AI data center 'frenzy' is pushing up your electric bill — here's why (www.cnbc.com)

U.S. electricity prices are rising in part because of a rapid build‑out of AI data centers, energy experts say, pushing residential retail rates up 7.4% year‑over‑year in September to about $0.18/kWh and outpacing inflation through at least 2026 in some regions. Large-scale server farms that power cloud AI are driving forecasted demand: data centers used 4.4% of U.S. electricity in 2023 and are projected by the Department of Energy to consume 6.7–12% by 2028. The International Energy Agency warns global data‑center demand could more than quadruple by 2030, with U.S. centers accounting for nearly half of domestic electricity‑demand growth. That surge is already feeding higher utility bills, increased household debt on overdue bills, and political pushback in states where grid constraints and weather‑related costs amplify price impacts. For the AI/ML community this has concrete technical and operational implications: rising power costs and strained grid capacity increase total cost of ownership and may slow expansion in high‑demand regions, forcing more careful site selection, demand‑response strategies, on‑site generation, and investments in energy efficiency (e.g., model sparsity, hardware accelerators, liquid cooling). It also raises regulatory and community risks as utilities pass infrastructure upgrades to consumers. Rethinking compute intensity, procurement of renewables, and coordination with grid planners will be essential to scale AI without exacerbating affordability and reliability issues.