Who pays for data center power is now a fight at FERC
Power availability gated site selection. Cost allocation is becoming the next gate.
Five state ratepayer advocates told the Federal Energy Regulatory Commission on July 17 that it is failing to keep the cost of data center power off ordinary electric bills, and the complaint turns a utility-rate question into a site-selection one.
Why it matters
For two years the gating factor on a data center site has been whether power is available, with interconnection queues running three to four years. A second gate is now forming next to it: who pays for the grid upgrades a hyperscale load triggers. When a project’s transmission bill lands on residents who get nothing from it, local opposition stops being about noise and water and starts being about money, and that is a far more durable objection.
The advocates argue that FERC’s December 2025 show-cause order to PJM and five other grid operators addressed how co-located loads connect but never fixed cost allocation, so ratepayers keep funding upgrades that serve data centers. Maryland’s advocate projects roughly $1.6 billion in added transmission costs to state ratepayers over the next decade, including $823 million on residential customers.
The numbers
The fight sits inside PJM, the largest US grid operator and the market covering the data-center-heavy Mid-Atlantic. The near-term risk for developers is not a single ruling but the direction: cost-causation reforms, if FERC orders them, would push more of the interconnection bill onto the load that caused it, changing the economics of siting in constrained PJM pockets versus greenfield power markets.
What’s next
Watch whether FERC reopens cost allocation rather than just co-location rules, and watch states cut their own deals. Michigan just secured a pledge from large tech users to shoulder grid costs directly, a template other states will copy. Underwrite new PJM sites assuming the load pays more of its own way. See our data center power guide.