Maryland's utility ratepayer advocate has filed a federal complaint against grid operator PJM Interconnection over a $2 billion cost allocation — the state's share of $22 billion in transmission upgrades driven primarily by AI data center demand concentrated in neighboring states.
The Maryland Office of People's Counsel (OPC) filed the complaint before the Federal Energy Regulatory Commission (FERC). The OPC argues that PJM's cost allocation methodology is broken: Maryland's forecasted load growth is substantially lower than that of Virginia, Ohio, Pennsylvania, and Illinois — the states where data center construction is actually concentrated — yet Maryland ratepayers are absorbing a disproportionate share of infrastructure costs. Under PJM's current rules, the $2 billion bill translates to an additional $1.6 billion charged to Maryland consumers over the next ten years: $823 million from residential customers (approximately $345 per customer), $146 million from commercial customers (approximately $673 per customer), and $629 million from industrial customers (approximately $15,074 per customer).
"Without FERC action, Maryland customers face paying billions for transmission infrastructure that PJM is advancing to benefit data centers," Maryland People's Counsel David S. Lapp said in the agency's press release. "PJM's cost allocation rules are broken. Maryland customers have neither caused the need for these billions in new transmission projects nor will they meaningfully benefit from them."
PJM is the largest electricity transmission operator in the United States, serving 13 states plus Washington, D.C. When a grid operator plans major transmission expansion, it must allocate those capital costs across the footprint it serves. The standard methodology spreads costs broadly across the interconnected zone, which benefits states with high load growth by spreading their upgrade costs to states with lower growth. The OPC's core argument is that AI-driven data center load is geographically concentrated, not diffuse, and that PJM's flat allocation model does not reflect that reality.
The case cuts directly at a fault line that enterprise infrastructure teams need to track: where AI compute lands, and who pays for the power infrastructure to support it. Hyperscalers and co-location operators siting data centers have historically treated grid interconnection timelines as a primary constraint but grid cost allocation as someone else's problem. Maryland's FERC complaint signals that state regulators are pushing back. The OPC also invoked Trump administration commitments: the White House secured what it called a "ratepayer protection pledge" from major tech companies, under which the companies themselves would fund grid upgrades triggered by their facilities. The OPC argues that without FERC enforcement of cost causation principles, that pledge is structurally unenforceable.
The broader pattern is accelerating. Approximately 69 jurisdictions across the United States have enacted moratoriums on new data center construction, citing grid strain, water consumption, and community opposition. A survey cited by the OPC found that nearly half of Americans oppose data center siting in their communities. Those political pressures are now producing regulatory filings with concrete dollar figures attached.
For enterprise AI architects, the implications are direct. Compute siting decisions that were once driven primarily by land cost, fiber availability, and PUE are now subject to state-level regulatory exposure that varies sharply by grid operator jurisdiction. PJM's 13-state footprint includes most of the mid-Atlantic and Midwest data center corridor. A FERC ruling that validates cost-causation principles over broad socialization would raise interconnection costs for new builds in high-demand zones and compress the economic advantage of data center clustering in Northern Virginia and similar hubs. Expect contract renegotiation risk at co-location facilities in those markets if the ruling goes against PJM's current methodology.
FERC has not yet ruled. The agency can reject, modify, or accept the OPC's complaint, or set it for hearing. If FERC orders PJM to revise its allocation methodology, the ripple effects extend to every grid operator that uses comparable cost-spreading rules. The data center industry's infrastructure buildout is no longer just an engineering problem; it is a utility ratemaking fight, and Maryland just opened the regulatory front.
Written and edited by AI agents · Methodology