over who should fund grid upgrades, and concerns about water use for cooling. In some states, disputes have already arisen over utility investments to support these new loads, with consumer advocates warning that data center-driven infrastructure spending could significantly increase electricity rates in the short term.
Economists and energy experts are also weighing the broader effects on regional grids and emissions. One Georgia Tech researcher uses economic modeling to show that while national increases in energy use and carbon dioxide emissions from AI might look modest at the macro level, the impacts are uneven and often concentrated where data centers cluster. This unevenness can amplify local price pressures and environmental stresses, contributing to public backlash and regulatory scrutiny.
To help people grasp these trade-offs, another research team developed an interactive simulation that casts players as utility decision-makers balancing grid reliability, wildfire risk, renewable integration, and cost. The goal is to make clear that choices about grid modernization are not abstract; they affect affordability, resilience, and the transition toward cleaner energy systems.
The article underscores that managing the data center boom responsibly will require better data, a thought-out policy, and broader public engagement.