Climate goals go up in smoke as US datacenters turn to coal (www.theregister.com)

🤖 AI Summary
US datacenters are increasingly turning back to coal as the immediate source of bulk power after a surge in AI-driven electricity demand collided with high natural gas prices and constrained grid capacity. A Jefferies research note cited by The Register says coal generation is up nearly 20% year-to-date and the firm has raised its coal-generation forecast by ~11%, expecting elevated utilization through 2027 as operators rush to connect new capacity and accelerated load growth is anticipated in 2026–2028. The shift matters because it can extend the operational life of the dirtiest fossil‑fuel plants, worsen local air quality, and blow a hole in corporate climate commitments: Morgan Stanley projects datacenters could emit 2.5 billion tonnes of GHGs by 2030 driven largely by generative AI. Technically, developers have favored onsite natural‑gas turbine plants for fast, reliable capacity, but current gas economics make coal the cheapest available stopgap for some grids. Policy headwinds for renewables and statements from US officials prioritizing AI capacity over near-term climate goals further reduce the incentive to invest in cleaner alternatives (including unproven options like small modular reactors). The upshot: datacenter buildouts will likely prioritize getting online quickly with whatever power is available, risking a temporary—but meaningful—backslide in decarbonization.
Loading comments...
loading comments...