Create a large-load tariff class with minimum take, long contract term and exit fees

The proposal

A state regulator establishes a separate rate class for customers above a threshold, typically 25 MW, with terms that make the customer bear the cost its connection causes: a minimum take obligation of around 85% of contracted capacity regardless of actual use, a contract term matched to the depreciation life of the network assets built for it, collateral or creditworthiness requirements, and an exit fee if the customer leaves early.

Why it would work

Two distinct mechanisms, and the second is the more valuable one.

Cost causation. Network and generation capacity built for a specific large load is paid for by that load rather than spread across the customer base. AEP Ohio's tariff, approved by the PUCO on 9 July 2025, requires at least 85% of contracted capacity with a four-year ramp and exit fees equal to three years of minimum charges, explicitly to prevent "cost shifting onto general ratepayers" (KJK analysis). Dominion's Virginia GS-5 class, effective 1 January 2027, requires 14-year contracts, at least 85% of transmission and distribution costs and 60% of generation costs, and $1.5 million per MW in collateral (Inside Climate News).

Forecast filtering, which is the striking result. After Ohio's tariff took effect, AEP's large-load forecast fell by half. A take-or-pay obligation makes duplicate speculative interconnection requests expensive, and much of the apparent demand crisis is duplication: ERCOT's large-load queue grew from 63 GW to 226 GW in a year against roughly 23 GW of new generation. Pricing the option to connect reveals which requests are real, which prevents utilities building generation against a phantom forecast. That is a benefit no moratorium delivers.

Adoption is now broad: 23 states have approved at least one large-load tariff with 7 more pending as of May 2026 (Columbia Climate Law).

Implementation path

A state utility commission proceeding, initiated by the utility or by consumer advocates. The contested design parameters are the MW threshold, the minimum take percentage, the term, and whether existing customers are grandfathered. Expect opposition: Amazon, Google, Meta, Microsoft and the Ohio Manufacturers' Association all opposed the Ohio tariff as discriminatory, the PUCO rejected rehearing, and an appeal to the state supreme court was signalled.

Trade-offs and limitations

• It does not recover everything. Even under Virginia's approved class, the Piedmont Environmental Council calculates that 61% of grid upgrade costs still fall on ratepayers after the contract term ends, and residential bills still rose about $16 a month. A tariff reduces the shift; it does not eliminate it.
• Term versus asset life. A 14-year contract against a 40-year transmission asset leaves a residual, which is precisely where the remaining cost shift sits.
• Legal risk. Discriminatory-rate challenges are live, and an adverse ruling could unwind the class.
• It addresses cost, not the physical harms. Noise, water and air quality are unaffected. This is the instrument for one specific facet, and communities that adopt it while expecting it to solve siting disputes will be disappointed.
• Well-designed terms can attract load. Clear, predictable large-load terms are also a competitive advantage for a jurisdiction, so this is not straightforwardly a deterrent, which cuts both ways depending on what the community wants.