Power utilities across the United States are leveraging artificial intelligence demand projections to justify unprecedented infrastructure spending, raising concerns about consumer cost allocation for potentially unnecessary generation capacity. Industry analysis reveals a significant disconnect between projected electricity needs and available computing hardware that would actually consume the power.

Data center developers routinely submit electricity requests to multiple utilities simultaneously, hoping to secure capacity in at least one jurisdiction. This practice creates inflated demand signals as each utility independently evaluates requests without visibility into parallel submissions. Constellation Energy executives compared this approach to casting multiple fishing lines, with developers pursuing numerous options for single projects.

The capacity planning challenge intensifies due to semiconductor supply constraints limiting actual data center deployment. Current GPU production from major manufacturers supports approximately 19 gigawatts of total electricity demand through 2028, significantly below the 711 gigawatts utilities claim they need to provide. Twenty-six major investor-owned utilities project data center requirements nearly equivalent to peak continental US summer demand, highlighting the scale of potential overcapacity.

Regulatory responses are emerging to address speculative demand inflation. Ohio implemented new requirements forcing data centers to pay for 85% of requested electricity regardless of actual usage, resulting in demand projections dropping from 30 gigawatts to 13 gigawatts. Similar large load tariffs under consideration across twenty states aim to eliminate purely speculative interconnection requests.

Electricity customers in PJM regional markets face $7.3 billion in additional capacity costs attributed primarily to data center growth projections. The total electricity supply auction reached a record $16.1 billion, with analysts citing data center load growth as the primary driver of tight supply conditions and elevated pricing across regional power markets.