Exelon filed the complaint on Tuesday. On November 4, the company announced it had asked the Federal Energy Regulatory Commission to require PJM Interconnection to hold its capacity auction two months before the delivery year, rather than three years ahead, arguing that the current long-lead schedule locks in high costs for customers on forecasts that keep changing after the price is set. The filing, docketed as EL27-11 and captioned “Complaint of Exelon Corporation v. PJM Interconnection, L.L.C.,” with the complaint itself dated November 3, contends that PJM’s three-year-forward Base Residual Auction is no longer just and reasonable. It is framed as a central piece of “The Exelon Promise,” Exelon’s affordability campaign; chief legal officer Colette Honorable said customers “cannot afford to wait” for a full package of market changes.
That sentence — filed, not ordered — is the discipline for everything that follows. FERC has not ruled on the complaint. No comment deadline had been set in anything reviewed as of this writing, and PJM had not publicly responded. What November 4 produced is a new venue for the affordability argument that has dominated PJM’s capacity market all year, and a testable proposition: that the timing of the purchase is what makes the market unjust and unreasonable.
The complaint, in its own terms
Exelon’s ask is deliberately small on its face. The company calls the change a “targeted” and “surgical” reform: move the buying decision closer to delivery, when PJM has better information about how much electricity the region needs and how much generation is available, and bring PJM in line with other U.S. capacity markets that already run prompt auctions. The release names no other market, so the analogy is an argument rather than a finding. The closest analogue is MISO, whose annual Planning Resource Auction is held in the spring, roughly two months before the June 1 delivery year — but MISO’s market differs from PJM’s in design and scale.
The evidentiary core of the complaint, as reported by RTO Insider, is an analysis by Emma Nicholson of Charles River Associates showing a positive correlation between shorter forward periods and the change in peak-load forecasts between the Base Residual Auction and the third Incremental Auction, held about four months before the delivery year. The logic: the further out the forecast, the more it moves before delivery, and the more of the bill is set on numbers that turn out wrong. A purchase two months before delivery prices capacity against a forecast that has had nearly three years to converge.
Exelon is equally careful about what the change would not do. “A move to a prompt auction schedule alone will not solve the affordability challenges facing the region,” the release states, and the company says PJM “urgently needs more generation” to keep pace with demand. The release gives no price figure, promises no household savings, and says nothing about which delivery year would be first or how the transition would work. Without a start date or a plan for the auctions already scheduled, the complaint asks the Commission to design the transition as well as approve the destination.
Why timing is the argument now
The complaint lands after a month in which PJM’s load forecast has been shown, twice and in opposite directions, to be wrong at the scale that decides billion-dollar procurements. On November 2, Charles River Associates told PJM’s Load Analysis Subcommittee that data-center demand reaches 30 GW by 2030 and 82 GW by 2047, with an 11.5 GW gap by 2030 between the load developers want and what the system can power at 1-in-10 reliability. Five days earlier, Jefferies analysts reported that PJM utilities including PPL Electric, FirstEnergy, and AES Ohio had lowered their near-term large-load forecasts by 6.5 GW — a cut that could shrink the refiled backstop from roughly 6.8 GW to about 630 MW. Both can be true at once: the long-run pipeline keeps growing while the near-term forecast falls, because cancellations and delays land in the near term first.
The monitor’s accounting gives the cancellation side a number. PJM’s Independent Market Monitor calculated that 4,017 MW — 42 percent — of the forecast increase in data-center load between the 2026/27 and 2028/29 delivery years had been canceled or delayed as of September 21. Nearly half the growth the July auction priced in had evaporated before the refiled backstop proposal was even written. That is the error a prompt auction is designed to remove: a July 2026 auction that cleared at the $325/MW-day cap yet procured only 138,318 MW, plus 10,864 MW under Fixed Resource Requirement arrangements, leaving a 6,831 MW shortfall against a 156,013 MW requirement.
But the same month’s evidence cuts the other way. CRA’s 11.5 GW gap is a warning that the forecast can undershoot as well as overshoot: if demand follows the soft trajectory and supply follows the status quo, load arrives faster than the system can serve it, and a prompt auction held two months before delivery cannot conjure generation a three-year-forward auction failed to finance. A prompt auction prices the forecast error out of the bill. It does not price the supply shortfall out of the system. The distinction is the difference between a cheaper auction and a reliable one, and Exelon’s filing, to its credit, does not pretend otherwise.
What a prompt auction would and would not change
Under a prompt schedule, the capacity price would reflect the load forecast and available supply as known roughly two months before the delivery year begins, rather than as projected three years earlier. The demand curve would be drawn against a peak forecast that has absorbed cancellations, delays, and the firm-growth data PJM is only now beginning to collect through the Large Load Registry FERC accepted on October 9. The reserve margin would be calibrated to accredited capacity that actually exists, rather than to projects that may or may not materialize. And the bill would no longer be set on load that, as the monitor documented, had a nearly even chance of never arriving.
What does not change is the harder problem. A capacity auction held two months before delivery can only buy what already exists, what is returning from retirement, or what is close enough to completion to commit. It cannot finance a new power plant, any more than a one-year commitment at the collar’s cap could: the July auction summoned roughly 525 MW of new generation and upgrades against a 6,831 MW shortfall. If the system is short two months before delivery, the prompt auction will say so at whatever price the demand curve produces — and the price it produces for a short system, without a collar to cap it, is the question nobody in this debate has yet answered in public. The gain is also modest: PJM’s third Incremental Auction already buys about four months before delivery, so a prompt schedule moves the main purchase only two months closer — while removing the forward price signal developers and lenders use to judge new entry.
The December auction and the March backstop are now in the way
Whatever FERC decides, the calendar will not wait for it. The Base Residual Auction for the 2029/30 delivery year closes December 15, capped at $337.83/MW-day, against a reliability requirement of 158,712 MW — 2,699 MW higher than the year before. It is the fourth and final auction under the price collar, and the first since FERC’s October 9 order accepted the Large Load Registry and rejected the curtailment and bring-your-own-capacity frameworks. December will clear on the existing demand curve — the registry tool is not expected until February 2027 — with generators weighing a one-year commitment at the capped price against holding their best offers for the backstop’s 15-year contracts at up to $555/MW-day. A prompt-auction order would arrive after December has already set the price for 2029/30, and the complaint says nothing about what happens to that auction or the ones scheduled behind it.
The backstop is the sharper collision. PJM’s amended Reliability Backstop Procurement, filed October 29, would buy the July shortfall on 15-year contracts: bid window March 1–28, 2027, final target April 28, results May 10. Its target is explicitly provisional, to be recalculated with updated information and reduced by bilateral placements — and the Jefferies arithmetic suggests it could land anywhere from roughly 630 MW to the full 6,831 MW. A prompt auction and a backstop answer the same question in opposite directions: the backstop buys long-term certainty for supply that does not yet exist; the prompt auction buys short-term accuracy for supply that already exists, on the theory that the forecast, not the fleet, is the problem. They cannot both be the market’s primary procurement mechanism — and the complaint does not say how the two would coexist.
The affordability argument has changed venues
For most of this year, the affordability fight has been waged in statehouses: governors backing the price collar, state commissions litigating cost allocation, the Pennsylvania commission pressing who pays for data-center growth, and rate cases carrying the capacity price into household bills. Exelon’s complaint moves that argument to FERC, and moves it in a form FERC is institutionally equipped to decide — a Section 206 complaint alleging that an existing tariff provision is unjust and unreasonable, with evidence attached, rather than a political demand for a lower price. That shift in register is why the filing reads as strategy rather than theater.
It also matters who filed it. Exelon is a load-side complainant: its operating utilities buy capacity for their customers across PJM rather than selling it. That gives the affordability argument a different kind of standing: the party asking for the change is the party whose customers pay the bill. But it also frames the opposition that will form. Generators and developers who rely on forward price signals to finance entry can be expected to argue that a prompt auction destroys the investment signal the market exists to provide. Consumer advocates may welcome the accuracy argument while fearing the volatility one: a prompt auction in a short system, uncapped, could produce prices higher than the collar ever allowed, with two months’ notice. Neither position has been filed yet; FERC’s docket will be where they are tested.
The state dimension has not disappeared, either. The backstop’s cost-allocation fight — zones paying on firm large-load growth between the actual 2026 summer peak and the 2028/29 forecast — is still pending before the same Commission, and states that have pressed cost causation will read a prompt auction through that lens: a more accurate forecast is also a fairer basis for allocating costs, if the growth figures are complete. The registry compliance filing, due November 9, is the first test of whether those figures will be complete. The complaint does not mention the registry, but its logic depends on it: a prompt auction is only as accurate as the load data available two months before delivery.
What to watch
Four things, in the order they will become visible. First, FERC’s handling of the complaint: a comment deadline, PJM’s answer, and interventions revealing who lines up on each side — and whether the Commission treats a narrow schedule change as separable from the wider redesign Exelon says stakeholders are still working toward. If FERC sees it as a standalone fix, it can move quickly; if it sees the schedule as inseparable from the demand curve, the collar’s expiration, and the backstop, the complaint becomes one more thread in a redesign with no deadline.
Second, the transition mechanics the filing does not supply: which delivery year a prompt auction would first govern, what happens to the December 2029/30 auction and those scheduled behind it, and whether the backstop survives, shrinks, or is superseded. FERC will have to answer them, and the answers will determine whether the prompt schedule is a reform or a disruption.
Third, December itself. The auction closing December 15 will be read as evidence for or against the complaint’s premise: a shortfall on a rising requirement would support the claim that the forward market is failing, while a strong supply response would support the claim that it is, slowly, working. Either way, December’s parameters will be the last forward-auction data point before FERC weighs whether there should be another one. The registry compliance filing due November 9 belongs on the same watchlist: the prompt auction’s case rests on better information close to delivery.
Fourth, and most consequentially, the post-collar design debate that December’s results will trigger and that this complaint now joins. The collar expires after the December auction. The first auction with no collar will be for the 2030/31 delivery year, on the normal three-year-forward schedule PJM is working back toward — unless FERC orders otherwise. Exelon has now asked the Commission to order otherwise, on affordability grounds, with evidence that the forward auction’s forecast is wrong in both directions at once. The honest reading holds two facts: a prompt auction would make the capacity bill more accurate, and it would remove the forward price signal on which new entry depends. December will update the numbers. It will not resolve the trade.
Sources: Exelon, “Exelon Seeks to Protect Customers from ‘Unreasonable Costs’ in New FERC Complaint” (Nov. 4, 2026); RTO Insider, “Exelon Asks FERC for Prompt Capacity Auction in PJM” (Nov. 4, 2026); FERC eLibrary, “Complaint of Exelon Corporation v. PJM Interconnection, L.L.C.,” Docket No. EL27-11 (complaint dated Nov. 3, 2026); PJM, 2028/2029 Base Residual Auction results (July 14, 2026) and 2029/2030 planning parameters; PJM, “PJM Files Amended Reliability Backstop Proposal,” Inside Lines (Oct. 29, 2026); FERC, 197 FERC ¶ 61,019, Docket No. ER26-3515-000 (Oct. 9, 2026); FERC, 196 FERC ¶ 61,245, Dockets ER26-3380-000 and EL26-108-000 (Sept. 29, 2026); Utility Dive, “PJM’s revised data center backstop procurement plan reduces utility risks: ICF” (Nov. 3, 2026), reporting Jefferies’ Oct. 28 note; Charles River Associates, “PJM Data Center Capacity & Peak Demand Forecast,” presentation to PJM Load Analysis Subcommittee (Nov. 2, 2026); PJM Independent Market Monitor, analysis of canceled/delayed data-center load (Sept. 21, 2026); prior Fracmoo coverage: “PJM Now Has Two Answers to the Same Question, and Both Run Into the Same Auction” (Nov. 4, 2026), “PJM Filed Its Backstop Fix. The Question Now Is Whether FERC Buys the New Way of Splitting the Bill.” (Oct. 31, 2026), and “December’s Auction Will Test What FERC’s Order Changed — and What It Didn’t” (Oct. 28, 2026).