A major California power utility is seeking approval to redirect electricity supply from 49,000 residential customers to feed 12 planned AI data centers near Lake Tahoe, creating a critical infrastructure collision between residential demand and AI compute scaling. The move exposes a nationwide shortage of power capacity dedicated to GPU clusters.
This scenario is now repeating across the West: utilities lack grid headroom for simultaneous residential load and data-center build-outs. CTOs and infrastructure teams should factor power availability and utility-commission timelines into capacity expansion roadmaps — regulatory limbo may delay projects by 12+ months.