What PJM's Curtailment Plan Actually Does
The operator of the biggest electrical grid in the country has decided that when supply gets tight, data centers will be among the first customers asked to step back. PJM Interconnection made the call after a capacity auction failed to secure enough new generation to cover expected demand, leaving the grid with a gap between what it can produce and what its growing roster of energy-hungry customers wants to draw.
The mechanism is straightforward. During periods when the grid is short on power, PJM will curtail supply to data centers and other large consumers rather than risk cascading failures that could darken homes and businesses across its footprint. The policy does not take effect immediately — curtailment begins in June 2027 — and it applies narrowly, covering only data center facilities rated at 50 megawatts or above. Smaller sites fall outside the threshold entirely. In the meantime, PJM has another capacity auction scheduled, which could change how much headroom the grid actually has when the policy goes live.
Who Gets Cut, and How Much Warning They Receive
This is not an improvised emergency measure. It borrows directly from demand response, a practice that grid operators have relied on for decades, historically involving heavy industrial users such as manufacturing plants. The arrangement is transactional rather than punitive: customers whose power is interrupted receive compensation for the disruption.
Notice periods vary with conditions. Depending on how demand is forecast, a facility might get anywhere from half an hour to several days of advance warning before a curtailment event. That range matters enormously for operational planning — thirty minutes is barely enough time to spin up backup systems, while a multi-day heads-up allows operators to shift workloads, reschedule maintenance, or arrange alternative supply.
|
Policy element
|
Detail
|
|
Effective date
|
June 2027
|
|
Facility threshold
|
50 MW and larger
|
|
Compensation
|
Yes — consistent with existing demand response programs
|
|
Advance notice
|
30 minutes to several days, based on demand forecasts
|
|
Precedent
|
Long-standing programs covering large industrial users
|
The On-Site Power Scramble This Will Trigger
The most predictable consequence is a rush toward self-generation. Operators planning new facilities — and quite possibly those already running existing ones — now have a concrete reason to build their own power sources rather than depend entirely on grid supply that can be interrupted by policy.
Not every operator will make that investment. Those that skip it are left leaning on backup generators, which carry two disadvantages: they cost more to operate than grid electricity, and they typically produce considerably more pollution.
Why Diesel Dominates Backup Generation
Diesel is the default choice across much of the industry, and the reasons are practical rather than ideological. The fuel is easy to obtain almost anywhere, and it can be stockpiled directly at the facility — no pipeline, no delivery dependency, no waiting. When the objective is guaranteed availability during an outage, on-site fuel storage is hard to beat.
Federal rules place limits on how much these units can run. Generators used for demand response events are capped at 50 hours annually. A separate, larger allowance of 100 hours per year covers emergency situations and maintenance activities. Those ceilings shape how curtailment-driven generator use fits into a facility's compliance picture.
The Air Quality Fight in Northern Virginia
The pollution question is not hypothetical, and it has already produced a public controversy. Vantage Data Centers drew criticism this week over what appeared to be coordination with Virginia environmental regulators aimed at undermining a report on diesel backup generation. That report concluded that diesel units at a 96 megawatt facility in Northern Virginia could produce health damages running into the tens of millions of dollars every year for nearby residents.
The episode illustrates the tension that curtailment policy creates. Push large loads off the grid during shortages, and some portion of that displaced demand reappears as combustion happening much closer to where people live.
PJM's Broader Reliability and Pricing Problems
Curtailment is one piece of a larger set of difficulties. The grid operator has faced sustained criticism in recent months over its handling of two related issues: bringing new generating capacity online, and managing the connection of large new consumers, data centers chief among them.
The scale involved explains why the stakes are high. PJM's territory stretches from Virginia west to Illinois and serves 67 million customers. Decisions made about who gets power and who gets curtailed ripple across a substantial share of the American population and economy.
Wholesale Prices and Where the Blame Is Landing
Cost pressure has been severe. Wholesale electricity prices across the grid have nearly doubled over the past year. PJM's independent market monitor attributed a large share of that increase to data centers — an assessment that puts the sector at the center of both the reliability problem and the affordability problem simultaneously.
The Demand Curve That Makes This Urgent
None of this is expected to ease. Projections indicate that data centers will consume four times as much electricity in 2035 as they do today. Against that trajectory, a curtailment policy covering facilities of 50 megawatts and up reads less like a one-time fix and more like an early adjustment to a structural imbalance between how fast computing capacity is being built and how fast generating capacity can follow.

