PJM stakeholders are to vote on a record-breaking number of proposals on how the RTO should integrate large loads without impacting resource adequacy. (See PJM Stakeholders Present CIFP Options for Meeting Rising Data Center Load.)
A dozen packages of changes are to be voted on at a special Members Committee meeting Nov. 19, which will immediately follow the Critical Issue Fast Path (CIFP) stage 4 meeting, in which sponsors will present to the PJM Board of Managers. The voting will be advisory to the board, which outlined its intent to direct PJM to make a December filing on a path forward for large loads in its letter initiating the CIFP process. The stage 4 meeting is closed to the media.
The bulk of the packages mix and match elements of several design components that have been developed across 10 meetings held since August.
Bring-your-own-generation or capacity (BYOG or BYOC) would incentivize, or require, new large loads to have resources to serve themselves. This could take the form of expedited interconnection, penalties for large loads that don’t self-supply or prohibiting interconnection. Proposals differ on whether the resource can be existing or must be new, as well as whether it must be located adjacent to the load.
Instituting queues for large loads also features prominently in some proposals, requiring them to hold off on interconnection until there is sufficient capacity to serve them or they procure their own capacity. Opponents have argued these models could impinge on state jurisdiction over retail interconnection.
Load flexibility would allow large loads that agree to curtail similar to demand response to either qualify for expedited interconnection or subject them to mandatory curtailment under new emergency procedures if they do not bring their own generation. Some proposals include limited-duration products with a maximum number of hours a customer could be dispatched during one event and across a delivery year. In the executive summary of its proposal, PJM said limited-duration DR would not be implementable until the 2029/30 Base Residual Auction (BRA).
PJM’s original CIFP proposal featured a mandatory non-capacity backed load (NCBL) model in which large loads would not pay for or receive firm service unless they brought their own generation; the RTO has dropped that concept, but versions have been adopted in alternative packages.
Bifurcating the capacity market would add a second phase to auctions where large loads would clear after all other RTO loads, potentially receiving a higher clearing price. They differ on whether the resources participating in the second phase would be limited to new resources or could include existing assets.
PJM Proposal
PJM’s proposal would create a 10-month expedited interconnection pathway for state-sponsored resources, with reduced readiness deposits for projects paired with large loads. It would also rework how price-responsive demand (PRD) is dispatched and add state review of large load adjustments (LLAs) before PJM determines if they will be included in its load forecast.
The RTO lowered the threshold for projects to qualify for the proposed expedited interconnection track (EIT) from 500 MW to 250 MW, which several stakeholders requested to allow a broader range of projects to qualify. It opted to retain the state-sponsorship element, requiring a letter from either the governor or siting authority for the state the project is in demonstrating “commitment to expedite consideration of permitting and siting.”
The requirements were loosened to allow standalone and uprate projects, not just resources paired with large loads. The readiness requirements for unpaired projects would be doubled at $20,000/MW. The resource in a paired configuration would need to be at least as large as the load, which would be required to have a signed electric service agreement (ESA) with its utility.
The changes to the load forecast would require utilities submitting LLAs to ask the customer requesting service if its project is duplicative of any other requests for service at different locations and, if so, to specify the number of sites and the share of the load that is duplicative. A concern that has been voiced throughout the discussion is that a significant portion of the load expected could include speculative or exploratory interconnection requests.
Outside review of PJM’s forecast would also be added, empowering the RTO to bring on a third-party to conduct a broader analysis of how its estimates fit into the broader national picture.
The changes to PRD would replace the dynamic retail rate with an energy market bid price and align the resource class with DR by requiring it to respond to dispatch regardless of bid price, subject it to performance assessment interval penalties and mirror their 30-minute energy bid price caps.
The proposal includes a request for the board to initiate a second phase of the CIFP process focused on changes to the reliability backstop and incentives for large loads to bring their own generation or participate in DR programs.
“To solidify such incentives, it will be important, among other things, to ensure that loads are prioritized appropriately when load shedding is required in order to maintain supply and demand balance in real-time operations,” PJM wrote.
IMM Proposal
The Independent Market Monitor’s proposal would establish a large load queue, in which PJM would study the projects for impacts to transmission security and resource adequacy. If a project is determined to compromise either, it would be prevented from coming online until the issue had been mitigated by network upgrades, new resources entering the capacity market or the load bringing new capacity covering its demand plus the reserve margin.
There would be an expedited interconnection pathway for BYOC resources, which would be required to go online at the same time as the load. Full deliverability would be mandated both to the customer and the PJM system.
In its executive summary, the Monitor said participation in PRD and DR does not provide the same value as new generation and would not count toward the BYOC process. The high strike price for PRD and Capacity Performance penalty structure do not present sufficient incentives for demand-side resources to regularly be deployed. If DR was to qualify, it would need to be dispatchable any time capacity is needed with no run hour limits, which could result in frequent deployments if forecasts of 30 GW of data centers are correct.
The Monitor stated that if PJM does not believe it has the authority to hold off on interconnecting load it cannot reliably serve, the RTO should seek clarification from FERC. Defending its position against arguments that putting requirements on large loads would be discriminatory, it argued the proposal would prevent one set of customers from shifting costs onto others.
“The options that accept the premise that PJM must interconnect new large data center loads that cannot be served reliably means by definition that reliability will be degraded. PJM will be in the position of allocating blackouts rather than ensuring reliability,” the Monitor wrote.
Joint Stakeholder Proposal
A joint package from Amazon, Calpine, Constellation Energy, Google, Microsoft and Talen Energy aims to improve the accuracy of the load forecast, create new forms of load flexibility and establish an alternative reliability backstop that would trigger if a capacity auction clears below 98% of the reliability requirement.
Large loads would be required to demonstrate they have made financial commitments supporting their interconnection before they are fully reflected in PJM’s load forecast. That can include entering into ESAs, funding infrastructure, entering into bilateral transactions for capacity or credit support. The ramp rate and utilization of the new load would also be captured in forecasting, and protections against double-counting projects would be added, as well as a “reality check” overview of PJM’s forecast comparing it to national trends and the availability of equipment needed for data center construction.
Two new voluntary DR products would be available for large loads that can provide some flexibility with a cap on the amount of curtailments they see in a year. The first would be limited to six-hour deployments with a maximum of 24 hours in a year, and the second would allow 10-hour deployments capped at 100 hours per year. The effective load-carrying capability (ELCC) rating for the products would be reduced compared to standard DR to reflect the lower availability.
Another form of DR would be created for large loads with backup generation, which would curtail their grid service as the final emergency procedure before manual load dump. The product would likewise have its ELCC rating reduced to account for the fewer deployments.
The alternative reliability backstop would allow certain resources to submit capacity offers for up to seven-year terms. Eligible resources would be new or reactivated resources; existing resources with offers higher than the maximum price for the BRA that cleared short; and traditional DR. The clearing price they receive in subsequent auctions would remain the same, and there would be prioritization for selecting offers with shorter commitment periods. It would be effective through the 2031/32 BRA and then sunset.
Data Center Coalition, Utility and Governor Proposal
Building off PJM’s proposal, an alternative from the Data Center Coalition (DCC), Exelon and PPL, as well as the governors of Maryland, New Jersey, Pennsylvania and Virginia, would add financial requirements for LLAs, introduce a limited DR product and loosen the requirements for EIT projects. It would also extend the collar on capacity market clearing prices by one year to the 2028/29 BRA to stabilize prices while the changes are implemented. (See FERC Approves PJM-Pa. Agreement on Capacity Price Cap, Floor.)
Large loads would be required to provide an ESA or transmission security agreement or pair with an EIT project to be included in the forecast, as well as provide information about potential duplication of their load and characteristics such as ramp and utilization rates.
The limited DR product would be capped at between 24 and 240 hours of curtailment a year and could specify daily maximums as well. Large loads could also opt-in to a voluntary program where they would be curtailed as the final emergency procedure before manual load dump.
In response to stakeholders arguing that any new DR products should be curtailed at the same time as existing DR participants, package sponsors said that would be the case with the limited DR option, while the additional emergency procedure would exist outside the DR paradigm.
The EIT rules would be relaxed to allow multiple resources to serve one large load, allow resources that would otherwise deactivate to qualify, resources that did not clear in the capacity market and generators switching their fuel type. The 10-per-year limit on EIT projects and minimum output qualification would both be removed.
PJM would be required to explore changes to its energy resource interconnection service pathway as an alternative for resources seeking faster time to market without immediately providing capacity.
Protecting Ratepayers Proposal
The Protecting Ratepayers proposal from the Natural Resources Defense Council and dozens of state legislators is based on the DCC proposal but would remove large loads from the capacity market and prevent them from receiving firm service unless they procure their own capacity.
Interruptible service would be allocated to states based on the amount their load exceeds committed capacity, with the relevant electric retail regulatory authority (RERRA) allocating interruptible service to customers, similar to PJM’s NCBL model.
New resources could be expedited through the bilateral integration of generation portfolios and load (BIGPAL) model proposed by Eolian Energy in the second phase of the CIFP. Resources adjacent to a large load would qualify for a shortened study process, bypassing full deliverability to the grid and forgoing capacity interconnection rights (CIRs). Participating resources could enter the standard interconnection process to receive CIRs. (See “Eolian BIGPAL Proposal,” PJM Stakeholders Present CIFP Options for Meeting Rising Data Center Load.)
Large loads could also receive firm service through participation in PRD or DR programs or by contracting other consumers in the same locational deliverability area to participate on their behalf. The limited DR product in the DCC proposal is copied over.
PJM would hold off on purchasing an amount of load in the BRA that matches the amount of new generation it expects to be completed by the third Incremental Auction, at which point the held back capacity would be purchased.
The proposal calls on PJM to initiate a stakeholder process for large loads to fund network upgrades needed for their interconnection.
Consumer Advocate Proposal
The consumer advocates for Pennsylvania and Maryland proposed a mandatory BYOC model in which load-serving entities would be required to report the amount of contractually committed LLAs they have and procure new capacity to serve them.
“A mandatory backstop ensures that service interruption to existing customers is minimized, while allowing LSEs to serve LLAs. If the stakeholders supporting voluntary BYOC are correct that LLAs will voluntarily bring sufficient capacity, then the proposed mandatory requirements and backstop would be harmless discipline at a time when clear rules of the road are needed,” they wrote in their executive summary.
The BYOC resources would participate in the capacity market and be subject to CP penalties if they do not meet their obligations during an emergency.
States would be able to participate in a pre-emergency curtailment program for large loads, which the advocates said would reduce the risk of manual load dump in those regions.
“If these curtailments do not happen because [electric distribution companies]/states opt not to align and coordinate with PJM in protecting residential consumers, this would mean that manual load dumps will likely affect LLAs concurrently with existing residential, commercial and industrial consumers, which can exacerbate the duration and recurrence of blackouts for residential consumers,” they wrote.
Dominion Proposal
Dominion Energy Virginia’s proposal seeks to orient capacity around bilateral transactions and re-establish BRAs making up any residual needs.
The proposal would require utilities to procure new capacity for at least 70% of LLAs in the BRA for which those loads are participating and to purchase the remainder in the third IA. Utilities that fail to do so would be subject to an insufficiency penalty equal to the maximum price for that auction times the shortfall between its capacity obligation and procurement.
“The objective of the penalty design is to incentivize [load entities] to proactively procure new generation capacity to meet their new large load additions and to not rely solely on the BRA. A behavioral change in how capacity is secured for new large loads is necessary for the [capacity market] to remain a functional and viable market for existing load,” Dominion wrote in its executive summary.
The proposal would modify the EIT to include resources being constructed through state integrated resource plans and projects already proceeding through the standard interconnection cycles that meet the EIT participation requirements.
Dominion wrote that PJM’s existing LLA process balances protecting consumers against speculative service requests in near-term forecasting without being overly conservative and allowing EDCs to have discretion on the loads they believe should be included in the long-term analysis. It supports adding a third-party review of the assumptions around data center load, so long as utilities are able to provide input on any changes PJM would make. It said the Independent State Agencies Committee is the proper venue for state regulators to review LLAs.
EKPC Proposal
The East Kentucky Power Cooperative proposed a model that aims to assign the risk associated with large growth to the utilities, LSEs and EDCs that serve them. It would establish a collateralized penalty for those that enter a BRA without enough supply, including imports, to meet its demand.
The penalty rate would be set at 1.5 times the BRA clearing price times the amount of new large load. The revenue would flow to utilities that did procure sufficient capacity.
The proposal adopts PJM’s load forecasting changes and EIT model, though it would remove the state sponsorship requirement for expedited resources. Large loads would not be included in PJM’s forecast until the utility that will serve them has been identified. In its executive summary, EKPC supported Dominion’s modifications to EIT.
Recognizing that the 2026 Load Forecast is already well underway, the cooperative proposed to hold a midterm adjustment to implement the LLA forecasting changes for the 2029/30 BRA.
The cooperative opposes PJM’s load flexibility components, stating that the mechanism for curtailing PRD is unclear and participants receive firm service funded by other customers.
LS Power Proposal
A proposal from LS Power would bifurcate the capacity auction to first clear existing “organic” load and large loads paired with new generation, then run a second phase to clear new large loads without contracted generation. The latter would pay an entry fee of about $1,800/kW.
It includes a seven-year price lock that resources can opt in to for longer commitment periods, which the executive summary said would address hesitation that investors may have when using one-year price signals to determine whether to back projects with long construction and capital recovery timelines.
An expedited interconnection process for dispatchable resources with ELCC class ratings above 60% is included, with lower entry fees for pairing with large loads.
DR Coalition Proposal
A coalition of DR providers will present a package that largely mirrors PJM’s proposal while adding a limited DR product available between 24 and 100 hours a year. It also adds DR to PJM’s BYOG model for LLAs seeking to be included in the load forecast as an offset.
PSEG Proposal
A proposal from Public Service Enterprise Group includes a modified version of the EIT without the state sponsorship requirement, substituting a site control requirement for the three-year in-service qualification, and a trigger for when it is effective.
The utility wrote that only initiating the EIT when there is a resource adequacy need would prevent impacts to cluster projects, and that being able to maintain site control is a preferable metric for determining that a project will be constructed. The proposal would also replace the 10-per-year limit on EIT projects with a state-by-state limit.
The proposal would break data center load out in PJM’s load forecast, with an outside consultant contracted similar to how electric vehicle load is analyzed. PJM’s guidance for LLA requests is included as a component, requiring that large loads have an ESA or construction commitment to be included in the three-year forecast and adding characteristics like ramp rates to the information utilities should include.
PSEG wrote that data center developers and operators are not PJM members and therefore not subject to the RTO’s rules around load forecasting, adding that only they can know whether a project is speculative.
The proposal calls for an issue charge for a second phase of the Sub-Annual Capacity Market Senior Task Force to explore how a sub-annual capacity market design could be implemented. The task force is currently charged with reviewing the work of a consultant drafting a report on the topic.
SMECO Proposal
The Southern Maryland Electric Cooperative proposed a variant of PJM’s proposal modifying its PRD components.
It would lower the strike price to $1,000/MWh, compared to PJM’s $1,849, and only subject PRD participants to CP penalties if the resource is dispatched when the strike price or PAI conditions have not been met. It would also require that the PRD provider have supervisory control over the load and the ability to curtail.
