$CEG $VST $TLN $NRG I don’t see how the merchant power names don’t trade up on this massive auction shortfall, but I’ve been proven wrong for the past 18 months on all the other indicators.
EXECUTIVE ASSESSMENT
The 2028/2029 PJM capacity auction confirms that the region remains structurally short dispatchable and reliably accredited capacity, supporting the medium-term cash-flow visibility and replacement value of incumbent generation portfolios. The more important investment conclusion, however, is not that capacity prices increased or that aggregate producer economics improved materially. The clearing price declined to $325/MW-day from $333.44/MW-day, total auction value remained approximately $16.4 billion, and almost all apparent supply improvement came from higher accreditation, reduced unoffered capacity, portfolio conversions, and previously existing assets rather than construction of substantial new generation. The auction therefore represents confirmation of persistent scarcity rather than a new upward earnings inflection for the entire independent power producer group.
The physical result was incrementally weaker than the headline comparison implies. PJM procured 138.318 GW of capacity through the Reliability Pricing Model, while Fixed Resource Requirement commitments contributed another 10.864 GW. Total committed capacity of 149.182 GW was 6.831 GW below the 156.013 GW reliability requirement. The deficit increased by approximately 315 MW, or 4.8 percent, from the prior auction’s 6.516 GW shortfall. The total reserve margin was 14.7 percent against a 20 percent target, equivalent to an 8.716 GW installed-capacity deficit relative to the target reserve margin. This does not imply that PJM will be unable to serve load during the delivery period. It indicates materially lower contingency protection, increased dependence on resource availability during extreme conditions, and a reduced margin for forecast error, forced outages, fuel constraints, transmission limitations, or weather-correlated renewable underperformance.
The $325/MW-day result was not an unconstrained competitive market-clearing price. It was the administratively imposed price ceiling. PJM’s static counterfactual estimates that the unconstrained Rest-of-RTO price would have been $554.72/MW-day and the COMED price would have been $776.69/MW-day. Those prices are 70.7 percent and 139.0 percent above the actual clearing price, respectively. The corresponding gross market value would have been approximately $29.7 billion, or $13.3 billion above the actual $16.4 billion result. Only 76.8 MW of additional capacity would have cleared in the static simulation, demonstrating an extremely inelastic near-term supply curve. The simulation is not a forecast because offers were held constant and higher prospective prices would alter bidding, retirement, investment, and demand-response behavior. It nevertheless establishes that the auction cap suppressed scarcity rents far more than it changed near-term procurement volume.
The earnings implications differ materially across Constellation, Talen, Vistra, and NRG. Based on disclosed cleared quantities, Talen has the strongest incremental gross-revenue read-through, Constellation has the greatest absolute capacity value and the largest exposure to latent COMED scarcity, Vistra largely preserves its prior elevated capacity run rate, and NRG’s outcome appears broadly consistent with the $325 assumption already embedded in its long-term framework. Constellation’s nuclear production tax credit exposure, Vistra’s and NRG’s retail load positions, portfolio hedges, bilateral contracts, and capacity-performance obligations prevent gross cleared value from translating directly into consolidated EBITDA. The Bloomberg Intelligence characterization of a uniformly positive earnings event is therefore too broad. The event is strongly positive for regional asset scarcity and earnings duration, but the near-term estimate revision differs substantially by company.
AUCTION ECONOMICS
At $325/MW-day, each 1 GW of cleared capacity produces $118.625 million of gross planning-year capacity value before hedges, contractual adjustments, performance settlements, taxes, and operating costs. The prior $333.44/MW-day price produced $122.039 million per GW because the 2027/2028 planning year included 366 days. The headline price declined by 2.53 percent, while planning-year gross value per cleared MW declined by 2.80 percent after accounting for the leap-day difference. Cleared RPM capacity increased by approximately 2.77 percent on the like-for-like PJM comparison, almost exactly offsetting the decline in per-MW value. This arithmetic explains why total auction value remained approximately unchanged at $16.4 billion despite a lower headline clearing price.
The $16.4 billion figure should not be interpreted as a direct incremental payment by retail customers or as aggregate generator EBITDA. A portion of load is self-supplied through the Fixed Resource Requirement construct, while other exposures are offset through bilateral transactions, retail hedges, full-requirements contracts, financial hedges, or common ownership of load and generation. PJM estimates that procuring sufficient capacity to meet the full reliability requirement at the capped price would have increased gross auction value by approximately $0.8 billion. The auction shortfall therefore reduced immediate capacity payments, but it also increased the probability that scarcity will be monetized through energy prices, ancillary-service prices, bilateral contracts, administrative backstop procurements, asset transactions, or future market-rule changes.
The clearing price declined because the reference resource’s unforced-capacity factor increased from 77 percent to 79 percent, mechanically reducing the price ceiling from $333.44/MW-day to $325/MW-day. The reduction did not result from lower demand, a material increase in newly constructed generation, or evidence of a looser reserve position. Net cost of new entry increased materially under PJM’s periodic review, including a 35.4 percent to 92.6 percent increase across the RTO and most modeled locational delivery areas, but the administrative collar prevented that increase from translating into the clearing price. The auction therefore delivered a misleading surface signal: the posted price fell while underlying replacement economics, reserve scarcity, and the estimated unconstrained price increased.
All modeled locational delivery areas cleared at the same $325/MW-day ceiling. This uniformity should not be interpreted as evidence that transmission constraints or locational scarcity disappeared. Broad RTO-wide scarcity caused the system price to reach the cap before local constraints could produce higher observable prices. PJM’s counterfactual indicates that COMED would have cleared at approximately $776.69/MW-day without the ceiling, materially above the $554.72/MW-day Rest-of-RTO estimate. COMED’s capacity emergency transfer limit also declined sharply because of generation deactivations, load growth, and external transmission-model effects. The administrative cap therefore concealed economically significant locational differentiation that remains relevant to asset valuation, bilateral contracting, backstop procurement, and future market redesign.
The market-power mitigation outcome reinforces the conclusion that the shortage is fundamentally physical and administrative rather than merely the result of opportunistic offer behavior. The RTO failed the 3-pivotal-supplier test, causing existing generation resources to be subject to market-power mitigation. Offers below applicable floors cleared, while offers above the ceiling did not. Concentration remains structurally important, but the capped result cannot be attributed solely to generators exercising unilateral pricing power through high offers. The combination of universal mitigation, a binding price ceiling, and a persistent procurement deficit indicates that the available accredited supply curve remained insufficient even after extensive administrative intervention.
THE SUPPLY SHORTFALL IS STRUCTURAL
Offered UCAP increased by 3.447 GW, from 136.148 GW to 139.595 GW, while cleared RPM capacity increased by approximately 3.733 GW on PJM’s like-for-like comparison. Only 317.1 MW of new generation and 207.6 MW of uprates cleared, for a combined 524.7 MW. That contribution was lower than the prior auction’s 774.3 MW and represented only 7.7 percent of the current 6.831 GW reliability deficit. Eligible installed capacity increased by only 1.294 GW, and offered installed capacity increased by approximately 528 MW. The gap between the large increase in cleared UCAP and the limited increase in physical installed capacity demonstrates that most of the apparent supply improvement reflected accreditation changes, portfolio reclassification, reduced withholding or nonparticipation, and improved recognition of existing capability rather than construction of new steel.
Required capacity that was not offered declined from approximately 809 MW UCAP to 193 MW UCAP, providing a meaningful but nonrepeatable improvement. Once most economically available existing capacity is participating, additional auctions cannot rely on another comparable reduction in unoffered capacity. Imports also declined, demand-response offers fell by approximately 224 MW of installed capacity, and no Price Responsive Demand or Order 2222 distributed-energy-resource offers were submitted. The absence of material demand-side and distributed-resource participation is particularly important because these categories should theoretically respond faster than greenfield generation to a high-capacity-price environment. Their limited contribution indicates that market rules, measurement requirements, customer economics, aggregation barriers, or implementation delays continue to constrain nontraditional supply.
The changing resource mix also supports a structural scarcity interpretation. Gas-fired accredited capacity increased by approximately 5.639 GW, largely because of accreditation changes, coal-to-gas conversions, and the return of previously nonparticipating capacity. Coal declined by approximately 2.941 GW, primarily because of retirements and conversions. Solar increased by approximately 651 MW, but its accredited contribution remained small relative to nameplate capacity and system need. Cleared supply was approximately 46 percent gas, 20 percent nuclear, 18 percent coal, 5 percent demand response, 4 percent hydro, 2 percent wind, 2 percent oil, and 1 percent solar. Approximately 84 percent of cleared capacity therefore remained gas, nuclear, or coal, emphasizing that PJM reliability continues to depend on dispatchable thermal and nuclear infrastructure despite rapid renewable development.
The delivery schedule is an additional constraint. The auction was conducted approximately 23 months before the June 2028 delivery start, substantially shorter than the traditional forward period of more than 3 years. That interval is insufficient for most large gas plants, nuclear uprates, transmission projects, or utility-scale generation developments to secure permits, interconnection agreements, equipment, financing, construction labor, and commercial operation. Even a theoretically adequate price signal cannot generate a rapid physical response when project lead times exceed the auction horizon. Existing facilities, advanced brownfield projects, uprates, delayed retirements, imports, demand response, and load reductions are consequently the only realistic sources of material incremental capacity before the next delivery periods.
The shortage is not solely a function of the latest annual increase in data-center load. PJM’s peak-load forecast increased by 1.375 GW, or approximately 0.8 percent, to 165.954 GW. The reliability requirement increased by a much larger 3.613 GW, or approximately 2.4 percent, to 156.013 GW. Approximately 1.273 GW, or 35.2 percent, of the requirement increase can be attributed mechanically to higher peak load at the prior forecast-pool requirement factor. Approximately 2.340 GW, or 64.8 percent, resulted from the increase in the forecast-pool requirement factor from 0.9260 to 0.9401, with the installed reserve margin remaining at 20 percent. Accreditation methodology and reliability modeling therefore contributed more to the year-over-year requirement increase than the latest load-forecast increment.
The broader multiyear demand trend remains dominated by data centers and other large loads. The Independent Market Monitor previously estimated that approximately 17.071 GW of existing and forecast data-center load increased 2027/2028 capacity-market revenues by approximately $6.498 billion, or 65.5 percent, relative to a scenario excluding that load. That prior-period estimate should not be mechanically applied to the current auction, but it demonstrates the cumulative scale of data-center demand already embedded in PJM’s load forecast and capacity economics. The relevant investment variable is not merely announced data-center capacity. It is the probability-adjusted, coincident, transmission-feasible, and contractually committed load incorporated into PJM’s reliability forecast. Forecast methodology and project attrition can therefore move capacity requirements by several GW even without an equivalent change in operating load.
RELIABILITY AND MARKET-DESIGN IMPLICATIONS
The 2028/2029 result represents the 2nd consecutive Base Residual Auction in which procurement was more than 1 percentage point below the installed reserve margin requirement. Under PJM’s tariff framework, a 3rd consecutive qualifying shortfall would require a Reliability Backstop Auction. The December 2026 auction for 2029/2030 therefore has significance beyond its direct clearing result. Another material deficit would demonstrate that the conventional annual auction cannot procure the target reserve margin and would activate a more interventionist reliability mechanism.
PJM has separately proposed a transitional Reliability Backstop Procurement intended to address the observed 2028/2029 shortage. The proposal contains a centrally administered procurement process and a facilitated bilateral matching mechanism. The central process is expected to begin in September 2026 and close in November 2026, subject to regulatory and implementation outcomes. Qualifying resources could receive 15-year capacity-only contracts beginning in 2028/2029, with pay-as-bid prices capped at $555/MW-day and settled as contracts for differences against future RPM clearing prices. Suppliers would retain accreditation and performance risk.
The coexistence of a $325/MW-day annual auction cap and a proposed $555/MW-day ceiling for 15-year procurement is the most consequential policy signal from the current market structure. It implicitly acknowledges that the administratively limited 1-year capacity price may be insufficient to finance new entry even though existing consumers are protected from the full short-run scarcity price. At $555/MW-day, 1 GW represents approximately $202.575 million of annual gross capacity value, versus $118.625 million at the regular auction cap. The proposed backstop thus creates a higher-priced, longer-duration administrative market for selected incremental projects while suppressing scarcity rents for the existing fleet in the general auction.
The backstop proposal is mixed for incumbent generators. Owners of permitted sites, existing interconnections, brownfield expansion opportunities, uprates, dormant units, repowering candidates, or advanced development projects could capture high-value contractual opportunities. Existing unmodified capacity would not automatically receive the same economics. Successful procurement would also add supply and reduce future auction scarcity, potentially diluting the medium-term capacity rents of the incumbent fleet. The proposal therefore transfers value from broad incumbent scarcity exposure toward project-specific development optionality. Companies with credible expansion pipelines and existing interconnection rights could benefit disproportionately.
Cost allocation remains unresolved and politically consequential. PJM has indicated that large-load customers, states, utilities, and consumers could face different allocations depending on the final framework. A mechanism that assigns most costs to incremental data-center load would preserve political support for procurement but could slow or redirect new-load development. Broad socialization across load would reduce the immediate economic burden on data centers but create higher consumer and regulatory risk. The ultimate allocation methodology could influence data-center siting, utility capital plans, state participation in PJM, retail rates, and the willingness of generators to commit capital.
EARNINGS TRANSMISSION
The delivery period runs from June 1, 2028 through May 31, 2029. At a constant daily rate, approximately 58.6 percent of planning-year capacity value will be recognized in calendar 2028 and approximately 41.4 percent in calendar 2029. The auction therefore has little direct effect on 2026 or 2027 reported earnings. Its immediate valuation significance is the reduction of uncertainty in outer-year cash-flow estimates, changes to forward hedging decisions, higher confidence in asset terminal values, and the signal provided for subsequent capacity, energy, ancillary-service, bilateral, and asset markets.
Mechanical cleared value should not be equated with incremental EBITDA. Recognized economics depend on ownership during the delivery period, bilateral capacity sales, retail supply obligations, hedge accounting, replacement-capacity purchases, capacity-performance bonuses and penalties, forced outages, fuel availability, operating costs, environmental compliance, tax credits, state support programs, and contractual pass-through provisions. A generation-only merchant portfolio captures capacity-price increases more directly than an integrated retail portfolio because the latter may experience offsetting increases in the cost of serving customer load. Nuclear portfolios can also experience reduced federal production tax credits when gross market receipts rise.
Capacity revenue nonetheless has high strategic value for an existing operating plant. It contributes toward fixed-cost recovery without requiring incremental dispatch, lowers the energy and ancillary-service margins required to remain economic, reduces retirement pressure, and can support maintenance and life-extension capital. A persistently high capacity price therefore raises the economic value of dispatchable assets even when the immediate year-over-year gross-revenue change is modest. The offset is increased policy risk: elevated consumer costs and repeated procurement deficits increase the probability of price caps, backstop contracting, accelerated transmission development, subsidized entry, revised accreditation, or other interventions that limit long-duration incumbent rents.
CONSTELLATION ENERGY
Constellation cleared 18,875 MW for 2028/2029, comprising 15,700 MW of nuclear capacity and 3,175 MW of fossil and other capacity. The geographic composition included approximately 10,200 MW in COMED, 6,550 MW in EMAAC, 1,725 MW in MAAC, 375 MW in BGE, and 25 MW in the Rest-of-RTO category. All plants offered by Constellation cleared, excluding assets classified as held for sale. The cleared portfolio increased by 925 MW, or 5.2 percent, from 17,950 MW in the prior auction.
Mechanical gross planning-year capacity value is approximately $2.239 billion, compared with approximately $2.191 billion for the prior planning year. The increase is approximately $48.4 million, or 2.2 percent. The higher cleared volume more than offsets the 2.8 percent reduction in gross value per MW. Constellation represents approximately 13.6 percent of RPM-cleared capacity and has the largest disclosed absolute capacity-revenue exposure among the analyzed companies.
The quality of the incremental earnings contribution is more complicated than the gross-value calculation suggests. Approximately 83.2 percent of Constellation’s cleared capacity is nuclear. Nuclear capacity revenue is included in gross receipts for purposes of the federal 45U nuclear production tax credit, which remains available through 2032 and phases down as market receipts increase. Constellation reported that estimated 2026 gross receipts exceeded the phaseout level for most units, resulting in no material 45U benefit during the initial 2026 period. Where a unit is already fully phased out, incremental capacity revenue can translate more directly into earnings. Where a unit remains inside the phaseout range, part of the higher capacity receipt may be offset by a lower tax credit. Certain state-support mechanisms can also require refunds or pass-throughs.
At an illustrative 93 percent nuclear capacity factor, the $118,625 annual capacity payment per MW is economically equivalent to approximately $14.55/MWh of gross receipts. Capacity revenue is therefore a meaningful component of the nuclear tax-credit calculation rather than a peripheral adjustment. The exact offset is unit-specific and depends on realized energy, ancillary-service, capacity, hedge, and contract receipts. Constellation’s gross capacity value should consequently be viewed as a measure of market support for the fleet, not as a direct EBITDA forecast.
The fossil and other portfolio increased by approximately 750 MW, while nuclear cleared capacity increased by approximately 175 MW. The fossil increase is likely to produce a more direct capacity-to-earnings conversion because it is not subject to the nuclear production tax credit, although portfolio hedging, customer-load positions, operating costs, and asset ownership changes remain relevant. Constellation’s completed acquisition of a large competitive generation and customer platform also complicates the comparison with the prior auction because part of the cleared-volume increase may reflect portfolio composition rather than organic accreditation improvement.
Constellation has the greatest disclosed exposure to the hidden COMED scarcity premium. Its approximately 10.2 GW COMED position represents more than 50 percent of its cleared PJM portfolio. Applying PJM’s static $776.69/MW-day counterfactual rather than the $325 cap would imply approximately $1.682 billion of additional annual gross value for that COMED capacity alone. This is not a realizable earnings forecast because the counterfactual holds offers and behavior constant, and nuclear tax-credit offsets would reduce the net benefit. It nevertheless quantifies the magnitude of scarcity value being suppressed by the administrative ceiling and highlights Constellation’s sensitivity to future COMED market design, bilateral contracting, transmission constraints, and backstop procurement.
The investment interpretation is therefore 2-sided. Constellation has the strongest absolute cash-flow visibility, the largest nuclear fixed-cost support, substantial COMED scarcity optionality, and a portfolio that should benefit from prolonged regional tightness. The immediate year-over-year gross-revenue increase is only approximately $48 million, and tax-credit interactions weaken 1-for-1 earnings conversion. The auction is more important for fleet value, retirement economics, contract negotiations, and terminal cash-flow duration than for a large near-term earnings revision.
