PJM Load Forecast 2026: What Data-Center Demand Does to Capacity and Energy Prices

Noreva, an energy market data and analytics provider (formerly Karbone Research), tracks the PJM load forecast against what the grid actually does each summer – and in 2026 the two stopped moving in lockstep. PJM’s 2026 Long-Term Load Forecast, posted January 14, 2026, projects RTO summer peak demand growing 3.6% a year over the next decade, driven largely by data centers (PJM, 2026 Load Forecast Report, January 14, 2026). Six months later, a July heat wave pushed PJM to an all-time peak load record and real-time prices above $1,600/MWh, while ERCOT set its own all-time demand record the same month without prices leaving double digits for most of the day.

This piece reads the PJM load forecast against the July 2026 grid data, then against what it means for the next capacity auctions and for how energy price curves should be built. A short ERCOT comparison shows how resource mix shapes the way a record peak shows up in price.

What Does PJM’s 2026 Load Forecast Actually Show?

PJM’s Resource Adequacy Planning Department publishes an independent, 20-year load forecast each January, covering peak load, net energy, distributed solar, electric vehicles and battery storage for every zone and the total RTO. The 2026 edition, built on hourly models estimated through August 2025 and 403 simulated weather scenarios, is explicit about the source of growth: summer peak load is projected to grow 3.6% per year over the next 10 years and 2.4% per year over the next 20 – up from a 10-year average of just 0.3% per year in the 2021 forecast (PJM, 2026 Load Forecast Report, January 14, 2026; PJM Inside Lines, January 14, 2026). RTO summer peak is forecast to reach 222,106 MW by 2036 and 253,077 MW by 2046 (PJM, 2026 Load Forecast Report, January 14, 2026).

The report names the source of that growth. PJM adjusted 14 zones upward specifically for data-center load: AEP, ATSI, APS, BGE, ComEd, Dayton, DLCO, JCPL, METED, PECO, PEPCO and PL for data centers alone, plus Dominion for data centers and a voltage optimization program, and Public Service (PS) for data centers and port electrification (PJM, 2026 Load Forecast Report, January 14, 2026). The zonal growth-rate table confirms the concentration: Dominion’s 10-year summer growth rate is 5.4%, PPL Electric (PL) is 6.4%, AEP is 5.3% and ComEd is 3.9% – all above the 3.6% RTO average – while Duquesne Light and Rockland Electric, with little data-center exposure, sit near flat (PJM, 2026 Load Forecast Report, January 14, 2026).

PJM also trimmed the near term. Compared with the 2025 Load Report, the 2026 forecast is lower by 2,564 MW (-1.6%) for the 2026 summer peak, 4,414 MW (-2.6%) for the 2028 peak used in the December 2025 capacity auction, and 1,630 MW (-0.8%) for the 2031 peak used in transmission planning – a direct result of PJM tightening its rules for large-load adjustment requests, as near-term years now need “firm” commitments such as Electric Service Obligations rather than speculative interconnection requests (PJM, 2026 Load Forecast Report, January 14, 2026). PJM is underwriting the near term more conservatively while holding the long-term trajectory steady, a tension Noreva’s coverage of data-center builders tracks from the load side of the queue.

PJM Summer Peak Load Forecast: Key Milestones and Sources

How Did PJM’s Actual Peak in July 2026 Compare to the Forecast?

PJM did not wait for 2030 to feel the load growth in its forecast. On July 2, 2026, the RTO recorded an all-time peak load, which PJM’s July 10, 2026 update put at an estimated 168,158 MW between 5 and 6 p.m., surpassing the prior record of 165,563 MW set in August 2006 (PJM Inside Lines, July 10, 2026). The region posted its hottest-ever RTO average air temperature, 97 degrees, that day, with 18,100-19,400 MW of forced generation outages running July 2-4 even as PJM maintained reliable operations (PJM Inside Lines, July 10, 2026).

Noreva’s own hourly grid dataset for the same window (72 hourly rows, PJM RTO, July 1-3, 2026) shows demand of 162,648 MW at the hour ending 18:00 on July 2 against a day-ahead forecast of 164,450 MW for that hour – actual demand came in modestly under forecast for that settlement hour, even as the day set the RTO’s operational record (Noreva hourly grid data, July 2026). The two figures measure different things: PJM’s record is an instantaneous metered peak, while the hourly series reports the average load over the full hour, which always sits below the peak instant inside it. Noreva’s series is an hourly average and is not intended to reproduce PJM’s operational peak. Both describe the same event: one of the tightest hours PJM has run this cycle.

At the peak hour, gas-fired generation supplied 70,435 MW, roughly 43% of PJM’s load at that hour, alongside 31,704 MW of nuclear, 28,761 MW of coal, 11,477 MW of solar and 2,733 MW of wind (Noreva hourly grid data, July 2026). The real-time RTO price at that hour was $1,621/MWh; the day’s maximum, $1,869/MWh, came an hour later at 161,875 MW of demand (Noreva hourly grid data, July 2026). Averaged over 24 hours, July 2 cleared at $496/MWh with 9 hours above $500/MWh; July 1 (peak 161,965 MW) averaged $191/MWh with 2 hours above $500; July 3 (peak 155,028 MW) averaged $360/MWh with 5 hours above $500 (Noreva hourly grid data, July 2026).

Nine hours above $500/MWh in a single day – sixteen across the three-day window – is not a rounding error on an annual average – it is the tail event that a merchant curve is built to price. A generator, storage asset, or load-serving entity that is short during those hours captures or pays a multiple of what the smoothed annual price would suggest. That is the practical argument for building PJM price curves around hourly dispatch and scarcity dynamics rather than annual or seasonal averages.

Why Did ERCOT’s 91 GW Peak Stay Below $210/MWh?

ERCOT set its own all-time peak demand record in July 2026: 91.1 GW on July 22, 6% above the prior record of 85.5 GW (EIA, Today in Energy, August 3, 2026). Its peak hour cleared at a real-time average of just $47.75/MWh (Noreva hourly grid data, July 2026) – a fraction of PJM’s $1,621/MWh – because the peak landed inside ERCOT’s solar window, while PJM’s gas-and-nuclear stack was running close to its ceiling. The gap is about resource mix and timing, not megawatts; for the hour-by-hour detail on both events, see Noreva’s comparison of the July 2026 PJM and ERCOT record days.

What Does This Load Growth Mean for the Next PJM Capacity Auctions?

PJM’s capacity market prices in the load forecast’s trajectory as much as its current level. The 2025/2026 Base Residual Auction cleared the RTO at $269.92/MW-day; the 2026/2027 auction cleared the entire footprint – including previously constrained zones like BGE and Dominion – at $329.17/MW-day, the FERC-approved cap, securing 146,244 MW UCAP against a reliability requirement it exceeded by just 139 MW, which PJM described as “underscoring the region’s tightening supply-demand balance”; and the 2027/2028 auction cleared at the FERC-approved cap of $333.44/MW-day, the second consecutive delivery year to clear at the ceiling (PJM news release, July 22, 2025). PJM’s forecasted peak load rose more than 5,400 MW year over year for the 2026/2027 Delivery Year, which PJM attributed largely to data-center expansion, electrification and economic growth (PJM news release, July 22, 2025). Full auction-by-auction detail is in Noreva’s coverage of PJM’s structural grid stress and why PJM isn’t ready to let capacity prices run.

With the 2026 Long-Term Load Forecast showing 3.6% annual summer peak growth sustained through the next decade, and data centers named as the driver behind upward adjustments in 14 PJM zones, there is little in the forecast that argues for near-term relief on the demand side of the next auctions. Supply response – new generation, deactivation withdrawals, and PJM’s Reliability Resource Initiative, which had attracted more than 11,000 MW of planned projects and upgrades as of the 2026/2027 BRA – is the variable to watch (PJM news release, July 22, 2025).

Does Hourly Price Shape or the Annual Average Matter More for PJM Valuation?

The July 2026 data makes the valuation case directly: Noreva’s PJM LMP curves are built on dispatch modeling, capacity-market interaction, nodal basis forecasting and scenario analysis because gas sets the marginal price in more than 60% of peak hours across the PJM footprint (Noreva, PJM LMP Merchant Curve coverage), and because a handful of scarcity hours – not the smoothed annual average – determine realized revenue for peaking capacity, batteries and load-following contracts. An asset valued off an annual average LMP would have badly mispriced the economics of July 2, 2026: a day that averaged $496/MWh in real time but touched $1,869/MWh at its peak, compared with July 1’s $191/MWh average from a lower but still record-adjacent peak. The gap between those two days, both inside the same heat event, argues for hourly-resolution forecasting over averaged annual price assumptions – particularly for assets whose revenue concentrates in the top few hundred hours of the year, where Western Hub, the standard PJM reference hub, remains the benchmark for merchant curve assumptions.

How Noreva Turns This Into a Forecast

Noreva’s PJM capacity curves combine fundamentals-based modeling with trader-informed pricing inputs from Karbone’s 15-plus years of capacity market trading and origination experience, spanning near-term BRA forecasts (1-5 years), 25-year merchant curves, low/base/high scenarios and LDA-level granularity across RTO, MAAC, EMAAC, SWMAAC, DOM, BGE and PSEG, delivered via API and Data Hub. On the energy side, Noreva’s PJM LMP methodology layers dispatch modeling, capacity-market interaction, nodal basis forecasting and scenario analysis across six hubs plus nodal delivery points, built for project finance, lender due diligence and PPA structuring; because gas remains the marginal fuel in most PJM peak hours, Noreva’s natural gas forecasting coverage feeds directly into how that marginal price behaves as data-center load builds through the rest of the decade. Noreva’s merchant curves refresh monthly as load, capacity accreditation, gas prices and realized dispatch move.

Frequently Asked Questions

What is PJM’s peak load forecast for 2030?

PJM’s 2026 Long-Term Load Forecast puts the RTO summer peak at 183,008 MW for 2030, part of a trajectory that reaches 222,106 MW by 2036 and 253,077 MW by 2046 (PJM Inside Lines, January 14, 2026, citing the 2026 Load Forecast Report).

How much of PJM load growth comes from data centers?

PJM’s 2026 Load Forecast Report explicitly adjusts 14 zones upward for data-center load growth, including Dominion, AEP, ComEd and PPL Electric (PL), all with 10-year summer growth rates well above the 3.6% RTO average (PJM, 2026 Load Forecast Report, January 14, 2026). The zones with the largest announced data-center pipelines carry the highest growth rates – the concentration a capacity forecast has to price.

Why did PJM prices spike in July 2026?

PJM set an all-time peak load record – an estimated 168,158 MW between 5 and 6 p.m. on July 2, 2026, above the prior 2006 record of 165,563 MW – during an extreme heat event that pushed forced generation outages to 18,100-19,400 MW (PJM Inside Lines, July 10, 2026). Noreva’s hourly grid data shows real-time RTO prices reaching $1,869/MWh that day, with gas supplying roughly 43% of PJM’s load at the peak hour and 9 hours across the day clearing above $500/MWh (Noreva hourly grid data, July 2026).

Is ERCOT’s July 2026 peak an all-time record?

Yes. ERCOT recorded an all-time peak demand of 91.1 GW on July 22, 2026, 6% above the previous record of 85.5 GW set August 10, 2023 (EIA, Today in Energy, August 3, 2026). Unlike PJM’s July 2 event, ERCOT’s peak hour cleared at a real-time average of $47.75/MWh, reflecting a resource mix weighted toward solar at the time of the peak; see Noreva’s comparison of the July 2026 PJM and ERCOT record days for the full hour-by-hour picture.

How does Noreva forecast PJM capacity and energy prices?

Noreva’s PJM capacity curves combine fundamentals-based modeling with trader-informed pricing inputs from Karbone, covering near-term BRA forecasts, 25-year merchant curves and LDA-level scenarios. Noreva’s PJM LMP curves layer dispatch modeling, capacity-market interaction, nodal basis forecasting and scenario analysis across PJM’s hubs and nodal points, refreshed monthly. Book a demo to see the current PJM capacity and energy curves.

Methodology note: figures cited as “Noreva hourly grid data, July 2026” come from an hourly-resolution dataset covering PJM RTO (July 1-3, 2026) and, where cited, ERCOT (July 21-23, 2026): average hourly demand, generation by fuel type, and real-time price (PJM RTO price; ERCOT hub average). The full ERCOT hourly breakdown – generation mix, battery behavior, and the evening ramp – is presented in Noreva’s companion piece on the July 2026 PJM and ERCOT record days.

Price the PJM Load Forecast With Noreva

Noreva’s PJM capacity and LMP curves are built on the hourly reality of the grid: dispatch, scarcity pricing and transaction data. As data-center demand continues to move PJM’s load forecast and the capacity auctions that follow it, Noreva’s monthly-refreshed merchant curves give project developers, traders and investors a forward view built for how the grid actually clears.

Book a demo to see Noreva’s PJM capacity and energy price forecasts.