PJM’s Base Residual Auction for the 2028/2029 Delivery Year closes on July 7, 2026. It is the third consecutive RPM auction to run under the $175 floor and $325 cap on unforced capacity that PJM first introduced for the 2026/2027 delivery year and that FERC extended on April 28, 2026 (Docket ER26-1556) through the 2028/2029 and 2029/2030 auctions, with annual adjustments. The prior two auctions both cleared at the cap. The 2026/2027 BRA settled at $329.17 per MW-day across the RTO (with the BGE and Dominion locational deliverability areas separating at $466.35 and $444.26 respectively). The 2027/2028 BRA, run in December 2025, settled at $333.44 per MW-day RTO-wide. Consensus expectation in the bidder community is that the 2028/2029 auction prints the cap again.
If that is right, the clearing price is not the read. The read is what cleared at that price, where it cleared, and what did not enter.
What the collar is actually doing
The collar exists because the auction’s pre-collar price cap, which scales off the cost of new entry of a reference combined-cycle, ran toward $550 per MW-day for the upcoming auction. PJM filed the collar in February 2026 with explicit political cover: a coalition of 13 governors, the White House National Energy Dominance Council, and the Department of Energy supported the extension. The framing was consumer cost, not market design. FERC accepted the filing on those grounds.
What the collar does on the demand side is bounded and explicit. It caps the bill impact of clearing into a tight reserve margin on every load-serving entity in the footprint. The 2026/2027 result alone moved capacity costs in PJM from roughly $14 billion the prior cycle to over $32 billion at the cleared price. The collar prevents that bill from doubling again on the same compositional supply curve.
What the collar does on the supply side is the question worth holding. The signal a competitive auction is supposed to send when reserve margins tighten is a price high enough to pull new entry through the construction window. The cap interrupts that signal at the level where most new gas, new storage, and reliability-must-run economics for old units sit, and substitutes regulatory and out-of-market mechanisms (state procurement, RMR contracts, capacity performance penalties, demand-side reliability programs) for the price-induced entry the model assumes. Whether enough megawatts will come in at $325 to balance the curve at the load forecast PJM is now carrying is the empirical question July 7 begins to answer.
The four lines in the result that actually matter
Strip the result down to four data points and the auction is legible.
New entry, in megawatts, accredited. PJM publishes the new-entry total in the post-auction report. The number to watch is not gross new-entry offers but accredited capacity from resources with a credible commercial operation date inside the delivery year. In the 2026/2027 auction the new-entry component was thin enough that the cleared volume sat almost entirely on existing capacity. In the 2027/2028 auction the same pattern held. The question for 2028/2029 is whether the additional 13 months of lead time and the FERC fast-track interconnection rules that took effect over the last twelve months are enough to bring new entry above 5,000 MW accredited. A new-entry print above that line would be the first auction in three cycles where the collar has not suppressed supply response. A print below it confirms that the constraint binding new entry is not the price.
Storage cleared, in megawatts, and at what ELCC. The ELCC accreditation values for 4-hour and longer-duration storage have moved each cycle. The 2028/2029 auction is the first to use PJM’s updated ELCC framework with the higher data-center-driven load profile. The cleared storage MW divided by nameplate gives an implied accreditation that can be benchmarked to the published class-average ELCC. A material gap (cleared MW well below the ELCC-adjusted potential supply) suggests storage developers either cannot deliver by 2028 or are sitting out the auction with the cap in place. A clear of the available ELCC-adjusted supply suggests storage is pricing aggressively into the cap and willing to take the regulatory risk of future collar changes.
Demand resources, in megawatts. The 2026/2027 auction cleared roughly 10.3 GW of demand resources, up from 7.8 GW the prior cycle. That growth came from large-customer interruptible load, much of it data-center, with the residential aggregator segment essentially flat. For 2028/2029 the line to read is whether DR keeps growing at the same gradient, with hyperscaler curtailment contracts and behind-the-meter battery enrollment now visible in the LSEs’ integrated resource plans. A 14 to 16 GW DR print would be the strongest data point for the thesis that capacity-market accreditation of demand-side reliability is the practical answer to the gas-turbine-OEM constraint between now and 2030.
LDA separations. When the RTO-wide cap binds, the locational deliverability areas with binding transmission constraints can separate at the LDA-specific cap (which is set higher). In the 2026/2027 auction, BGE and Dominion separated at locational caps in the mid-$400 range. The pattern reflects load growth in Northern Virginia and the Baltimore-Washington corridor running ahead of internal generation and import capability. For 2028/2029 the LDAs to watch are Dominion, BGE, PSEG (New Jersey), and ATSI (northern Ohio), where data-center load is concentrated and where transmission build-out remains slower than the load curve. The number of LDAs that separate, and the spread between RTO clearing and the binding LDAs, is the cleanest measure of where the transmission queue is failing to keep up.
What is genuinely uncertain
Three things could break the consensus that the cap binds.
A demand response surprise on the upside would lower the implicit demand curve enough to clear below the cap. The DR offer cap rules tightened over the last filing cycle, but the practical compliance and the hyperscaler contract terms now in the system suggest the DR offer book could be materially larger than the bidder community expects. If the DR clear runs above 15 GW, the auction could clear below $325.
A coal-and-old-gas retirement deferral surprise on the upside would do the same. The reliability-must-run conversation, the Brandon Shores and Wagner cases in the BGE zone, and the wave of state-level commission actions to keep aging units in service past their announced dates have all moved through public proceedings over the last six months. Net retirement volume for 2028 is meaningfully lower than the late-2024 IRP filings assumed. Higher retained existing supply means a softer clear.
A new-entry surprise on the downside is the third possibility, with the opposite effect on price but the same effect on credibility. If new-entry MW comes in materially below 5,000 accredited, the supply curve clears at the cap but the post-2029 outlook for the next auction (the 2030/2031 cycle, which the current collar does not cover) becomes the active concern. The market would print at the cap and the political conversation would immediately turn to whether the collar gets extended a third time.
Where this sits in the broader 2027 to 2029 picture
The 2028/2029 auction is the third reading of an experiment. The hypothesis behind the collar was that the prior auction prices were sending a signal to enter at a level that the supply chain (gas turbines, interconnection, EPC, financing) could not honor, and that capping the signal would not lose meaningful new entry because the new entry was not coming anyway. If the July 7 print produces 5,000 MW or less of accredited new entry, that hypothesis looks confirmed. If it produces 8,000 MW or more, the hypothesis looks wrong and the case for keeping the collar past 2029/2030 weakens substantially.
Either reading reshapes the planning math. Confirmation of the collar hypothesis pushes the burden of reliability further onto the four levers covered in earlier posts: storage acceleration, coal life-extension, demand-side reliability, and behind-the-meter customer deals. Rejection of the hypothesis puts price-based new entry back on the table and the collar’s days become numbered.
The clean energy thesis cuts both ways. A higher storage clear with credible ELCC values is the cleanest evidence that the storage build-out is becoming the load-following backbone of the eastern interconnection. A higher gas new-entry print is a sign that the OEM constraint is loosening faster than the IRPs assumed. Either is material to the 2027 to 2030 reliability picture. The auction result on or about July 21 (PJM typically posts results two weeks after close) will set the frame for the second half of 2026.
Watch list, in priority order
- Accredited new-entry MW (target: above 5,000 to break the prior pattern).
- Storage cleared MW divided by nameplate, against the published ELCC class average.
- Demand-resource cleared MW, against the 10.3 GW baseline.
- Number of LDAs that separate at locational caps and the RTO-to-LDA spread.
- Total cleared MW vs. PJM’s installed reserve margin requirement; the residual is what the collar is buying or hiding.
The clearing price will be the lede in the trade press. It is the least informative line in the report.
Sources
- PJM Inside Lines, “FERC Approves Capacity Auction Price Collar to Next Two Capacity Auctions”, April 28, 2026.
- PJM, 2026/2027 Base Residual Auction Report, July 22, 2025.
- PJM, 2027/2028 Base Residual Auction Report, December 17, 2025.
- FERC Docket ER26-1556 (Price collar extension).
- Utility Dive, “FERC approves PJM capacity auction price cap, floor.”