Electricity demand in PJM is on the rise.

After years of flat forecasts, PJM’s 2025 outlook projects peak demand climbing from 160 GW to 210 GW by 2035. This projected growth is almost entirely due to the continued buildout of data centers.

​However, while high demand growth is likely, growth at the scale of PJM's most recent forecast is infeasible.

Supply chain bottlenecks, permitting delays, queue reforms, and PJM’s proposed Non-Capacity-Backed-Load category mean much of this “paper demand” won’t appear as 'scheduled'.

A more realistic view puts peak demand closer to 185 GW by 2035 (or 177 GW if NCBL goes through). That’s still substantial growth of more than 15%, but far short of PJM’s 50 GW, 31% jump.

Development constraints will keep PJM’s demand growth outlook in check

​PJM’s 2025 outlook, updated with utility filings, points to 66 GW of large load additions by 2035. But while demand projections keep rising, the supply side ultimately sets the ceiling.

PJM has struggled to add new capacity in recent years. While queue reform will shorten interconnection timelines for generators, it won’t enable demand growth at the scale PJM projects.

A realistic projection of new generation over the next decade - based on the current queue, PJM’s cluster study timelines, and assumptions around attrition and completion rates - shows PJM adding far less than its demand outlook requires.

This forecast is anchored in what the capacity market is designed to ensure - procuring enough supply to cover peak demand, plus the reserve margin.

If PJM can’t maintain that margin, reliability is at risk - and new demand is unlikely to connect.

Projected additions from the queue add high nameplate capacity, but with low ELCC

​The queue might enable 72 GW of new installed capacity (ICAP) in PJM by 2035.

​Most of this new capacity is solar, with batteries adding a growing slice.

​Solar and batteries are crucial to reduce net load and meet energy needs. Storage can provide a significant contribution to capacity calls in PJM, with an Effective Load Carrying Capability (ELCC) of 50% for four-hour duration batteries. However, solar’s ELCC in PJM’s capacity market is low - just 8-11%.

Conversely, as supply chain pressures ease and natural gas turbine procurement becomes more feasible, the forecast assumes ~5.8 GW of natural gas generation to be online between 2029-2032. This adds ~4 GW of unforced capacity (UCAP) to the capacity market supply stack each year, with a 70% ELCC.

Once derated for ELCC, the 72 GW pipeline provides only 21 GW of firm capacity that counts towards PJM’s reserve margin. ​

​Generation deployment won't be the only constraint to large load development

​The Modo Energy load forecast’s downward adjustment reflects real-world headwinds.

If generation projects can’t come online fast enough, there won’t be enough capacity to serve every new campus.

Beyond generation and storage deployment limits, other factors have the potential to restrict data center growth. Transformers remain one of the biggest bottleneck with delivery times stretching 3 to 5 years.

​Additionally, local siting and permitting battles - especially in Northern Virginia - add delays.

For data centers, these hurdles are decisive. Against that backdrop, full realization of the projected 66 GW of large loads in PJM is unrealistic.

​Large loads drive most demand growth but make projections more uncertain

​Without large loads like data centers, PJM’s peak demand is projected to stay flat over the next decade.

But large loads create the potential for rapid demand growth - and widen the range of possible outcomes.

Since PJM’s January forecast, utility filings and earnings calls push the headline number higher - raising large load additions from 59 GW to 66 GW by 2035.

​But many requests are speculative or duplicative, and therefore never fully realized.

Assuming only 35% of anticipated capacity materializes, large load growth by 2035 drops to 23 GW.​

How the forecast is built: from installed to qualified capacity

To project feasible demand growth, the first step is to establish how much new capacity can realistically come online.

From there, a view of how much demand growth can be supported under PJM’s capacity market is constructed.

​But many requests are speculative or duplicative, and therefore never full. (ICAP) for retirements and new builds. Then, derate the installed capacity by each technology’s ELCC to determine it’s qualified capacity.

When tested against PJM’s 2025 load forecast, this stack shows a severe shortfall, growing from 10 GW to 30 GW, between supply and the reliability requirement. Against the Modo Energy forecast, the gap narrows but doesn’t close.

PJM’s proposed Non-Capacity-Backed Load (NCBL) offers a workaround for connecting new demand. If implemented in its current form, PJM would assign this gap as NCBL - demand exposed to curtailment.

The NCBL framework lowers the bar for new load to connect, but shifts curtailment risk onto customers.

​​A recent joint proposal from Amazon, Microsoft, Constellation, and others suggests alternatives: voluntary demand response, backup generation, or curtailment only in rare events.

In that design, new load acts as demand response and earns a high ELCC in the capacity market. ​If adopted, the capacity market would back more new demand, reducing the share left exposed under NCBL.

This underscores why 210 GW by 2035 is infeasible, but the peak is more likely to land at 177-185 GW.

AEP, Dominion, and COMED utility zones capture most of the growth

​Large load growth is uneven across PJM - and the balance of demand is starting to shift.

The Modo Energy load forecast projects AEP to lead new large load additions. With ~6 GW of industrial growth in Ohio and West Virginia by 2035 - including a new chip processing plant and data centers.

COMED and Dominion follow close behind, each projected to add ~4 GW of new large loads.

Dominion’s growth comes from campus expansions in Northern Virginia’s data center corridor, alongside continued residential and commercial electrification.

​When considering total peak load - existing demand plus new additions - together these 3 zones make up 43% of PJM’s total peak demand by 2035.

​Capacity scarcity caps demand growth - but boosts the case for new build

PJM’s capacity market already indicates the system has limited available capacity to meet peak demand, and that frames the outlook for demand.

Load growth will be significant, but the real constraint is the pace of queue additions and the ability to clear bottlenecks in transformers, turbines, and permitting.

Generators will likely delay retirements, as large loads continue to push against supply limits.

The result is a system where peak demand still rises meaningfully - to 177-185 GW by 2035 - but is ultimately defined by the practical limits of what the grid can deliver.