PJM’s board has proposed a backstop capacity auction and a data-center curtailment framework, handing FERC the power to decide how the largest U.S. wholesale power market pays for its last tranche of reliability. The combination matters now because PJM’s existing capacity auction already cleared at record levels, data-center load is growing in concentrated zones, and the design choices FERC makes will determine which entities pay for reserve margins and which projects get built.
Why PJM’s board reached for a backstop and curtailment plan
PJM runs a forward capacity market in which suppliers bid to provide reliability three years ahead. The auction produces a single market-clearing price that is meant to signal where new generation, storage or demand response gets built. Notably for this proposal, the board is now signalling that the regular auction may not be enough: if the auction clears below the reliability target, a backstop capacity auction would provide a separate, potentially out-of-cycle way to secure the rest.
Also part of the package is a way for data centers to be connected on terms that tolerate interruption. Called a curtailment framework, this would allow a large data center to take a service agreement that permits PJM to drop or reduce its load during grid emergencies, in exchange for lower costs or a different contracting route. That matters for queue processing; if a hyperscaler can agree to be interrupted, its interconnection can move faster because it does not require the worst-case peak capacity on the local grid.
What is crucial is what the board did not decide: the specifics. Under the source, the proposal will go to FERC where the parameters, price caps and scrap rules will become the true test. As ClearView Energy Partners put it, FERC’s response will influence utility investment, project schedules and costs.
The context is that PJM’s forward capacity market has already produced a record price for the 2025/26 delivery year, commonly reported around $270/MW-day – up dramatically from single- and low-double-digit levels when supply was looser. This has nearly made new generation more rewarded, the same dynamic has exposed that auctions alone cannot produce a physical commitment faster than interconnection queues, and this gap is why an extra “backstop” helps.
Cap and curtailment in one package
If the backstop auction is conceived as a rarely used emergency valve, it will instead matter most for the expected price-setting dynamic: capacity suppliers that are unsure of clearing the basic auction could hold back some ties, knowing the backstop auction might pick them up at a price determined administratively. That will load some of the biggest market questions onto FERC and the PJM Board…
That is different to maintain: For data centers, the curtailment option can lower capacity commitment in the auction and, in a more precise reading, do what demand response has long been: turn a load into a supply-like commodity. By pre-signing an interruption agreement, a data center can reduce coincident peaks and lower system capacity points. In comparison to the old single-first, this is a move from extreme-baseline rating to day-of delivery mandate.
The two proposals are related because they allocate reliability risk to the customer in the same location. PJM’s reserve margins are set against a peak hour that data centers are increasingly pushing, so when the board gives data centers to curtail, part of the price risk of peak demand is instead transferred to the data center’s willingness to be flexible.
The magnitude matters. In rough arithmetic a 1,000 MW of capacity at the current auction level is on the order of $100 million per year in capacity revenue for that one project, not counting unit-contingents. For planners those sums will represent the difference between a speculative generation market and real anchor contracts.
Cross-cutting: control beyond the RTO
The proposal arrives as other regions are designing similar tools, though with a different timeline. Texas and other markets are experimenting with data center load curtailment and price-based subscription options, and capacity markets in MISO and SPP are facing the same focus on a “technology-neutral” capacity – and the same problem of generator retirements before replacement on line. PJM is not the first to wrestle with this, but it is the one with the most to lose: its footprint is the largest, and interconnection costs are climbing with reconnections and smaller foundations.
This is leading to a second big theme: interconnection access during construction. Data center developers want specific capacity commitment at peak; utility planners have to choose whether to invest in network upgrades or in a third-party generator bringing peak power to the market. The backstop auction helps the merchant generator secure revenue certainty: by offering a fallback, it also gives FERC authority over the timestamps…
From a merchant-project perspective, the curtailment proposal is better read as the first positive step for flexible supply. If `load equivalent` can voluntarily interrupt, the resulting offers gain the same value in the capacity auction as a gas plant’s availability bid – with the benefit of microsecond response. But exactly how this is measured – whether such curtailment is a blockable priority, or it must be interrupted before other firm load – will decide if it is one resource or many in the future.
Who this affects
- Utility planners – a backstop that caps risk means an increased reason to build around a known peaker reference. Expect them to rework load forecasts into scenarios with and without voluntary curtailment, and to adjust network upgrade orders for connection of data center service.
- Capacity and storage developers – the backstop auction with its administrative price could replace the surprise of a “lost” year in the underlying base market; developers should price in the chance that some of their expected annual revenue will be set by the backstop rather than the base market, and prepare sites that satisfy both bidding processes.
- Data center / procurement teams – a curtailment option converts power purchase negotiations into risk trade-offs: pay premium for full firm service or accept a minute-scale interruptible contract that FERC may require to be structured identically; delays in negotiations can be caused.
- Policy/economic investigators for states – the FERC response will also determine if costs of the backstop auction are spread across all load or left to winners; retain charge-schedules. that’s the line between reliability justice and economic development.
What to watch next
- Whether PJM formally submits the backstop auction and curtailment tariff to FERC, and whether they are filed as a single package or two separate proceedings.
- How FERC treats price cap and rate for the bull hurried tariff – decisions will come in public orders, but the bigger tell is whether FERC allows trigger terms to be set in advance or only on a case-by-case basis.
- Any protests from utilities or utilities associations, especially seeking transparent allocation of backstop auction costs, or from data centers complaining about mandatory churn.
- Next PJM capacity auction: the market’s clearing levels after the proposal is known will reveal whether bidders expect a rate of “backstop” always clear away the floor price shown in the core market.
Bottom line
The board gives FERC a chance to do the one legal action it cannot do on its own: keep PJM a market, not a black box – and define explicitly how much preserving that costs. The entity that determines the price cap on backstop capacity and the granularity of the data center curtailment reward will set the center of gravity for all PJM externalives for the backend of the 2020s.
Read the full report at Utility Dive.
Note: facts and figures attributed above to Utility Dive reflect that outlet's original reporting. Broader context, cross-sector connections, and forward-looking scenarios reflect independent analysis by our editorial team.
About this article: Drafted by Energy Ai with AI-assisted research and writing based on public reporting, then reviewed under our editorial process before publication.
Leave a Reply