The rapid buildout of AI data centers is poised to reshape U.S. electricity demand, raising the prospect of higher power bills for households. As AI training and inference workloads surge, data centers could account for as much as 9%–10% of U.S. power consumption within a decade, up from roughly 2%–3% five years ago, according to industry projections cited in the piece. Meeting that demand will require new generation, transmission and grid upgrades—costs that utilities have historically recovered from ratepayers through higher tariffs and surcharges. Tech giants are racing to secure dedicated power supplies, exploring everything from massive renewable PPAs to small modular reactors, while utilities in high-growth states warn of mounting grid stress. With residential electricity prices already elevated in recent years, the AI-driven load could add persistent upward pressure unless supply expansions keep pace.





























