AI Boom Drives Up Electricity Costs for Ratepayers

The price of electricity in a 13-state region managed by PJM Interconnection reached its maximum cap for the third consecutive auction, signaling that the rapid expansion of artificial intelligence infrastructure is driving up costs for consumers. On July 14, the capacity auction for the 2028-29 period cleared at $325 per megawatt-day, the highest possible price, despite a shortfall of approximately 6.8 gigawatts needed for grid reliability. This situation indicates that the costs associated with building new power generation to meet increased demand are being socialized across all customers rather than being borne by the new, large energy consumers like data centers.

A report from Moody's Ratings on July 22 highlighted that the current system within PJM Interconnection lacks sufficient mechanisms to ensure new entrants, such as hyperscale data centers, cover the costs of new supply. Instead, these costs are distributed among all ratepayers. Moody's, an agency focused on pricing risk for investors, noted that other power markets in the U.S. require large-load customers to secure power through direct supply contracts that recoup the expenses of new generation and infrastructure over time. This suggests that PJM's current rules are out of step with practices in other regions.

The demand for electricity has surged, with PJM experiencing a record peak demand of 168.2 gigawatts on July 2, surpassing a previous record set nearly two decades ago by almost 3 gigawatts. Data centers are identified as the primary contributors to this escalating demand. In the latest auction, the total new generation capacity that cleared was only 525 megawatts, which is roughly half of what cleared in the preceding auction, according to an analysis by Syso Technologies. This reduced capacity clearing, coupled with record demand, exacerbates the strain on the grid and contributes to higher electricity prices.