Answer: PJM Interconnection plans to curtail new data centers and other large loads of at least 50 MW during capacity shortages beginning June 1, 2027, if they have not brought or otherwise secured enough power supply. The proposal places those loads ahead of existing pre-emergency demand-response resources in PJM’s emergency sequence.
Which data centers would PJM curtail from June 2027?
PJM’s proposed Interim Resource Adequacy Service applies to new Large Loads entering service after June 1, 2027, when the load has not brought or otherwise secured sufficient supply. PJM defines a Large Load as at least 50 MW of cumulative peak demand at a single site, including specified affiliated facilities within one mile.
The original report on PJM’s data center power cuts described the measure as a response to capacity shortages. PJM’s subsequent board announcement clarifies that curtailment targets uncovered new Large Loads rather than every existing data center above the threshold.
This represents an inflection point for data center economics. Operators losing grid access—even temporarily—face two costly options: build captive generation or rely on backup diesel generators.
Why is PJM creating a large-load curtailment framework?
PJM is creating the framework because forecast large-load growth is outpacing qualifying supply. The 2028/2029 capacity auction finished 6,831 MW below PJM’s Reliability Requirement. PJM also forecasts about 70 GW of new Large Load demand by 2038 while roughly 15 GW of generation has retired since 2022.
| Policy element | What PJM says | Operational meaning |
|---|---|---|
| Affected load | New Large Loads with at least 50 MW at a qualifying site | Site aggregation and supply status determine exposure |
| Start date | June 1, 2027 | Operators need a supply or curtailment plan before service begins |
| Trigger | Capacity shortages | Curtailment is conditional, not a continuous disconnection |
| Emergency sequence | IRAS before Pre-Emergency Load Management | Applicable new loads reduce demand before existing pre-emergency resources |
| Compensation | Subject to a FERC-approved rate and state jurisdiction | Payment terms depend on regulatory approval and location |
How would PJM curtailment affect data center operations?
The practical effect is an obligation to reduce uncovered load when PJM calls an IRAS action. Compensation will depend on a FERC-approved rate and state jurisdiction. TechCrunch reported that demand-response notice can range from 30 minutes to several days, depending on forecast conditions.
Operators therefore need a documented sequence for workload shedding, backup-power transfer, fuel availability, staffing, customer notification, and restart. Data centers favor diesel over other backup fuels because fuel availability and on-site storage are well-established.
Federal rules cited by TechCrunch allow diesel generators up to 50 hours per year for demand-response events and up to 100 hours per year for emergencies and maintenance. The same reporting also highlights the environmental and public-health scrutiny surrounding diesel backup at large facilities.
How can operators reduce exposure to PJM power cuts?
Operators can reduce exposure by bringing new generation, securing supply through bilateral arrangements, enrolling qualifying demand-side resources, or engineering a curtailment plan that can protect critical workloads. The right choice depends on site size, interconnection timing, contract availability, capital, and the load that can be interrupted.
- Bring generation: Pair the site with qualifying new supply or self-supply.
- Secure supply: Contract bilaterally for capacity that PJM can recognize.
- Design flexibility: Use storage, distributed resources, or controllable demand where those assets meet program requirements.
- Prepare for interruption: Define which workloads can be reduced and how critical services transfer to backup power.
What is PJM and how did the capacity shortage develop?
PJM Interconnection serves 67 million people and operates wholesale markets and grid reliability functions across its region. It has managed demand response for decades, but the scale and speed of new data center connections have changed the planning problem: large, high-utilization sites can add demand faster than generation, transmission, and capacity commitments become available.
PJM’s proposed Interim Resource Adequacy Service, previously discussed as Connect and Manage, adds a curtailment step for uncovered new Large Loads during capacity shortages. Its proposed Large Load Registry would track each qualifying site’s location, megawatt requirement, service area, and whether the customer brings supply. That information is intended to improve forecasting, transparency, settlement, and decisions about which load is exposed.
The supply-side response is a Reliability Backstop Procurement designed to cover the remaining auction shortfall after recognized bilateral contracts and self-supply are counted. PJM proposed running a one-time Reliability Backstop Procurement from September 30 through October 21, with results in early December. The proposal allows commitments of up to 15 years and caps accepted supply offers at $555/MW-day.
What should data center operators and investors monitor next?
Operators should monitor FERC decisions, the final compensation rate, state implementation, Large Load Registry disclosures, bilateral contracting, and the Reliability Backstop Procurement result. Investors should separate sites with recognized supply from projects that depend on interruptible grid access, because the distinction affects commissioning risk, backup-power cost, and revenue continuity.
The broader signal: grid-constrained regions are entering a new phase where data center growth no longer depends solely on fiber connectivity and real estate availability. Power supply strategy is now a first-order competitive factor.
About the Author
