{"document":{"slug":"filing-20260925-5208","title":"Answer and Motion for Leave to Answer of the Independent Market Monitor for PJM Interconnection, L.L.C. to the Answer of PJM Interconnection, L.L.C. filed on 09/10/2026","issued":"2026-09-25","accession":"20260925-5208","docket":"ER26-3380-000","group":"PJM filings","filer":"Monitoring Analytics, LLC","kind":"answer","file":"sources/originals/20260925-5208_01.pdf","url":"https://elibrary.ferc.gov/eLibrary/filelist?accession_Number=20260925-5208","sha256":"88400af045854979c5c49d3369c43e3c159d22db2df918b782fc5d8600a5d4ee","pages":33,"tokens":11223,"emptyPages":[]},"pages":[{"page":1,"startToken":0,"endToken":253,"text":"UNITED STATES OF AMERICA \nBEFORE THE \nFEDERAL ENERGY REGULATORY COMMISSION \n \nPJM Interconnection, L.L.C. \n) \n) \n) \n \nDocket No. ER26-3380-000 \nANSWER AND MOTION FOR LEAVE TO ANSWER \nOF THE INDEPENDENT MARKET MONITOR FOR PJM \nPursuant to Rules 212 and 213 of the Commission’s Rules and Regulations,1 \nMonitoring Analytics, LLC, acting in its capacity as the Independent Market Monitor \n(“Market Monitor”) for PJM Interconnection, L.L.C. (“PJM”),2 submits this answer to, and \nmoves for leave to answer, the answer submitted by PJM on September 10, 2026 (“September \n10th Answer”). \nThe proceeding concerns the filing submitted by PJM Interconnection, L.L.C. (“PJM”) \non July 31, 2026, proposing (at 1) “to establish the Reliability Backstop Procurement \n(“RBP”)—a one-time procurement of new capacity to address resource adequacy challenges \nin the PJM Region arising from insufficient commitments of capacity in the 2027/2028 and \n2028/2029 Reliability Pricing Model (“RPM”) Auctions, primarily driven by the \nunprecedented addition of large loads, most notably data centers” (“RBP Filing”). \nDespite the clearly stated public policy goals of elected officials, PJM’s stated purpose \nis not to assign the costs and risks of data centers to data centers. PJM’s stated purpose is not \nto protect other PJM customers from paying the costs and bearing the risks caused by data \n \n1 \n18 CFR §§ 385.212 & 385.213 (2025). \n2 \nCapitalized terms used herein and not otherwise defined have the meaning used in the PJM Open \nAccess Transmission Tariff (“OATT”), the PJM Operating Agreement (“OA”) or the PJM Reliability \nAssurance Agreement (“RAA”)."},{"page":2,"startToken":253,"endToken":638,"text":"- 2 - \ncenters. The shortfall identified in the RBP Filing is a result solely of data center demand.3 \nPJM’s approach will continue to impose data center costs on other customers. PJM fails to \ncorrectly identify the source of the capacity market and reliability problem. PJM fails to \ncorrectly identify the appropriate solution to the problem. As a result of the incorrect framing \nof the problem, the details of the RBP Filing are largely irrelevant. The details are about \nimplementing a design that cannot achieve the correct objective of ensuring that data centers \nbear their own costs. The significant issues with the design details further reveal the flaws \nwithin PJM’s approach. \nPJM actually denies the fact that the shortfall faced by PJM is the result of data center \nload. PJM states (September 10th Answer at 64): “It is fallacious to suggest that a just and \nreasonable cost allocation approach must either identify specific loads (which are not FPA-\njurisdictional actors) and not the LSEs serving those loads or be limited to one type of load \nwhen the resource adequacy shortfall is caused by all types of loads that comprise the subject \nLoad Adjustments” [emphasis in original].) \nThe underlying theme and narrative of PJM’s RBP and IRAS Filings is that data center \nload is just load and the normal PJM capacity market and energy market mechanisms can \nhandle the impact of data center load. PJM even goes a step beyond and incorrectly asserts \nthat the identified shortfall is a result of all types of loads. In a word, PJM asserts that it is just \nsupply and demand.4 PJM fails to recognize that the massive influx of data center load \nrequires a paradigm shift in thinking about capacity. PJM fails to recognize that the massive \ninflux of data centers requires a policy decision. That policy decision cannot reasonably fall \nback on simple high school economic nostrums. The policy instruction from a wide range of \nsources, including the elected representatives of the 67 million residents in PJM’s market area, \n \n3 \nSee Monitoring Analytics, LLC “Analysis of the 2027/2028 RPM Base Residual Auction – Part A,” \n(January 5, 2026). \n4 \nSee Proposal of PJM Interconnection, L.L.C. to Establish Interim Resource Adequacy Service and a \nLarge Load Registry to Address System Reliability, Docket No. ER26-3515-000 (2026) (“IRAS Filing”)."},{"page":3,"startToken":638,"endToken":955,"text":"- 3 - \nis clear. Data centers should pay for their own costs. A significant part of data center costs is \nthe cost of the capacity required to serve them. PJM’s proposals fail to conform to that policy \ninstruction. \nPJM attempts to use its current capacity cost allocation mechanism for the obligations \nthat result from the RBP with EDCs/LSEs at the center and in which EDCs/LSEs receive an \nallocation of capacity costs from PJM and then allocate those costs to customers on a peak \nload basis. The current cost allocation process is clearly not adequate to address the \nassignment of data center costs to data centers. Rather than directly addressing the problem, \nPJM’s proposal would push the issue to the states by suggesting that states can create \nallocation rules that the EDCs/LSEs must follow. \nThe September 10th Answer fails to address these underlying issues primarily by \nignoring the real policy issue and focusing on narrow technical issues of how to implement \nthe misguided RBP approach. \nOn August 21, 2026, the Market Monitor submitted a protest demonstrating that the \nRBP Filing has not been supported as just and reasonable. The Market Monitor’s protest also \ndemonstrated that the RBP Filing is unjust and unreasonable, and that attempting to \nimplement it will severely damage PJM markets, impose data center costs on other customers \nas an intended consequence, create unintended consequences, and compound that damage \nfor a period of 15 years. The Market Monitor concluded that, for all these reasons, the RBP \nFiling should be rejected. \nThe RBP Filing and PJM’s September 10th Answer would continue to impose the costs \nof data center development on other customers. The RBP Filing and PJM’s September 10th \nAnswer would continue to add to the $29.4 billion of capacity market costs that have resulted \nfrom the inclusion of data centers and PJM’s inaction to date by leaving 20,539.7 MW of data"},{"page":4,"startToken":955,"endToken":1323,"text":"- 4 - \ncenter load in the 2029/2030 BRA to be run in December 2026.5 6 While appearing to recognize \nthat data center load is the source of the reliability issues facing PJM, PJM’s actual proposal \nwould impose the costs and risks of the proposed RBP solution on other customers. The \nshortfall identified by PJM is entirely the result of existing and forecast data center load, yet \nPJM has failed to verify the underlying details of the forecast and refuses to reduce the \nforecast, and therefore the shortfall, based on known changes to data center plans. The costs \nof addressing the shortfall within the RBP approach are inflated as a result of PJM’s failure \nto include any market power mitigation rules, failure to use a single clearing price auction \nand the creation of crushing collateral requirements for EDCs/LSEs and their customers. \nThe September 10th Answer entirely fails to address the most fundamental issues \nraised in the Market Monitor’s protest. The central error in the RBP Filing was not addressed \nin the September 10th Answer. PJM assigns the responsibility for the identified shortfall to \nindividual EDCs that have data centers in their areas and to their customers despite the \nirrefutable fact that the RBP shortfall is entirely the result of data center load. PJM’s approach \ndirectly contradicts the principle that data center costs should be paid by data centers and \nnot by other customers. PJM does not explain why it is just and reasonable to assign the costs \nof data center loads to EDCs/LSEs. PJM proposes to penalize EDCs and their customers for \nthe locational decisions of data centers. As a result of the fact that PJM is avoiding taking \nresponsibility for the final allocation of data center costs by EDCs, it is impossible to know \nwhat the final outcome of PJM’s RBP Filing will be. \n \n5 \nTotal data center load in 2029/2030 BRA (27,371.0 MW) minus PJM’s shortfall (6,831.3). See PJM. Total \nLoad \nAdjustments \nBreakdown \n. \n6 \nSee Monitoring Analytics, LLC, 2026 Quarterly State of the Market Report for PJM: January through June, \nSection 5: Capacity Market (August 13, 2026)."},{"page":5,"startToken":1323,"endToken":1656,"text":"- 5 - \nPJM has the authority to require LSEs to condition wholesale service to data centers \non a requirement that data centers enter into bilateral contracts with generators that protect \nother customers from the data center costs.7 \nEven if PJM were to allocate costs to other customers, PJM has the authority to allocate \nthe RBP costs to all PJM customers consistent with the allocation of backstop auction costs in \nthe current tariff.8 \nPJM asserts that it is not really allocating costs to the EDCs and their customers by \nasserting that states can require the EDCs to allocate those costs to data centers. PJM is \nattempting to evade its responsibility to directly ensure that data centers pay their own costs \nby shifting that responsibility to the states and the EDCs. PJM denies responsibility for \nallocating data center costs to data centers (at 56–58) by stating that this is not PJM’s job but \nis the job of the states, based on an incorrect understanding of the FERC/PJM jurisdictional \nauthority. Despite all that, PJM does include a specific allocation of RBP data center costs to \nLSEs (EDCs here) in the event that no alternative is created by individual states and \nEDCs/LSEs are expected to allocate those costs to their customers per the method currently \nused to allocate capacity costs.9 10 \nPJM focuses on the defined shortfall while ignoring the much larger issue of total \nexisting and forecast data center load. On a UCAP basis, the total data center load included \nin the 2028/2029 BRA was 20,183.9 MW which is approximately three times the shortfall \nidentified by PJM. The September 10th Answer does not address this fundamental issue, \n \n7 \nSee Protest of the Independent Market Monitor for PJM, Docket No. ER26-3380-000 (August 21, 20261, \n2026) at 28–32. \n8 \nSee PJM OATT Attachment DD § 16.4(e)(ii). \n9 \nSee RBP Filing, Attachments (A&B) Proposed OATT Attachment DD § 18.3(b)(iii). \n10 \nSee RAA Schedule 8 § D."},{"page":6,"startToken":1656,"endToken":1999,"text":"- 6 - \nincluding the fact that data center load will continue to result in a significant increase in \ncapacity market costs for all customers in the upcoming 2029/2030 BRA in December 2026. \nThe Market Monitor identified significant implementation issues in the RBP Filing \nthat the September 10th Answer fails to address. These issues, including the explicit failure to \naddress the exercise of market power, and the bafflingly complex implementation by PJM, \nare all created by PJM’s misguided attempts to blame EDCs and their customers for the \naddition of data center load rather than simply assigning the costs and risks of data center \nload to data centers. For all these reasons, the RBP Filing is unjust and unreasonable and \nshould be rejected. \nThe Market Monitor recognizes the urgency of the situation facing the PJM markets. \nThe Market Monitor continues to recommend that the Commission reject the RBP Filing and \ndirect PJM, under the Commission’s FPA Section 206 authority, to file, within two weeks, a \nsolution with the following elements: \n• \nPJM should, working with LSEs and EDCs, immediately identify data center load \nnot interconnected by June 1, 2026, and remove it from RPM capacity market \nauctions. \n• \nPJM should establish criteria to be included in the RAA defining the obligations of \nLSEs to require that new data center loads have 15 year bilateral contracts with \nnew generation that do not permit any cost or risk shifting to other customers. The \ncontracts would be based on general terms and conditions in accordance with \ntariff defined criteria with the cost and risk specifics defined by the data centers \nand generators. \n• \nFor LSEs with data centers unable to make independent arrangements meeting the \ncriteria, PJM should run a full single clearing price auction using the BRA capacity \nmarket clearing model for new data center load. LSEs would require data centers \nto participate in the reliability backstop auction if they do not have a bilateral \ncontract for new generation that meets the requirements of such contracts. When \nthe clearing process is complete, the LSEs would require that generators and data"},{"page":7,"startToken":1999,"endToken":2203,"text":"- 7 - \ncenter loads enter into 15 year bilateral contracts that do not permit any cost or \nrisk shifting to other customers. The contracts would be based on general terms \nand conditions in accordance with tariff defined criteria with the cost and risk \nspecifics defined by the data centers and generators. \n• \nPJM should establish curtailment obligations for wholesale service to data centers \nduring a defined transition period that would allow all data centers the time to \ncontract for new generation. \nThese requirements meet the goals of the Pledge and Principles and can be \nimplemented immediately. \nI. ANSWER \nA. PJM’s Reliance on “Cooperative Federalism” Is Misguided and Has Resulted in \na Flawed and Incomplete Proposal. \nThe September 10th Answer defends (at 56–65) what it alleges is a “cost allocation \napproach” in the RBP Filing with references to “cooperative federalism.” The RBP Filing does \nnot propose a cost allocation approach, it proposes to procure the 6,831.3 MW shortfall of \ncapacity through a flawed process that, by default, imposes the costs and risks of meeting \ndata center load on other customers. PJM hopes that the states will come up with additional \nrules for allocating costs consistent with the broad consensus represented by the Pledge and"},{"page":8,"startToken":2203,"endToken":2699,"text":"- 8 - \nPrinciples.11 12 PJM relies on speculation about what states may or may not do to complete its \nproposal. The RBP Filing should be rejected as incomplete because it does not clearly state \nhow the costs and risks of meeting data center load must be assigned to customers. PJM failed \nto file a complete regional solution consistent with its role as an RTO. PJM’s balkanized \napproach is inefficient and ineffective at creating consistent regional markets. The \nCommission has the jurisdiction needed to approve a complete proposal. \n \n11 \nSee The White House, “Ratepayer Protection Pledge” (March 4, 2026), , \nProclamation No. 11014, 91 Fed. Reg. 11439 (Mar. 4, 2026) (“Today, pursuant to the Ratepayer \nProtection Pledge, leading United States hyperscalers and AI companies guarantee that data centers’ \nenergy needs will not increase household electricity costs for American citizens. Instead, these \ncompanies will build, bring, or buy the new generation resources and electricity needed to satisfy \ntheir energy demands, and pay for all new power delivery infrastructure upgrades to service their \ndata centers. They will voluntarily negotiate new, separate rate structures with their utilities and \nrelevant State governments, and pay those rates, and for their infrastructure, whether they use the \nelectricity or not. They will also invest in local communities and coordinate with grid operators to \ncontribute to a more reliable grid.”) (“Pledge”). On July 23, 2026, over 200 additional entities joined \nthe pledge. See The White House, “President Trump’s Ratepayer Protection Pledge Secures American \nAI Dominance, Protects Consumers” (July 23, 2026), . \n12 \nNational Energy Dominance Council, Statement of Principles Regarding PJM (January 16, 2026) \n \n(“Principles”). \nTwo \nprinciples apply to the contemplated procurement of new generation. First, the procurement should \n“[p]rovide 15-year price certainty for new capacity resources” [emphasis]. The statement notes that \nthis can be done in through a “Reliability Backstop Auction,” but does not require that this auction \nbe a centralized procurement by PJM or that PJM customers assume risks. Second, “PJM should \nallocate the cost of any new capacity through the aforementioned Reliability Backstop Auction to \nload serving entities (LSEs) with new data centers that have not self-procured new capacity or agreed \nto be curtailable.” The principles, agreed to by the Governors of PJM States Indiana, Ohio, Maryland \nPennsylvania, Virginia, West Virginia, Delaware, Illinois, Michigan, New Jersey, Tennessee, North \nCarolina and Kentucky, call for PJM to do the allocation and to match data centers to the costs not \nexcluding risks of new resources or commitments to curtail. Additional principles that going beyond \nthe RBF Proposal, the Market Proposal, or potential alternatives, include: (i) extending the price collar \nfor two additional RPM Base Residual Auctions, (ii) PJM improving load forecasting, (iii) accelerating \ngenerator interconnections, i.e., the Reliability Resource Initiative (RRI), and (iv) continued efforts to \nreform the capacity market."},{"page":9,"startToken":2699,"endToken":3094,"text":"- 9 - \nPJM should ensure that the costs and risks of acquiring capacity to meet all data center \nload, and not just the identified shortfall, are allocated to the data centers creating those costs \nand risks in the PJM Region, consistent with the Pledge and Principles. Data centers and the \nowners of new generation built to serve them are best positioned to understand and manage \nthe risks associated with building data centers and securing their power supply. It does not \nserve the public interest to assign risk to customers who have no understanding and ability \nto manage the associated risks. What states may or not do is unknown, but it cannot be \nassumed the state regulators are better able to understand and manage these risks than data \ncenters and generators. The design flaws in the RBP Filing are required by PJM’s failure to \nassign risks to those that create them. \nPJM reveals that it rejects the policy determination that underlies the Pledge and \nPrinciples.13 Although it attempts to hide behind an incorrect jurisdictional analysis, PJM \nreveals here that it does not intend the RBP Filing to actually address the issue that it initially \nidentified as the reason for submitting the RBP Filing.14 Failing to recognize the extraordinary \nnature of data center demands treats existing customers unjustly and unreasonably. PJM \ncannot show that its proposal is just and reasonable while failing to distinguish between new \ndata center load and existing load. \nPJM’s attempts to justify its incomplete RBP Filing based on an improper framing of \nthe Commission’s jurisdiction. States have jurisdiction over generation, and over retail sales \n \n13 \nCf. September 10th Answer at 64 (“It is fallacious to suggest that a just and reasonable cost allocation \napproach must either identify specific loads (which are not FPA-jurisdictional actors) and not the \nLSEs serving those loads or be limited to one type of load when the resource adequacy shortfall is \ncaused by all types of loads that comprise the subject Load Adjustments” [emphasis in original].); \nRBP Filing at 1 (PJM proposes “to establish the Reliability Backstop Procurement (“RBP”)—a one-\ntime procurement of new capacity to address resource adequacy challenges in the PJM Region arising \nfrom insufficient commitments of capacity in the 2027/2028 and 2028/2029 Reliability Pricing Model \n(“RPM”) Auctions, primarily driven by the unprecedented addition of large loads, most notably data \ncenters”). \n14 \nId."},{"page":10,"startToken":3094,"endToken":3528,"text":"- 10 - \nin their own states. States do not have jurisdiction over neighboring states. PJM does not \nshow that states have the capability to perform the allocation consistent with the Pledge and \nPrinciples. States cooperating with (or failing to cooperate with) other states to regulate \ninterstate commerce, displacing Federal regulation, is not cooperative federalism. PJM \nshould not attempt to transfer its responsibility to operate the regional capacity market to the \nstates. The Supreme Court has recognized that the Commission’s jurisdiction extends to rules \nrelated to retail customers’ participation in the wholesale power markets as sellers.15 The \nSupreme Court recognizes Federal jurisdiction over the terms and conditions of retail \ncustomers’ participation in wholesale markets even as they remain retail customers. \nPJM apparently believes that because data centers are retail customers, PJM cannot \nrequire that the LSEs who serve them place terms and conditions on service to data centers.16 \nPJM has not reconciled this position with the jurisdiction exercised under the Reliability \nAssurance Agreement (“RAA”). The RAA requires LSEs to meet their obligations from \ncapacity provided by generators that meet the terms and conditions of the RAA.17 LSEs are \nnot permitted under the RAA to rely on Generation Capacity Resources that do not meet \n \n15 \nSee FERC v. EPSA, 577 U.S. 260, 281 (2016) (EPSA) (“Yet a FERC regulation does not run afoul of \n§824(b)’s proscription just because it affects—even substantially—the quantity or terms of retail sales. \nIt is a fact of economic life that the wholesale and retail markets in electricity, as in every other known \nproduct, are not hermetically sealed from each other. To the contrary, transactions that occur on the \nwholesale market have natural consequences at the retail level. And so too, of necessity, will FERC’s \nregulation of those wholesale matters. Cf. Oneok, Inc. v. Learjet, Inc., 575 U. S. 373, 387… (2015) (noting \nthat in the similarly structured world of natural gas regulation, a “Platonic ideal” of strict separation \nbetween federal and state realms cannot exist). When FERC sets a wholesale rate, when it changes \nwholesale market rules, when it allocates electricity as between wholesale purchasers—in short, \nwhen it takes virtually any action respecting wholesale transactions—it has some effect, in either the \nshort or the long term, on retail rates. That is of no legal consequence.”). \n16 \nSee RBP Filing at 64 (“It is fallacious to suggest that a just and reasonable cost allocation approach \nmust … identify specific loads (which are not FPA-jurisdictional actors) …”). \n17 \nSee RAA Schedules 9, et seq. & 10 (rules for PJM to evaluate capability and deliverability of \nGeneration Resources)."},{"page":11,"startToken":3528,"endToken":3900,"text":"- 11 - \nthese terms and conditions regardless of state views about the capability of the facilities.18 \nResources included in Fixed Resource Requirement (“FRR”) Plans under Schedule 8.1 to the \nRAA must also meet requirements specified in the RAA.19 \nPJM does not explain why it cannot propose terms and conditions related to data \ncenters through the LSEs that serve them when it can and does include terms and conditions \nthrough LSEs (and other intermediate entities such as Curtailment Service Providers) on \nother retail customers and on generation facilities. The reliability backstop procurement \nwould be a part of PJM’s rules designed to ensure resource adequacy through the capacity \nmarkets and related rules. The capacity market rules are terms and conditions of wholesale \nenergy sales. \nThe case for jurisdiction is stronger in this case than it was in EPSA. This case concerns \nthe procurement of capacity under proposed capacity market rules. The RBP Filing is a \nmodified version of the existing rules for reliability backstop auctions.20 The purpose of the \nRBP is to address the particular and extraordinary demands related to resource adequacy \nattributable to the rapid interconnection of data centers in the PJM Region. The Supreme \nCourt has recognized PJM and federal responsibility in this area to the point that such \nresponsibility preempts state action.21 Data centers are receiving wholesale sales along with \nthe associated terms and conditions of such sales from PJM through Load Serving Entities. \n \n18 \nSee id. \n19 \nSee RAA § 7.2. \n20 \nSee RBP Filing at 28 (“’[B]ackstops to . . . market design based solutions that are required to insure \nreliability’ are nothing new to PJM.[citation omitted] In fact, the RBP bears several significant \nsimilarities to the existing Reliability Backstop Auction provisions set forth in Tariff, Attachment DD, \nsection 16.”). \n21 \nSee Hughes v. Talen Energy Mktg., LLC, 578 U.S. 150, 163 (2016) (“We agree with the Fourth Circuit’s \njudgment that Maryland’s program sets an interstate wholesale rate, contravening the FPA’s division \nof authority between state and federal regulators. As earlier recounted, … the FPA allocates to FERC \nexclusive jurisdiction over ’rates and charges . . . received . . . for or in connection with’ interstate \nwholesale sales. §824d(a). Exercising this authority, FERC has approved the PJM capacity auction as"},{"page":12,"startToken":3900,"endToken":4195,"text":"- 12 - \nPJM allocates costs to EDCs based on the level of data center load in their areas that \nthe EDCs have provided to PJM through the large load adjustment process.22 While PJM \nasserts that it does not have jurisdiction over the allocation of costs to customers by the EDC, \nPJM does include rules defining that allocation to EDCs that would apply in the event that \nthe states do not develop allocation rules.23 EDCs/LSEs are expected to allocate those costs to \ntheir customers per the method currently used to allocate capacity costs.24 \nPJM denies responsibility for allocating data center costs to retail loads while making \ncritical decisions about that allocation. The RBP Filing would allocate the costs of meeting \ndata center load to EDCs rather than to data centers. \nPJM’s attempt to justify its incomplete proposal based on an incorrect jurisdictional \nanalysis and to instead rely on speculation about actions that states may take is insufficient \nto support the RBP Filing as just and reasonable. PJM’s approach is unjust and unreasonable. \nThe RBP Filing should be rejected for that reason. \nB. The September 10th Answer Fails to Address the Issues Raised by the Market \nMonitor and Others. \n1. The Shortfall is Entirely the Result of Data Center Load. \nIt is a matter of fact that the apparently very precise 6,831.3 MW of shortfall is the \nresult solely of data center load and that the shortfall MW are not correct. \n \n \nthe sole rate setting mechanism for sales of capacity to PJM, and has deemed the clearing price per \nse just and reasonable”). \n22 \nSee RBP Filing, Attachment A, Proposed OATT Attachment DD § 18.3(a). \n23 \nSee id., Proposed OATT Attachment DD § 18.3(b). \n24 \nSee RAA Schedule 8 § D."},{"page":13,"startToken":4195,"endToken":4393,"text":"- 13 - \nPJM set the RTO-wide RBP procurement target as the shortfall observed in the July \n2026 RPM Base Residual Auction (“BRA”) for the 2028/2029 Delivery Year.25 PJM defined this \nshortfall as the difference between the total cleared capacity (in UCAP MW) and the RTO \nRPM reliability requirement for that auction. Table 1 shows PJM’s calculation of the shortfall \nthat will be used to set the RBP procurement target. PJM’s peak load forecast for the 2028/2029 \nDelivery Year, which is based on PJM’s 2026 Load Forecast Report, included 21,470.0 MW \n(ICAP) of data center load. The total organic load in PJM’s peak load forecast was only \n144,483.5 ICAP MW for the 2028/2029 Delivery Year. The data center load’s contribution to \nPJM’s reliability requirement (20,183.9 UCAP MW) was approximately three times (2.95 \ntimes) the shortfall (6,831.3 UCAP MW) in the 2028/2029 BRA run in July 2026. There would \nhave been no shortfall in the RPM BRA if not for the large data center load included in the \npeak load forecast. \n \n25 \nSee RBP Filing, Attachment A (Marked Tariff Records), Proposed OATT Attachment DD § 18.1(a)."},{"page":14,"startToken":4393,"endToken":4731,"text":"- 14 - \nTable 1 Shortfall and Data Center Load in the 2028/2029 RPM BRA26 \n \n2. The Shortfall MW Value Is Not Correct and PJM Refuses to Correct It. \nIt is a matter of fact that PJM’s MW shortfall target is not correct. PJM provided no \nactual response to this point in the September 10th Answer other than to state that it is ok to \nbe incorrect. PJM plans to reduce the shortfall for opt out MW but refuses to reduce the \nshortfall based on reductions in identified data center loads. \nThe shortfall procurement target is based on the 2028/2029 BRA held in July 2026. PJM \nset the procurement target to the deficit calculated as the difference between the RTO \nreliability requirement and the total cleared MW in the 2028/2029 BRA. In the 2028/2029 BRA, \nthe deficit was 6,831.3 MW. The data center load forecast for the 2028/2029 BRA was based \n \n26 \nFor this calculation, it is assumed that the FRR load does not include data center load. \n2028/2029 RPM BRA\nFormula\nOrganic Load ICAP (MW)\n144,479.5\n \nA\nData Center Load ICAP (MW)\n21,474.0\n \nB\nForecast peak load ICAP (MW)\n165,953.5\n \nC=A+B\nFRR peak load ICAP (MW)\n11,556.0\n \nD\nPRD ICAP (MW)\n0.0\nE\nInstalled reserve margin (IRM)\n20.0%\nF\nPool-Wide Accredited UCAP Factor\n78.34%\nG\nForecast pool requirement (FPR)\n0.9401\nH=(1+F)*G\nCleared UCAP (generation and DR)\n138,317.8\n \nI\nCleared ICAP (generation and DR)\n176,560.9\n \nJ=I/G\nRPM peak load ICAP (MW)\n154,397.5\n \nK=C-D-E\nRPM peak load UCAP (MW)\n120,955.0\n \nL=K*G\nOrganic reliability requirement UCAP (MW)\n135,825.2\n \nM=A*H\nData Center reliability requirement UCAP (MW)\n20,187.7\n \nN=B*H\nFRR Obligation UCAP (MW)\n10,863.8\n \nO=D*H\nRPM reliability requirement UCAP (MW)\n145,149.1\n \nP=M+N-O\nShortfall UCAP (MW)\n6,831.3\n \nQ=P-I"},{"page":15,"startToken":4731,"endToken":5125,"text":"- 15 - \non large load adjustments received by PJM before September 5, 2025.27 Since then, several \ndata center projects have been cancelled. The Market Monitor calculated that 4,016.9 MW or \n42 percent of the forecasted increase in the data center load between the 2026/2027 Delivery \nYear and 2028/2029 Delivery Year have been cancelled or delayed as of September 21, 2026.28 \n29 Of the 4,016.9 MW of cancelled or delayed data center load, the Dominion EDC accounted \nfor 51 percent and the AEP-Ohio EDC accounted for 29 percent. There are no provisions in \nthe RBP Filing to revise the data center load forecast and the target procurement prior to the \nRBP procurement process. \nThe assertion in the September 10th Answer (at 18) that the RBP procurement target of \n6,831.3 MW is “deliberately conservative” and “allowing procurement of capacity beyond \nthe initial 2028/2029 Delivery Year decreases the risk of over procurement” is not supported. \nPJM arrived at the 6,831.3 UCAP MW as the shortfall in 2028/2029 Base Residual Auction that \nused the load forecast that included EDC provided large load adjustments, which were based \non inputs from data center developers (see Table 1). These estimates are understood to \ninclude speculative load and have been significantly revised in the past and are likely to be \n \n27 \nSee PJM Presentation at Load Analysis Subcommittee, “Large Load Adjustment Discussion,” \n (October 28,2025). \n28 \nPJM included 11,993 MW of data center load in the 2026/2027 BRA and 21,474 MW in the 2028/2029 \nBRA (see Table 2). The cancelled or delayed data center load of 4,016.9 MW as of September 21, 2026, \nrepresents 42 percent of the 9,481 MW increase (21,474 MW minus 11,993 MW) between the two base \nresidual auctions. \n29 \nThe Market Monitor calculation is based on tracking of datacenter project status by Industrial Info \nResources (IIR) corroborated by publicly available data . The \nMarket Monitor does not have access to project level data behind large load adjustments requested \nby EDCs. The Market Monitor considered only cancelled and delayed data center projects that were \noriginally scheduled to be in service by 2029 as indicated by the IIR tracking data."},{"page":16,"startToken":5125,"endToken":5487,"text":"- 16 - \nrevised.30 PJM never did a detailed evaluation of whether each data center was likely to be \nbuilt, leaving the details to the EDCs. \nThe September 10th Answer argues (at 22) that the Commission does not require the \nRBP procurement target to be continuously adjusted to reflect updated information. PJM cites \nthe Commission’s decision in which the Commission accepted information in certain contexts \nbased on a perceived need for balance.31 Neither case involved condoning reliance on \ninaccurate or stale information over available accurate or fresh information. The argument \nthat PJM seeks to evade is that PJM must support its 6,831.3 MW target, inherently a precise \nnumber, based on accurate and contemporaneous information. The cases cited by PJM \ninvolved acceptance of proposals intended to achieve balanced results and of rules \npotentially remaining in place indefinitely.32 PJM has not explained how reliance on stale and \ninaccurate information would contribute to a balanced procurement or any other reason. The \nRBP Filing is for a one time special reliability backstop auction in response to an emergency \nshortfall. PJM is not entitled to propose any number that it can conjure using inaccurate and \nstale information. The one time auction will place significant obligations on PJM customers \nfor 15 years. Under these circumstances, PJM should recognize its obligation to procure the \n \n30 \nAccording to a recent report by Capgemini Research Institute, “Around one in five data center power \nrequests \nmay \nnever \nmaterialize, \ndistorting \nforecasts,” \n (June 25, \n2026). \n31 \nId. at 22, citing PJM Interconnection, L.L.C., 128 FERC ¶ 61,157, at P 77 (2009); Midcontinent Indep. Sys. \nOperator, Inc., 189 FERC ¶ 61,108 (2024). \n32 \nSee PJM,128 FERC ¶ 61,157 at P 77 (rejected argument opposing market design proposal because an \nalternative approach, that was not proposed, that could have addressed alleged “bias in estimation”); \nMISO, 189 FERC ¶ 61,108 at 64 (Approving RTO MW Value based on statement: “[W]hile an 85% \nThreshold MW Value does not eliminate these risks, they believe that it strikes the best balance \nbetween mitigating oversubscription and undersubscription risk.”)."},{"page":17,"startToken":5487,"endToken":5902,"text":"- 17 - \namount needed based on reliable evidence showing what is needed. No such procurement \nshould be approved as just and reasonable unless that obligation is met.33 \nPJM’s RBP Filing improperly uses stale information that will have a substantial \nfinancial impact on rate payers. Any procurement to serve a data center load that is not \nexpected to be online at the time of the delivery year represents a fully avoidable financial \nburden for ratepayers. PJM is well positioned to obtain updated information to adjust the \nlarge load forecast and accordingly adjust the procurement target. Failure to do so would \nresult in unjust and unreasonable outcome. \nIn direct contradiction to PJM’s assertion, even with PJM’s RBP framework, the failure \nto conduct the procurement process without updated accurate information imposes \nunnecessary costs and risks on EDCs and their customers. The evidence suggests substantial \nslowdown in the data center project development.34 \nTable 2 shows the data center load included in the summer peak load forecast for the \n2025/2026 through 2028/2029 RPM BRAs. PJM forecast that the load from the planned data \ncenters and the forecast load growth from the existing data centers (above embedded) would \nbe 16,436 MW higher for the 2028/2029 BRA. \nTable 3 shows the calculated shortfall for a range of data center load assumptions. In \nScenario 1, the data center peak load is assumed to be fixed at 11,993 MW, the level forecast \nby PJM for 2026 (Table 2). In Scenario 1, PJM would have a 1,096.5 MW surplus rather than a \nshortfall. In Scenario 2, it is assumed that only 50 percent of the forecasted data center peak \n \n33 \nSee N. Plains Res. Council, Inc. v. Surface Transp. Bd., 668 F.3d 1067, 1086 (2011) (“Reliance on data \nthat is too stale to carry the weight assigned to it may be arbitrary and capricious,” citing Lands \nCouncil v. Powell, 395 F.3d 1019, 1031 (9th Cir. 2005)). \n34 \nThe Market Monitor calculated that 4,016.9 MW have been cancelled or delayed as of September 21, \n2026. This calculation is based on tracking of datacenter project status by Industrial Info Resources \n(IIR) corroborated by publicly available data . The Market Monitor \ndoes not have access to project level data behind large load adjustments requested by EDCs. The \nMarket Monitor considered only cancelled and delayed datacenter projects that were originally \nscheduled to be in service by 2029 as indicated by the IIR tracking data."},{"page":18,"startToken":5902,"endToken":6368,"text":"- 18 - \nload would be added to the PJM system between the 2026 Summer and 2028 Summer.35 This \nscenario represents the possibility that part of the data centers would not come online as \nprojected by the 2028/2029 Delivery Year. The 50 percent assumption is illustrative and not \nbased on actual data. In Scenario 2, the shortfall would be reduced to 2,374.7 MW. Scenario 3 \nassumes that the data center load identified by the Market Monitor as cancelled or delayed \nas of September 21, 2026, will not come online by the 2028/2029 Delivery Year. The Market \nMonitor calculated that 4,016.9 MW or 42 percent of the forecasted data center load between \n2026/2027 Delivery Year and 2028/2029 Delivery Year has been cancelled or delayed as of \nSeptember 21, 2026.36 37 In Scenario 3, the shortfall would be reduced to 3,055.0 MW. \nTable 2 Embedded and above embedded data center load for the 2025/2026 through 2028/2029 \nRPM BRAs38 39 \n \n \n35 \nThe forecasted increase in data center peak load between 2026 and 2028 is 9,481 MW (21,474 MW – \n11,993 MW). A 50 percent increase equals 16,733.5 MW (11,993 + 0.5*9,481). \n36 \nPJM included 11,993 MW of data center load in the 2026/2027 BRA and 21,474 MW in the 2028/2029 \nBRA (see Table 2). The cancelled or delayed data center load of 4,016.9 MW as of September 21, 2026, \nrepresents 42 percent of the 9,481 MW increase (21,474 MW minus 11,993 MW) between the two base \nresidual auctions. \n37 \nThe Market Monitor calculation is based on tracking of datacenter project status by Industrial Info \nResources (IIR) corroborated by publicly available data . The \nMarket Monitor does not have access to project level data behind large load adjustments requested \nby EDCs. The Market Monitor considered only cancelled and delayed datacenter projects that were \noriginally scheduled to be in service by 2029 as indicated by the IIR tracking data. \n38 \nEmbedded load means existing data center load at the time of the forecast. Above embedded includes \nforecast increases in load for both existing data center load and new data center load. \n39 \nSee PJM. 2026 PJM Long-Term Load Forecast Report (February 6, 2026). Previous PJM Long-Term \n \n2025/2026 RPM BRA (Based on 2024 Load Forecast)\n4,654\n \n7,927\n \n2026/2027 RPM BRA (Based on 2025 Load Forecast)\n7,892\n \n11,993\n \n2027/2028 RPM BRA (Based on 2025 Load Forecast)\n13,018\n \n17,071\n \n2028/2029 RPM BRA (Based on 2026 Load Forecast)\n16,436\n \n21,474\n \nAbove Embedded Data Center Load\n(MW)\nAll Data Center Load\n(MW)"},{"page":19,"startToken":6368,"endToken":6863,"text":"- 19 - \nTable 3 Shortfall under different data center load assumptions \n \n3. PJM Incorrectly Allocates Data Center Costs to EDCs/LSEs and Their \nCustomers. \nPJM’s assertions about the role of LSEs obscures the fact that every EDC is an LSE but \nnot every LSE is an EDC. Competitive LSEs exist in restructured states and compete with the \nzonal EDC to serve retail customers. While the exact rules vary, if there are no competitive \nLSEs, then EDCs become the default LSE. Given that PJM’s RBP Filing would impose \nsignificant risks on LSEs, it is very unlikely that a competitive LSE would agree to serve a \ndata center load. As a result, EDCs will bear the risks that the PJM RBP Filing would impose \non what PJM terms LSEs. EDCs are the PJM Transmission Owners plus public power entities. \nTable 4 shows the list of EDCs that may be allocated a shortfall MW under PJM’s RBP \nproposal. In this filing, the Market Monitor uses the term EDC to include this group. \n \nLoad Forecast Reports . \n2028/2029 RPM BRA\nScenario 1\nScenario 2 Scenario 3\nFormula\nOrganic Load ICAP (MW)\n144,479.5\n \n144,479.5\n \n144,479.5\n \n144,479.5\n \nA\nData Center Load ICAP (MW)\n21,474.0\n \n11,993.0\n \n16,733.5\n \n17,457.1\n \nB\nForecast peak load ICAP (MW)\n165,953.5\n \n156,472.5\n \n161,213.0\n \n161,936.6\n \nC=A+B\nFRR peak load ICAP (MW)\n11,556.0\n \n11,556.0\n \n11,556.0\n \n11,556.0\n \nD\nPRD ICAP (MW)\n0.0\n0.0\n0.0\n0.0\nE\nInstalled reserve margin (IRM)\n20.0%\n20.0%\n20.0%\n20.0%\nF\nPool-Wide Accredited UCAP Factor\n78.34%\n78.34%\n78.34%\n78.34%\nG\nForecast pool requirement (FPR)\n0.9401\n0.9401\n0.9401\n0.9401\nH=(1+F)*G\nCleared UCAP (generation and DR)\n138,317.8\n \n137,332.5\n \n138,317.8\n \n138,317.8\n \nI\nCleared ICAP (generation and DR)\n176,560.9\n \n175,303.2\n \n176,560.9\n \n176,560.9\n \nJ=I/G\nRPM peak load ICAP (MW)\n154,397.5\n \n144,916.5\n \n149,657.0\n \n150,380.6\n \nK=C-D-E\nRPM peak load UCAP (MW)\n120,955.0\n \n113,527.6\n \n117,241.3\n \n117,808.2\n \nL=K*G\nOrganic reliability requirement UCAP (MW)\n135,825.2\n \n135,825.2\n \n135,825.2\n \n135,825.2\n \nM=A*H\nData Center reliability requirement UCAP (MW)\n20,187.7\n \n11,274.6\n \n15,731.2\n \n16,411.4\n \nN=B*H\nFRR Obligation UCAP (MW)\n10,863.8\n \n10,863.8\n \n10,863.8\n \n10,863.8\n \nO=D*H\nRPM reliability requirement UCAP (MW)\n145,149.1\n \n136,236.0\n \n140,692.5\n \n141,372.8\n \nP=M+N-O\nShortfall UCAP (MW)\n6,831.3\n \n(1,096.5)\n \n2,374.7\n \n3,055.0\n \nQ=P-I"},{"page":20,"startToken":6863,"endToken":7088,"text":"- 20 - \nTable 4 List of EDCs with Potential Allocation of Shortfall MW40 \n \n4. PJM Allocates Financial Risks Created by Data Centers to EDCs and \nTheir Customers. \nThe fundamental design of PJM’s RBP Filing assigns the risk for data center costs to \nEDCs and their non data center customers. PJM would require EDCs to post unprecedented \nlevels of collateral in order to cover the risks that PJM’s RBP Filing would inappropriately \nimpose on EDCs and their customers. \nThe September 10th Answer fails to address or resolve the substance of the protests \nrelated to the proposed RBP credit requirements. The protests question the size of the buyer \nside credit requirement, the basis for the calculation, the lack of adherence to accepted \nstandards and the interaction with current RPM credit requirements. In addition, PJM gives \nno consideration to the carrying cost of the LSE collateral requirement or the effect of carrying \nthis amount of debt will have on the LSE’s credit rating.41 \n \n40 \nSee RBP Filing, Attachment C (“Croop Affidavit”) at 12. \n41 \nSee Protest of East Kentucky Power Cooperative, Inc., Attachment B – Coscia Affidavit at 122 and \n142, Docket No. ER26-3380-000 (August 21, 2026) (“EKPC Protest”). \nZone\nEDC\nBGE\nBGE\nPECO\nPECO\nPPL\nPPL\nPSEG\nPSEG\nAEP\nAEP-Ohio\nAPS\nAPS-PE\nATSI\nATSI-PP\nATSI\nATSI-Ohio\nCOMED\nCOMED\nDAY\nDAY\nDOM\nDOM\nDOM\nNVEC\nDOM\nODEC\nDOM\nREC"},{"page":21,"startToken":7088,"endToken":7409,"text":"- 21 - \nRather than directly address the overwhelming deficiencies in the RBP Filing related \nto credit requirements, PJM suggests that the “LSE credit rules do not need to be effective for \nthe RBP to be conducted” and that PJM “would consent to the Commission resolving the \napplicable buyer-side LSE RBP Credit Rate through additional procedures.”42 The Market \nMonitor disagrees. Simply ignoring the massive impact of the RBP Filing on risk, credit and \ncollateral and addressing that impact only after it has been made inevitable if the rest of the \nPJM RBP Filing is accepted is not a plan. PJM’s RBP Filing must be evaluated in its entirety \nincluding its intended consequences for credit and risk and its unintended consequences for \nthe markets. The credit provisions are fundamental to the backstop auction and should not \nbe artificially separated from the backstop auction design in order to win FERC approval. \nThe issues with the buyer side credit provision stem from PJM’s proposed RBP design. \nThe failure to match the individual suppliers with the individual buyers creates the credit \nproblem by spreading the risk to EDCs and their customers rather than limiting the risk to \nthe parties to bilateral contracts. PJM’s proposed solution is to impose a credit requirement \nthat could total $10.2 billion on EDCs.43 This is an unprecedented level of collateral that \nwould almost double the total collateral held by PJM for all markets.44 \nThe September 10th Answer has no substantive response to the issues raised in the \nprotests. The overwhelming credit and collateral requirements are a direct and unavoidable \nresult of the RBP design. \n \n42 \nSeptember 10th Answer at 93. \n43 \nEqual to the Initial Target MW (6,831.3) x $1.5 million. See EKPC Protest, Attachment B (Coscia \nAffidavit) at 7. \n44 \nIn support of the EKPC Protest, Dr. Coscia notes that PJM held $10.9 billion in collateral for all \nmarkets on July 30, 2026."},{"page":22,"startToken":7409,"endToken":7830,"text":"- 22 - \nPJM’s credit requirement for EDCs with data center load is $1.5 million per MW.45 46 \nPJM arrived at the $1.5 million per MW value by first estimating the total value of a 1 MW \nRBP commitment for 15 years at an RBP average procurement cost of $555/MW-day, or $3.0 \nmillion (= 15 years x $555 per MW-day x 365 days).47 With no supporting argument based on \ncredit markets, PJM simply used half of the $3.0 million as the buyer side credit requirement. \nPJM notes that the present value of 15 annual payments ($220,575 = $555 per MW-day x 365 \ndays), discounted at 9.5 percent, is approximately $1.5 million per MW.48 49 PJM does not \nassert that this would provide adequate credit to cover the liability that the RBP Filing would \ncreate. PJM provides no analytical basis for the 9.5 percent discount rate. If a discount rate \nwere to be used, it should reflect what the provider of collateral is paid in interest. PJM’s \ndiscount rate clearly does not meet that standard. PJM’s interest payments to its customers \non held collateral is far below 9.5 percent. \nPJM’s credit requirements are significantly asymmetric between load and generation. \nWhile imposing these requirements on EDCs and their customers, PJM would require much \nlower credit for sellers of generation under the RBP Filing. PJM proposes that RBP sellers \nprovide collateral equal to 20 percent of the first three years of their RBP sell offers.50 Under \nassumptions consistent with the buyer side obligation derivation, namely that 1 MW is \noffered at $555 per MW-day, the RBP seller’s credit requirement is $121,545 per MW (3 years \n \n45 \nRBP Filing, Attachment A (Marked Tariff), Proposed OATT Attachment DD § 18.8(b)(i)(B)(1). \n46 \nRBP Obligation is an LSE’s allocation of RBP UCAP MW. See RBP Filing, Attachment A – Marked \nTariff, Definitions. \n47 \nRBP Filing, Attachment D (Kelly Affidavit) at 25. \n48 \nRBP Filing, Attachment D (Kelly Affidavit) at 27. \n49 \nDr. Coscia points out that this calculation is meritless because PJM used a discount rate (9.5 percent) \nequal to the weighted average cost of capital for merchant generation investment that PJM assumes \nin its cost of new entry analysis. PJM offers no justification as to why this discount rate would be \nappropriate for LSEs with data center load. See EKPC Protest, Attachment B Coscia Affidavit at 129. \n50 \nRBP Filing, Attachment A (Marked Tariff), Proposed OATT Attachment DD § 18.8(a)(iv)."},{"page":23,"startToken":7830,"endToken":8211,"text":"- 23 - \nx $555 per MW-day x 365 days x .20) or 4 percent of the total value of the RBP obligation. This \nmeans that the proposed credit requirements are $1.5 million per MW for the buyer and \n$121,545 per MW for the seller. The credit requirement is 50 percent of the total value of the \nRBP obligation for the buyer and 4 percent for the seller. PJM provides no justification for the \nbuyer’s credit requirement being 12.5 times higher than the seller’s. \nThe proposed provisions regarding the return of collateral would create another \nsignificant disparity between buyer and seller. The RBP Filing proposes that the collateral \nposted by the RBP seller be returned one year after the resource achieves commercial \noperation.51 RBP buyers must maintain their collateral for the full 15 year term. The return of \ncollateral for buyers is less well defined with proposed tariff language stating that the credit \nrate “may be reduced … when total posted RBP Collateral is greater than the total remaining \nRBP Charges obligations.”52 53 The RBP Filing would ignore generator operating risk. \nPJM’s asymmetric proposal does not represent a likely result of a bilateral contract in \nwhich the risk was limited to the buyer and the seller. The RBP Filing imposes significant \nrisks on the EDCs and therefore their customers, a comparatively small amount of risk on the \nsuppliers, and no risk on the data centers. The EDCs and their customers should not bear the \nrisk of these transactions. The risk should be borne by the principals of the transaction, the \ndata centers and the generators. As the Market Monitor has proposed, the buyer and seller \nwould have to agree to credit and breach of contract provisions, and the EDCs and the LSEs \nwould not be party to the agreement. \nThe EKPC Protest also identified an issue with the timing of the buyer credit \nprovisions that is not addressed by the September 10th Answer. Coscia states the “credit \n \n51 \nRBP Filing, Attachment A (Marked Tariff), Proposed OATT Attachment DD § 18.8(a)(iii)(B). \n52 \nRBP Filing, Attachment A (Marked Tariff), Proposed OATT Attachment DD § 18.8(b)(i)(A). \n53 \nIn support of the EKPC Protest, Dr. Coscia notes that this deviates from established standards and \ncurrent PJM practices."},{"page":24,"startToken":8211,"endToken":8606,"text":"- 24 - \nrequirement is absent during the period of greatest uncertainty.”54 The period to which \nCoscia refers is the time between posting the RBP results and the date that collateral is due, \nwhich is 45 days prior to the start of the delivery year.55 For the 2028/2029 Delivery Year, the \nLSE credit requirement is due the middle of April 2028. The RBP obligations for each of the \n14 Electric Distributor zone/areas with data center load (Table 4) will be fixed at the \nconclusion of the RBP. For the period from the close of the RBP to 45 days prior to the start \nof the 2028/2029 Delivery Year, no collateral secures the EDCs’ obligation and in fact, the \nEDCs will not know their credit requirements until they are assigned an RBP Obligation. If a \ndata center load fails to materialize during this period, there is no collateral to fall back on \nand the RBP charges will fall to other customers in the Electric Distributor zone/area. \nIn the RBP Filing, the RBP charges are first allocated to the Electric Distributor \nzone/areas based on their pro rata shares of the Final RBP Target MW.56 The initial RBP Target \nMW is 6,831.3 MW. This value will be reduced by Qualifying RBP Offset UCAP MW and \nUCAP MW satisfying the opt out provisions.57 58 Qualifying RBP Offset UCAP MW \ncorrespond to data center load that was included in the shortfall that has since secured its \nown capacity. The UCAP MW eligible for the opt out provision are in an Electric Distributor \nzone/area that is not regulated by the state and the EDC is unable to allocate the RBP costs, \ni.e. some public power entities. The Final RBP Target MW will be determined just prior to the \nRBP and each “zone/area’s assigned share of the Final RBP Target MW shall remain constant \nthroughout the fixed, 15-year term.”59 \n \n54 \nEKPC Protest, Attachment B (Coscia Affidavit) at 72. \n55 \nRBP Filing, Attachment A (Marked Tariff), Proposed OATT Attachment DD § 18.8(b)(i)(B)(1). \n56 \nRBP Filing, Attachment A (Marked Tariff), Proposed OATT Attachment DD § 18.1(a). \n57 \nRBP Filing, Attachment A (Marked Tariff, Proposed OATT Attachment DD § 18.1(b). \n58 \nSee example calculation in RBP Filing, Attachment C (Croop Affidavit) at 46–50. \n59 \nId."},{"page":25,"startToken":8606,"endToken":8983,"text":"- 25 - \nThe next phase of the settlement process is to allocate the RBP charges assigned to \neach Electric Distributor zone/area to EDCs. PJM states that the allocation to EDCs or end use \ncustomers will be based on values “provided to PJM by the Electric Distributor designated \nto coordinate with the states and/or RERRAs to the extent that states and/or RERRAs wish to \nprovide the Electric Distributors with information on how costs are allocated to LSEs.”60 If \nPJM does not receive RBP Obligation MW values from an Electric Distributor zone/area then \nPJM will “assign the RBP UCAP MW using the Daily Unforced Capacity Obligation MW load \nratio shares in the respective zone/area that are assigned RBP UCAP MW.”61 The RBP Filing \ndescribes this last provisions as the “backup default cost allocation.”62 \nUnder the so called backup default cost allocation, PJM will allocate the RBP Charges \nfor the area to all LSEs (EDCs here) in the area.63 In other words, the RBP charges will be \nallocated to LSEs (EDCs here) in the same manner as RPM charges. In the case where there \nare other LSEs in the EDC with no data center load, they will also be charged for the cost of \ncapacity to meet the data center load. EDCs/LSEs are expected to allocate those costs to their \ncustomers per the method currently used to allocate capacity costs.64 \nIn the case that a data center fails, the corresponding obligation will fall to other \ncustomers in the Electric Distributor zone/area. This is true whether the LSE has posted \ncollateral or not. The collateral posted by the EDC is only accessed in the case that the EDC \nactually defaults on RBP charges. If the EDC defaults, the proposed OATT Attachment DD § \n18.9(a) would apply and the charges would be “allocated to all remaining Load Serving \nEntities responsible for paying RBP Charges” and “may be offset by the RBP Collateral \n \n60 \nRBP Filing, Attachment C (Croop Affidavit) at 51. \n61 \nRBP Filing, Attachment C (Croop Affidavit) at 52. \n62 \nRBP Filing, Attachment C (Croop Affidavit) at 41. \n63 \nRBP Filing, Attachment A (Marked Tariff), Proposed OATT Attachment DD § 18.1(b)(iii). \n64 \nSee RAA Schedule 8 § D."},{"page":26,"startToken":8983,"endToken":9332,"text":"- 26 - \nposted by the defaulting” LSE.65 Otherwise, the charges would be assigned under proposed \nOATT Attachment DD § 18.3(b)(i)-(ii) to other customers in the Electric Distributor zone/area. \n5. PJM Ignores Fundamental Locational Market Requirements. \nThe RBP Filing and September 10th Answer ignore the locational requirements that \nare an essential part of the capacity market. The PJM Capacity Market has been, from the \ngenesis of the RPM design in 2007 based on locational reliability differences, been a locational \nmarket. The PJM Capacity Market includes Locational Deliverability Areas (“LDA”). PJM \ncalculates a reliability requirement for every LDA with Capacity Emergency Transfer Limit \n(“CETL”) that exceeds the LDA’s Capacity Emergency Transfer Objective (“CETO”) by 115 \npercent or meets other criteria. PJM has been using CETO and CETL metrics to account for \ntransmission network constraints that prevent capacity in one part of the system from reliably \nserving load in another part of the PJM system. PJM’s failure to account for the actual location \nof data centers and the generation to serve them will create additional and unnecessary costs \nfor all customers. \nThe central objective of PJM’s Reliability Pricing Model is to limit potential loss of load \nevents to approximately one event every 10 years at the RTO level. PJM applies a stricter \ncriterion to ensure that the transmission constraints do not reduce reliability. PJM’s standard \nis approximately one event every 25 years for every modeled LDA. \nMany protestors have raised this issue.66 PJM’s assertion that “all capacity in PJM is \ndeliverable to all loads in PJM until transmission constraints are met,” misses the point as \nPJM is aware.67 The deliverability analysis conducted when a new resource first interconnects \nto the PJM grid is used to identify necessary transmission upgrades and award Capacity \nInterconnection Rights. However, this type of interconnection study cannot and does not \n \n65 \nRBP Filing, Attachment A (Marked Tariff), Proposed OATT Attachment DD § 18.9(a). \n66 \nSee IMM Protest at 11; EKPC Protest at 31–33; JCA Protest at 28–29; AEP Protest at 29. \n67 \nSee PJM Answer at 60 n.179."},{"page":27,"startToken":9332,"endToken":9675,"text":"- 27 - \nreplace a local resource adequacy assessment in the capacity market based on CETO and \nCETL metrics. Unlike the engineering evaluation conducted to determine how much energy \na new generator could inject, the LDA level CETO/CETL study examines power flows under \nspecific emergency conditions caused by generation and transmission outages. The \nCETO/CETL analysis establishes the MW of capacity needed to meet the LDA’s reliability \ncriterion. PJM’s response is inconsistent with its own practice and is simply wrong. The RBP \nFiling and September 10th Answer simply ignore the real issues that would be created if PJM \nis permitted to ignore the fundamental locational realities of the capacity market. \nPJM will continue to conduct CETO/CETL analysis for each capacity market auction. \nIf the RBP results in procuring capacity that cannot be delivered to the data center load in \nsuch a way that the LDA reliability needs are met, the LDA where the data center is located \nwould be forced to procure more capacity in the capacity market auction. If the new capacity \nis not available, the LDA with the data center would pay the maximum price for capacity \n(shortage pricing). \nPJM’s argument that accounting for locational requirements for a 15 year forward \ncontract is speculative would simply ignore the implications of PJM’s RBP Filing.68 Ignoring \nlocational requirements would result in excess capacity in some LDAs and a shortage of \ncapacity in others. \nUnder the RBP Filing, a data center load in BGE would be allowed to procure capacity \nlocated in a distant LDA, e.g. ComEd. BGE, located within the SWMAAC LDA has been one \nof the constrained LDAs in PJM.69 In this scenario, BGE LDA’s load would increase without \noffsetting generation and without additional transmission capacity to import power from \noutside BGE under emergency conditions. In this scenario, BGE’s reliability requirement \n \n68 \nSee September 10th Answer at 21. \n69 \nIn the 2028/2029 BRA, BGE’s internal generation accounted for only 53 percent of BGE’s reliability \nrequirement. The BGE LDA price separated in the 2022/2023 through 2025/2026 BRAs."},{"page":28,"startToken":9675,"endToken":10007,"text":"- 28 - \nwould not be satisfied and BGE customers would pay the maximum price for capacity as a \nresult. \nAllowing capacity and load to be added without accounting for locational resource \nadequacy would also result in higher energy market prices and increased transmission costs. \nIn the example, in the BGE LDA with higher load and lack of offsetting generation and \ntransmission capacity to import power would result in higher prices in the energy market. \nThis would trigger transmission upgrades. \n6. PJM’s Procurement Process is Flawed. \nThe September 10th Answer fails to support the maximum average price of $555/MW-\nday. The September 10th Answer asserts that because some commenters would like a higher \nnumber and some would like a lower number, that PJM must be correct. PJM’s illogical \napproach does not support and cannot be the basis for selecting an exact number that will \naffect the costs of many PJM customers and the revenues of the new generators. \nPJM provides no supportable reason for its failure to update the forward prices used \nin the net revenue calculation included in the Net CONE value used to calculate Point A of \nthe Variable Resource Requirement (“VRR”) curve. The failure to update the forward prices, \nusing PJM’s own method, resulted in an overstatement of the RBP price by $82/MW-Day \nUCAP.70 \nMore importantly, PJM provides no supportable reason for its failure to provide a \nmechanism to update net revenues over the contract term. PJM fails to recognize that the net \nrevenue offset will vary significantly over the 15 year life of the contracts. This has nothing \nto do with the contract for differences settlement mechanism that simply assures that \ngenerators will be paid their contract prices. The $555/MW-day cap assumes energy and \nancillary services net revenues are exactly equal to $361/MW-day ICAP in every one of the \n \n70 \nSee Protest of the Independent Market Monitor for PJM, Docket No. ER26-3380-000 (August 21, 2026) \nat Technical Appendix, included as Attachment B for more details."},{"page":29,"startToken":10007,"endToken":10339,"text":"- 29 - \n15 years in the contracts. The PJM mechanism does not include a true up mechanism to \nensure that generators receiving a capacity payment under the RBP are neither underpaid \nnor overpaid compared to their offered gross costs. PJM’s RBP approach ignores the \nfundamental basis for the capacity market which is to provide the missing money. The same \nmissing money logic should apply to the RBP. Generators need a total revenue level to cover \nthe entire gross cost of units for 15 years. Total revenues include energy and capacity market \nrevenues. Despite PJM’s assertions, the procurement mechanism based solely on an \nincorrectly calculated and static capacity only payment creates significant risk for generators \nand for customers by failing to account for energy market net revenues during the contract \nterm. \nPJM’s RBP Filing fails to account for the likely scenario in which data center load \nincreases energy prices and increases net revenues yet customers have to pay the RBP price \nthat does not reflect the increased revenues to generators. \nThese critical elements of the RBP design are not consistent with the design of the \ncurrent backstop auction. The current backstop auction rules are based on the actual net costs \nof generators and reflect actual net revenues over the 15 year life of the contracts.71 The \ncurrent rules provide for payment of 100 percent of the required revenue, accounting for both \nenergy and capacity revenues. PJM fails to provide a reason for deviating from the current \ndesign. \n7. PJM’s Definitions of Capacity That Can Satisfy the Requirements to \nMeet Data Center Load are Inconsistent. \nIn the IMM Protest at 19-22, the Market Monitor explained the differences in PJM’s \ndefinitions of the qualifying criteria for RBP resources and the RBP offset. In PJM’s September \n10th Answer, PJM fails to address any of the concerns raised by the Market Monitor. The \nproblem is that PJM proposes to treat RBP offset capacity as equivalent to RBP supply for \n \n71 \nOATT Attachment DD § 16.4(e)(ii)."},{"page":30,"startToken":10339,"endToken":10727,"text":"- 30 - \npurposes of reducing the regional reliability requirement, even though the two have \nmaterially different contractual protections and performance obligations. \nIn the September 10th Answer at 29, PJM states, “To be clear, the eligibility \nrequirements for Qualifying RBP Offset UCAP and RBP Resources are not the same.” PJM \nstates that both the RBP resources and the RBP offset would contribute to addressing the \nresource adequacy shortfall. But that contribution is not the same. In the September 10th \nAnswer at 32, PJM claims that comparing the RBP resource and RBP offset requirements is \nan erroneous comparison. However, PJM is proposing to treat them as equivalent for \nreducing the procurement target. The comparison is relevant because the RBP Filing treats \nthem as substitutes. \nIn the September 10th Answer at 33-34, PJM argues that the less stringent requirements \nfor the RBP offset are meant to encourage parties to voluntarily enter into bilateral \narrangements rather than reliance on the RBP framework. PJM is choosing the preferred \napproach to be bilateral contracts in defining their proposal. In the RBP Filing (at 40), PJM \nstates that “bilateral agreements may not need to be longer than five years to develop new \ncapacity [Italics added].” PJM does not provide any substantive support for the less stringent \n5 year term of the RBP offset or for the differing criteria for RBP resources and the RBP offset. \nPJM does not explain how this contributes to reliability or to price stability over the next 15 \nyears or how it interacts with the other elements of the RBP Filing. \nIn addition, the RBP Filing and IRAS Filing have inconsistent definitions of capacity \nthat can be used to satisfy the requirements to meet the data center load. The IRAS Filing \nallows 10 year contracts for Bring Your Own New Capacity (“BYONC”) that could terminate \nafter three years, whereas the RBP auction would require 15 year contracts but would also \nallow a five year bilateral contract to meet the same obligation under the RBP offset option. \nThe RBP Filing includes inconsistent definitions of capacity that will result in \nunintended consequences for reliability and capacity market prices in the future. One result \nof the weaker requirements will be that data center load is returned to the capacity market \nsooner and will impose additional costs on other customers sooner. While this may be PJM’s"},{"page":31,"startToken":10727,"endToken":11018,"text":"- 31 - \ngoal, it is not consistent with the policy objective of requiring data centers to pay for their \nown costs rather than imposing their costs on other customers. \nII. MOTION FOR LEAVE TO ANSWER \nThe Commission’s Rules of Practice and Procedure, 18 CFR § 385.213(a)(2), do not \npermit answers to protests, answers, or requests for rehearing unless otherwise ordered by \nthe decisional authority. The Commission has made exceptions, however, where an answer \nclarifies the issues or assists in creating a complete record.72 In this answer, the Market \nMonitor provides the Commission with information useful to the Commission’s decision \nmaking process and which provides a more complete record. Accordingly, the Market \nMonitor respectfully requests that this answer be permitted. \nIII. CONCLUSION \nThe Market Monitor respectfully requests that the Commission afford due \nconsideration to this answer as the Commission resolves the issues raised in this proceeding. \n \nJoseph E. Bowring \nIndependent Market Monitor for PJM \nPresident \nMonitoring Analytics, LLC \n2621 Van Buren Avenue, Suite 160 \n \nRespectfully submitted, \n \nJeffrey W. Mayes \n \nGeneral Counsel \nMonitoring Analytics, LLC \n2621 Van Buren Avenue, Suite 160 \n \n72 \nSee, e.g., PJM Interconnection, L.L.C., 119 FERC ¶61,318 at P 36 (2007) (accepted answer to answer \nthat “provided information that assisted … decision-making process”); California Independent System \nOperator Corporation, 110 FERC ¶ 61,007 (2005) (answer to answer permitted to assist Commission in \ndecision-making process); New Power Company v. PJM Interconnection, L.L.C., 98 FERC ¶ 61,208 (2002) \n(answer accepted to provide new factual and legal material to assist the Commission in decision-\nmaking process); N.Y. Independent System Operator, Inc., 121 FERC ¶61,112 at P 4 (2007) (answer to \nprotest accepted because it provided information that assisted the Commission in its decision-\nmaking process)."},{"page":32,"startToken":11018,"endToken":11158,"text":"- 32 - \nEagleville, Pennsylvania 19403 \n(610) 271‐8051 \njoseph.bowring@monitoringanalytics.com \nEagleville, Pennsylvania 19403 \n(610) 271‐8053 \njeffrey.mayes@monitoringanalytics.com \nAlexandra Salaneck \nSenior Analyst \nMonitoring Analytics, LLC \n2621 Van Buren Avenue, Suite 160 \nEagleville, Pennsylvania 19403 \n(610) 271‐8050 \nalexandra.salaneck@monitoringanalytics.com \nJohn Hyatt \nSenior Economist \nMonitoring Analytics, LLC \n2621 Van Buren Avenue, Suite 160 \nEagleville, Pennsylvania 19403 \n(610) 271‐8050 \njohn.hyatt@monitoringanalytics.com \n \nDevendra R. Canchi \nSenior Analyst \nMonitoring Analytics, LLC \n2621 Van Buren Avenue, Suite 160 \nEagleville, Pennsylvania 19403 \n(610) 271‐8050 \ndevendra.canchi@monitoringanalytics.com \nPaul G. Scheidecker \nSenior Analyst \nMonitoring Analytics, LLC \n2621 Van Buren Avenue, Suite 160 \nEagleville, Pennsylvania 19403 \n(610) 271‐8050 \npaul.scheidecker@monitoringanalytics.com \nKeri Dorko \nAnalyst \nMonitoring Analytics, LLC \n2621 Van Buren Avenue, Suite 160 \nEagleville, PA 19403 \n(610) 271-8050 \nKeri.Dorko@monitoringanalytics.com \n \n \nDated: September 25, 2026"},{"page":33,"startToken":11158,"endToken":11223,"text":"CERTIFICATE OF SERVICE \nI hereby certify that I have this day served the foregoing document upon each person \ndesignated on the official service list compiled by the Secretary in this proceeding. \nDated at Eagleville, Pennsylvania, \nthis 25th day of September, 2026. \nJeffrey W. Mayes \nGeneral Counsel \nMonitoring Analytics, LLC \n2621 Van Buren Avenue, Suite 160 \nEagleville, Pennsylvania 19403 \n(610) 271‐8053 \njeffrey.mayes@monitoringanalytics.com"}],"sections":[{"p":1,"at":0,"title":"Opening text","method":"opening text","endPage":7,"endAt":609,"startToken":0,"endToken":2093},{"p":7,"at":609,"title":"I. ANSWER","method":"PDF bookmark","endPage":7,"endAt":620,"startToken":2093,"endToken":2095},{"p":7,"at":620,"title":"A. PJM’s Reliance on “Cooperative Federalism” Is Misguided and Has Resulted in a Flawed and Incomplete Proposal.","method":"PDF bookmark","endPage":12,"endAt":1113,"startToken":2095,"endToken":4083},{"p":12,"at":1113,"title":"B. The September 10th Answer Fails to Address the Issues Raised by the Market Monitor and Others.","method":"PDF bookmark","endPage":12,"endAt":1213,"startToken":4083,"endToken":4100},{"p":12,"at":1213,"title":"1. The Shortfall is Entirely the Result of Data Center Load.","method":"PDF bookmark","endPage":14,"endAt":77,"startToken":4100,"endToken":4407},{"p":14,"at":77,"title":"2. The Shortfall MW Value Is Not Correct and PJM Refuses to Correct It.","method":"PDF bookmark","endPage":19,"endAt":74,"startToken":4407,"endToken":6378},{"p":19,"at":74,"title":"3. PJM Incorrectly Allocates Data Center Costs to EDCs/LSEs and Their Customers.","method":"PDF bookmark","endPage":20,"endAt":76,"startToken":6378,"endToken":6875},{"p":20,"at":76,"title":"4. PJM Allocates Financial Risks Created by Data Centers to EDCs and Their Customers.","method":"PDF bookmark","endPage":26,"endAt":194,"startToken":6875,"endToken":9015},{"p":26,"at":194,"title":"5. PJM Ignores Fundamental Locational Market Requirements.","method":"PDF bookmark","endPage":28,"endAt":502,"startToken":9015,"endToken":9753},{"p":28,"at":502,"title":"6. PJM’s Procurement Process is Flawed.","method":"PDF bookmark","endPage":29,"endAt":1533,"startToken":9753,"endToken":10253},{"p":29,"at":1533,"title":"7. PJM’s Definitions of Capacity That Can Satisfy the Requirements to Meet Data Center Load are Inconsistent.","method":"PDF bookmark","endPage":31,"endAt":169,"startToken":10253,"endToken":10755},{"p":31,"at":169,"title":"II. MOTION FOR LEAVE TO ANSWER","method":"PDF bookmark","endPage":31,"endAt":787,"startToken":10755,"endToken":10848},{"p":31,"at":787,"title":"III. CONCLUSION","method":"PDF bookmark","endPage":33,"endAt":0,"startToken":10848,"endToken":11158},{"p":33,"at":0,"title":"CERTIFICATE OF SERVICE","method":"printed heading","endPage":33,"endAt":449,"startToken":11158,"endToken":11223}]}