Key Highlights
- PJM has filed a proposed framework under which new large loads that do not bring or contract for new capacity — and are not covered through the Reliability Backstop Procurement — could face curtailment during capacity shortages.
- Under the proposed Interim Resource Adequacy Service, or IRAS, PJM would assign load-reduction obligations to affected zones, while states and electric distributors would determine how those reductions are implemented among individual retail customers.
- Bring Your Own Capacity (BYONC) enables data centers to contract for new generation or storage, integrating procurement into development timelines and reducing reliance on regional supply.
- PJM is accelerating generation interconnection through a temporary expedited track, supporting projects capable of entering service within three years, with significant state involvement.
- A proposed long-term capacity backstop procurement aims to fill the 6,831 MW shortfall identified in 2028/2029, with contracts extending up to 15 years to provide revenue certainty for developers.
PJM Interconnection has now filed one of the most consequential proposed changes yet in the relationship between data centers and the electric grid. Rather than simply treating a new hyperscale or AI facility like any other customer whose demand will be backed through regional capacity procurement, PJM is proposing a framework under which the largest new loads would need to be supported by new capacity, have their needs covered through the Reliability Backstop Procurement, or face potential curtailment when the regional power system is short of supply.
The approach has been developing since PJM launched its Critical Issue Fast Path process for large loads in 2025, but it became substantially more concrete in late July and August 2026. PJM filed its proposed Reliability Backstop Procurement with FERC on July 31 and began accepting applications that day for its FERC-approved Expedited Interconnection Track. On Aug. 13, PJM filed its proposed Interim Resource Adequacy Service, or IRAS, along with the Large Load Registry that would support it.
The immediate numbers explain the urgency. PJM's July 2026 capacity auction for the 2028/2029 delivery year procured 138,318 MW of unforced capacity through the centralized auction. Even after including Fixed Resource Requirement resources, however, PJM came up 6,831 MW short of its reliability requirement. The auction cleared at the FERC-approved $325/MW-day price cap. It was the second consecutive auction in which the PJM region failed to procure its full reliability requirement, something that had not happened before these two auctions.
That gap is occurring while demand continues to accelerate. PJM's 2026 long-term forecast projects summer peak demand growing at an average 3.6% annually over the next decade, compared with just 0.3% in the comparable forecast issued in 2021. Summer peak demand is projected to rise by nearly 66 GW over 10 years. Data centers are a major component of the large-load adjustments driving that change, as PJM's latest resource-adequacy planning work says prospective large-load additions could reach as much as 70 GW by 2038.
The Core Idea: Firm Power is No Longer Automatic
The most significant piece for data center developers is PJM's proposed Interim Resource Adequacy Service, or IRAS, previously known as Connect and Manage.
Under the filed IRAS proposal, a new large load that does not bring its own new capacity — or have its capacity needs covered through the Reliability Backstop Procurement — could still connect. If PJM faces a capacity shortage, however, it would direct affected zones to reduce demand in proportion to their share of new Large Load not supported by new generation. States and electric distributors would determine which retail customers fulfill that reduction before PJM calls on broader Pre-Emergency Load Management resources.
That distinction could become important to the economics of future AI campuses. Developers would effectively have a choice between securing enough new capacity to support their load—or accepting some degree of interruptibility during extreme system conditions.
PJM is also proposing a consequential change beginning with the 2029/2030 capacity auction: incremental new Large Loads that do not bring their own new supply would be excluded from the demand used to determine the capacity PJM procures in future auctions. The change is intended to prevent existing consumers from bearing higher capacity costs created by new large loads that have not brought or contracted for the supply needed to serve them.
The proposed IRAS treatment would apply beginning June 1, 2027, to new Large Loads that have not brought their own new capacity or otherwise secured supply. PJM defines a Large Load as an end-use customer with cumulative peak demand of at least 50 MW at a single site behind one or more delivery points or points of interconnection within a one-mile radius. The Large Load Registry would track both existing and new Large Loads, although IRAS curtailment obligations are aimed at incremental new demand.
Importantly, PJM is not proposing that every unbacked data center automatically be shut down whenever the grid is tight. IRAS reductions would depend on the actual capacity shortage and the amount of new load in the affected areas. PJM says only the megawatts required to address the real-time resource-adequacy shortage would be subject to reduction. Transmission owners and electric distributors, coordinating with states, would determine which individual customers fulfill a zonal reduction requirement.
IRAS is therefore more targeted than a blanket ‘data centers get disconnected first’ policy.
Bring Your Own Capacity Becomes a Development Strategy
The alternative is what PJM initially described as Bring Your Own New Generation and now increasingly frames as Bring Your Own New Capacity, or BYONC.
BYONC does not require a hyperscaler to build generation beside the data center. A load-serving entity can instead identify qualifying new supply associated with the load, including contractual arrangements that add accredited capacity to the PJM system. PJM has also launched a facilitated bilateral matchmaking process designed to connect large-load customers seeking long-term power with developers of new generation, storage and demand-side resources.
In practical terms, power procurement moves earlier in the development cycle. A company planning a 500 MW or 1 GW campus may need to advance IT infrastructure, utility interconnection and generation procurement in parallel. Sites with credible capacity strategies could therefore carry materially different reliability and commercial profiles from projects simply waiting for regional generation to catch up.
PJM Is Also Trying to Build the Missing Supply
Demand management is only one half of the strategy. PJM is simultaneously trying to accelerate construction of generation.
FERC approved PJM's temporary Expedited Interconnection Track in June, and PJM began accepting applications July 31. The program can advance as many as 10 qualifying generation projects per calendar year during 2026 and 2027. Projects must provide at least 250 MW of unforced capacity, have support from an appropriate state siting authority and be capable of entering commercial operation within three years. PJM expects successful applicants to reach a Generation Interconnection Agreement within roughly 10 months.
The mechanism is deliberately aimed at shovel-ready generation rather than speculative projects. State involvement is particularly significant because PJM can accelerate its own interconnection studies, but it cannot eliminate state permitting, siting, financing, equipment procurement or construction delays. FERC Commissioner David Rosner made the same point in his concurrence on the EIT order, emphasizing that state authorities, load-serving entities and developers must still take the financing, procurement, permitting and construction steps needed to turn faster interconnection into actual generating projects.
For data center developers, that creates another option: align new campuses with generation projects already advanced enough to qualify for the expedited track.
A Multi-year Backstop for the 6.8 GW Gap
PJM is also proposing a more direct intervention: procuring additional capacity through a one-time backstop.
On July 31, PJM filed its proposed Reliability Backstop Procurement with FERC. Assuming regulatory approval, PJM plans a one-time procurement with bidding running from Sept. 30 through Oct. 21 and results expected by Dec. 2. The initial target is tied to the 6,831 MW shortfall identified in the 2028/2029 capacity auction, although PJM would reduce that target to account for new supply secured through bilateral contracts, integrated resource plans or qualifying demand-side arrangements.
The contracts could extend as long as 15 years, an unusually long commitment compared with PJM's normal capacity-market structure. Eligible resources must generally represent new capacity and be capable of entering service no later than June 1, 2032. PJM proposes a maximum volume-weighted average cost of $555/MW-day for the portfolio selected through the procurement.
Those long-term commitments are intended to provide the revenue visibility that major new generation projects may not get from PJM's traditional capacity-market structure, bridging today's rapidly rising demand with the supply needed to support it.
Who Pays Could Be the Most Contentious Question
PJM's plan also attempts to address one of the strongest criticisms of the data center boom: that residential and ordinary commercial customers should not be forced to finance generation required primarily because of large new loads.
PJM proposes to allocate Reliability Backstop Procurement costs first among zones or service areas based on their forecast load additions, using the difference between their 2028 and 2026 summer load forecasts. Those costs would then be allocated to Load Serving Entities through a new Reliability Backstop Obligation. PJM does not have jurisdiction to determine which individual retail customers ultimately bear those costs; electric distributors and state regulators would determine how the assigned costs are allocated among data centers, other commercial customers or the broader rate base.
That jurisdictional boundary may prove as important as PJM's market reforms themselves. The Board has explicitly endorsed the principle that new large loads should bear the costs they cause, but PJM acknowledges that implementing that principle at the individual data center level will require state action.
The result could be a patchwork across the PJM footprint. Virginia, Pennsylvania, Ohio, Maryland, New Jersey and other PJM states could adopt different tariffs, capacity requirements, peak-shaving programs and cost-allocation policies even while operating inside the same regional wholesale market.
The Large Load Registry May Quietly Become One of the Biggest Changes
Less visible, but potentially just as consequential for planning, is PJM's proposed Large Load Registry.
Utilities would report detailed information about qualifying projects, including location, expected in-service dates, demand levels and ramp schedules, associated new capacity, allocated backstop capacity and even backup-generation capability. PJM says aggregate information would be made public where confidentiality rules allow.
That addresses a persistent planning problem: the same project can appear in multiple utility pipelines, change its requested capacity or repeatedly move its energization date. PJM has already tightened its forecasting methodology to separate more committed near-term projects from speculative development. The registry should give PJM, utilities and regulators a more consistent view of which gigawatts are credible and when they may actually arrive.
A Fundamentally Different Data Center Power Market
Taken together, PJM's measures amount to something larger than another interconnection reform.
For hyperscalers and other large-load developers, the region is moving toward a different development equation. Developers seeking firm service will have stronger incentives to secure new capacity, while those willing to operate flexibly may gain earlier grid access at the cost of curtailment risk. Generation able to enter service quickly gets an expedited path through PJM’s studies, while the backstop gives PJM another way to secure supply if the competitive market still comes up short.
The most important change is therefore philosophical. PJM is no longer approaching extraordinary data center growth solely by asking how the grid can produce enough electricity to serve every megawatt requested. It is increasingly asking a second question: what responsibility should the customer adding hundreds or thousands of megawatts assume for making sure that power exists?
In this June 2026 CSIS discussion, Equinix’s Arthur Haubenstock, former FERC Chairman Neil Chatterjee and other energy-policy experts examine how new federal large-load rules could accelerate data center grid connections while addressing reliability, cost allocation and ratepayer impacts.
At Data Center Frontier, we talk the industry talk and walk the industry walk. In that spirit, DCF Staff members may occasionally use AI tools to assist with content.
Keep pace with the fast-moving world of data centers and cloud computing by connecting with Data Center Frontier on LinkedIn, following us on X/Twitter and Facebook, as well as on BlueSky, and signing up for our weekly newsletters using the form below.
About the Author



