Data centres in the United States are becoming major consumers of electricity as the artificial intelligence (AI) revolution accelerates. In 2024, they accounted for about 4% of the nation’s total electricity use, a figure expected to rise sharply in the coming years. Experts warn that as demand for AI-driven computing power continues to grow, data centre energy consumption could more than double by 2030. The expansion of generative AI models, which require immense computational resources, is the primary factor driving this surge in energy demand.
Despite these projections, the exact future energy use of U.S. data centres remains uncertain. Studies suggest that by 2030, their consumption could reach between 300 to 400 terawatt-hours (TWh) per year, potentially accounting for 6–12% of the country’s total electricity use. These estimates vary due to multiple factors, such as advances in energy-efficient hardware, availability of renewable power, and regional variations in infrastructure development. Policymakers and researchers are paying close attention to these dynamics, recognizing that small shifts in efficiency or energy sourcing could have major national implications.
The rapid rise in electricity demand from AI data centres also poses significant infrastructure and sustainability challenges. As these centres expand, they place heavy loads on local power grids and water supplies needed for cooling systems. Regions hosting large clusters of data centres are already experiencing strain on energy infrastructure, prompting local utilities to reassess grid capacity and long-term planning. Balancing technological growth with sustainable energy management has become a pressing issue for both regulators and energy providers.
While the AI boom offers enormous potential for innovation and economic growth, it also highlights the urgent need for responsible energy governance. The Pew Research Center emphasizes that adopting renewable energy sources, improving transparency in data centre operations, and investing in efficiency measures are essential steps forward. If managed wisely, the AI-driven expansion of data infrastructure could align with sustainability goals — but without such measures, it risks deepening power shortages and environmental pressures across the U.S.
