Exploring PJM's Zonal Hourly Load

Each panel highlights one month of 2025. The solid line traces the mean hourly load across all days in that month, representing the typical shape of electricity demand over a 24-hour period.

Notice how the profile shifts throughout the year. If you imagine these curves as the surface of a lake, winter months look a bit choppy, with a load bump in both the morning as people wake and thermostats recover from overnight setbacks, and the evening as temperatures drop.

Summer profiles are almost surfable by comparison. Air conditioning load builds through the hottest hours of the day, producing a significant late afternoon peak. These months have the steepest ramp, starting with relatively low demand around 4 AM and climbing to a peak around 5 PM. That ramp rate matters for operators, who need to bring generation online fast enough to keep up with demand.

The shoulder months (April, May, October, and November) sit in between with lower loads overall and gentler swings throughout the day. Neither heating nor cooling dominates the load profile in a significant way, and the grid is about as relaxed as it gets.

Also visible in every month is the baseline load that never approaches zero. Even at 3 AM, industrial processes, streetlights, refrigeration, and always-on equipment sustain a floor of roughly 12 to 14 GW across the AEP zone.

The darker band shows the interquartile range, the middle 50% of observed loads for each hour. The lighter band extends to the middle 90%.

Consider August at 4 PM. On a mild, overcast afternoon, the AEP zone might draw around 16 GW. On a hot day, that same hour could push past 22 GW.

This variability is what these bands represent. Variability is driven by weather, weekday versus weekend rhythms, and broader economic activity.

While not universally true, winter and summer months tend to have the widest spread. Cold snaps might produce dramatic swings in heating demands, while mild winter days may barely stress the system at all. Similarly, a string of 95°F days will stack high loads consistently, while a cool front temporarily drops demand back toward the mean.

The width of these bands matters for resource planning. PJM must maintain enough generation capacity to serve not just the average hour, but the range of plausible outcomes.

The dashed turquoise line shows the 24‑hour profile of each month's single highest demand day. These are the days with the most potential to push the transmission and generation system to its limits.

Peak days are driven by extreme weather: arctic cold in the winter, and heat and humidity in the summer. On these days, nearly every home and business is running its HVAC system at the same time, compounding demand that might otherwise be non-coincident.

Notice that the peak day's profile doesn't sit uniformly above the mean. During overnight hours, even the most extreme day may fall back within the normal band. Often the stress is concentrated in a handful of hours: the morning ramp in winter, and the afternoon plateau in summer.

In 2025, AEP's peak of peaks occurred on a cold January morning. In another year, the summer could easily take that title instead. Either way, these few hours matter more than any other for how the grid is built and how costs are allocated. The entire capacity infrastructure of the system, billions of dollars in generation, transmission, and distribution, is sized to reliably serve those few hours.

Together, these three layers tell the story of how electricity demand behaves across the AEP zone over the course of a year. The mean profile might shape energy procurement and rate design. The variability bands quantify some of the risk that planners must hedge against. The peak days highlight the limits the system is built to handle.

For commercial and industrial customers, performance during those peak hours directly determines their share of transmission and capacity charges for the following year. A customer who can reduce load during the system peak, or who has invested in on-site generation, may save hundreds of thousands of dollars annually.

Understanding how and when demand peaks form is the starting point for evaluating on-site generation, demand response participation, and long-term energy planning. The next question is how that picture might change with a significant influx of new load onto the grid.

The orange dashed line scales each month's peak day profile in line with PJM’s most recent load forecast, representing a scenario in which data center and large load interconnections materially increase system demand by 2030. Click '2030' above to cycle through the projection years from 2026 to 2030.

This represents a significant departure from the past decade, during which overall electricity demand in the region was largely flat or declining. AEP's current interconnection queue includes over 38 GW of data center load requests, of which AEP expects 11.2 GW to materialize by 2030.

The projected peaks are striking. Summer months that already push 23 GW could approach 32 GW under this scenario. Winter morning peaks would likewise intensify further as data center baseload stacks on top of weather-driven demand.

Also noticeable is how little the relative shape of demand changes; rather, the profile is just shifted up. This is likely because data centers are expected to have high load factors, and unless these facilities curtail during peaks, their usage is expected to add to the baseload.

The consequences touch every layer of grid economics: generation capacity requirements, transmission expansion costs, and ultimately the rates that residential, commercial, and industrial customers pay. The question of how to plan and pay for a grid that can reliably serve this load, as well as evaluating how much of this projected load is likely to materialize, is one the industry is actively working through.