Tech companies seek federal oversight of electric grid amid rising AI energy demand

What This is

• AI-driven data-center growth is sharply increasing electricity demand, straining power grids and prompting calls for more oversight and planning.
• A United Nations University report estimates data centres already use nation-scale electricity and water and projects energy use and emissions could roughly double by 2030 as AI grows.
• State and local governments are debating rules for new data centers, including moratoriums, reporting requirements, cooling and noise standards, and changes to tax incentives.
• In Seattle, the city council is moving toward a one-year moratorium on new large data centers to write rules on energy rates, pollution standards, labor, and land-use requirements.
• In New York, lawmakers are advancing a one-year statewide pause on new large data centers while regulators study grid, water, and emissions impacts, despite Gov. Kathy Hochul’s skepticism.
• Utilities and grid operators face long interconnection queues and other bottlenecks, and firms are pursuing faster connection processes, grid analytics, and alternative power arrangements to bring capacity online.
• Companies and investors are shifting toward “AI power” infrastructure such as on-site generation, backup power, electrical equipment, and energy storage to meet data-center load growth.
• Water use is a major flashpoint, with Google proposing industry guidelines that aim for data-center operators to return more water to local watersheds than they consume by 2030.
• Governments are experimenting with constraints and conditions on data-center expansion, such as Ireland requiring new data centres to secure most of their power from new renewable generation before grid connection.

Why This Matters

• Grid capacity constraints can slow AI deployment and force choices between industrial growth, electrification goals, and household affordability.
• How governments assign the costs of new transmission, substations, and backup power will shape who pays—ratepayers, data-center operators, or taxpayers.
• The issue tests whether climate targets can hold as AI increases electricity use, cooling needs, and upstream pollution from added generation.
• Local decision-making over land, noise, and water pits community consent against national competitiveness arguments tied to AI leadership.