Particle.news
Download on the App Store

AI Data Centers Hit Power Crunch as Senators Press White House on Soaring Bills

A new Cornell study warns the U.S. AI buildout could add tens of millions of tons of CO2 and strain water supplies by 2030.

Overview

  • In Santa Clara, completed Digital Realty and Stack Infrastructure facilities remain unpowered as city utility Silicon Valley Power undertakes a $450 million grid upgrade scheduled through 2028.
  • Five senators led by Richard Blumenthal sent a letter to OSTP and the Commerce Department criticizing “sweetheart” data‑center deals and requesting consumer protections and cost transparency.
  • The lawmakers cited rising electricity prices and asked who will pay for grid upgrades and water impacts, setting a Nov. 21 deadline for the administration’s response.
  • A peer‑reviewed Nature Sustainability analysis projects U.S. AI servers could add 24–44 million metric tons of CO2 and use 731–1,125 million cubic meters of water annually by 2030, with smart siting, cleaner power, and better cooling able to cut impacts by roughly 70–85%.
  • Utilities are delaying coal retirements and turning to natural gas and quick‑deploy generators, including a 10‑gigawatt gas request in Georgia, as forecasts show data centers could consume about 7–12% of U.S. electricity by 2030 and interconnection waits stretch up to seven years in hubs like Virginia.