Six Gigawatts Bought in Three Weeks

PJM has filed a two-part plan with FERC, submitted on 31 July and 7 August, to run a one-time Reliability Backstop Procurement auction for approximately 6 GW of new capacity — the amount it says is needed to restore its reliability position to breaking even. The auction window runs from 30 September to 21 October and offers 15-year capacity contracts for new power plants including natural gas, nuclear, clean energy projects and battery storage. The context is the December 2025 capacity auction for the 2027/2028 delivery year, which cleared at the FERC-approved price cap of $333.44 per MW-day, a 1,053 percent increase on the 2024/2025 auction's $28.92. Resources that can be built quickly, particularly battery storage, are expected to have a significant advantage. The NRDC broadly supports the backstop component and strongly supports the interim resource adequacy service, but plans to protest at FERC over the absence of locational requirements — arguing that a plant procured in Ohio to serve a Virginia data centre would have the data centre pay the plant's upfront cost and contract but potentially not the multi-state transmission needed to connect it.

The thousandfold price move is the figure to sit with, because it is not a market quirk. A capacity price is what a system pays to ensure enough generation exists to meet peak demand with a reserve margin. When it rises to the cap, the market is signalling that supply is not arriving fast enough relative to demand, and in PJM the demand is substantially data centre load appearing faster than generation or transmission can be built. That cost lands on every customer in the zone, which is why industrial energy users in the region have found their capacity charges rising for reasons entirely unrelated to their own consumption.

The build-speed advantage for storage is a structural point worth generalising. A gas plant is a four-to-seven year project including interconnection; a battery is twelve to twenty-four months. When a market is short and paying at a cap, the resource that can arrive inside the window captures the value, regardless of which resource would be cheaper over thirty years. That is why storage keeps winning procurements aimed at near-term adequacy, and it is a reason to expect the same pattern wherever load growth outruns construction.

The NRDC's locational argument is the one industrial readers should follow, because it determines who pays. Capacity procured without a locational requirement can be built where land and interconnection are easiest rather than where the load is, and the transmission to connect the two is socialised across ratepayers rather than charged to the load that caused it. For a manufacturer in the zone, that is the difference between a data centre paying for its own grid and a manufacturer paying part of it through network charges — and it is being decided at FERC over the next few weeks rather than in a distant policy process.