Maryland Ratepayers Absorb Grid Upgrade Costs for Out-of-State AI Data Centers

Maryland has filed a complaint with federal energy regulators over transmission upgrade costs being passed to in-state ratepayers for infrastructure that largely benefits out-of-state AI data center operators. The reported bill runs into the billions of dollars.

The core dispute is cost allocation. Regional transmission organizations socialize grid upgrade expenses across ratepayers in their footprint. When large new loads — AI data centers drawing hundreds of megawatts — interconnect to the grid, the upgrades required to support that load get spread broadly, including to customers in states that see none of the economic benefit.

Maryland's complaint centers on a broken ratepayer protection pledge. The state argues the current cost allocation violates prior commitments about how transmission costs would be distributed. Whether FERC rules in Maryland's favor depends on how those prior commitments were structured and what the relevant tariff language actually says.

For engineers and founders building on cloud infrastructure, this is a slow-moving constraint worth tracking. AI training and inference workloads concentrate massive power demand in specific grid regions. That demand is now surfacing as a political and regulatory cost that states are unwilling to absorb quietly. Expect more challenges like this.

The practical effect: utilities serving AI-heavy regions will face harder interconnection negotiations. States without data center tax revenue have little incentive to subsidize transmission for operators they don't host. This creates geographic friction in where large GPU clusters can economically land.

Data center siting decisions increasingly need to account for transmission cost allocation regimes, not just land and power availability. A favorable power purchase agreement means less if ratepayer backlash triggers regulatory changes that shift costs back onto the operator mid-contract.

This is infrastructure policy catching up to a demand curve it did not anticipate. The regulatory lag is now measurable in dollars.

Source