The fall of the midday peak
It's coming everywhere, even if the utilities don't build it
In over 100 years of commercial electricity production as a business has needed to solve one problem: how to get people to use power at night. Over the last 25 years, this has changed in a big way. The rise of solar has created the so-called Duck Curve where solar is depressing the midday load and businesses need to adapt. I decided to find out just how quickly things are changing in different parts of the US. What I have found is what Wilson Gibson already knew: the future is already here – it’s just not very evenly distributed.
Why care about load profiles?
In the early days of electric power, electric generators had a demand problem. Thanks to Faraday’s Law, bigger generators meant more power. Bigger generators meant more fuel going to one place, so the supply chain was simpler to manage. Electricity was already popular, so the grid of wires to send power to customers was an easy investment to make. The problem was that people slept at night and didn’t need those electric motors, fans, and shop lights that drove the demand during the work day meant there was no consumer for electricity overnight. Stopping generators meant wasting partially spent fuel as combustion shut down, losing momentum in ever larger turbines, and kick starting the process with more fuel in the morning. As demand continued to grow during the day, but people kept sleeping at night, these generator operators eventually hit a point near the turn of the 20th century where keeping the power on just made more sense. Electric companies created incentives such as free night time power just to keep from needing to shut down generators in hopes that eventually overnight demand would catch up enough to the daytime for round the clock profits. Even through the best efforts of companies like General Electric to spend the century finding new ways to use power at night such as the very useful refrigerators, dishwashers, laundry machines, and the not so useful electric mousetrap, industry with round the clock loads drives power at night. Time-of-use rates at night often are half what they are during the day in the US here in 2025.
I’ve written about the load profiles changing with the seasons throughout the year in the past. Renewable rich shoulder seasons followed by fossil fuel-based resources kicking back up during the summer and winter peak seasons. Now I want to take a look at how the load profile for the June timeframe looks today versus at the beginning of the century. For the first time in history, the cheapest time to produce electricity is when the sun is up and it costs nothing to turn off the power when the sun goes down. This is a game changer from the days when Thomas Edison was pitching electric lights to replace gas lamps as a safer, cheaper option to light up the night.
NYISO’s collapsing midday
New York is a nice power market to look at because not much has changed with the scope of the market over the last 25 years. While other power markets add and lose territory, NYISO has basically the same geographic shape, so it is possible to do a comparison of not just the load profile but also the absolute load. Back in 2001, data from NYISO shows that June’s load profile crosses 20 GW around 9AM and basically stays flat at 22-23 GW until 5PM when load drops as people commute home, flattens while they do whatever it is they do while awake at home, then falls back below 20 GW as they head to bed by 10PM. In 2025, the story is quite different. Load does ramp throughout the day starting at 6AM as it did in 2001. However, as the sun rises, load grows slowly, without flattening, until it peaks at 6PM, right about 23 GW as it did in 2001, when the sun’s intensity starts to fall off. Load draws down slowly at first and then again falls as it did in 2001 at 10PM when folks are off to bed.
What’s happening here? As we’ll talk about with Texas, net load show be falling down if the impact is a result of utility scale solar or wind, but it stays very much in line with the load and load forecast. Overall the total load for a day is simply lower than it was in 2001, meaning we’re not really seeing the impact from a growing HVAC or EV boom over the past 24 years. My take is that this is a whole bunch of behind the meter solar sitting on top of warehouses and homes that are narrowing the midday peak substantially. There’s no other way to account for 7-10 GW of demand reduction during the morning and early afternoon hours that completely disappears by the time 6PM rolls around.
It’s worth noting that electricity rates in NYISO have skyrocketed over the past ten years, nearly doubling the average retail rate statewide to around 26 cents per kWh with NYC feeling the brunt of that impact where prices have crossed 30 cents per kWh in general. That makes a pretty good case to put up some solar panels if you can afford the up front investment!
ERCOT net load craters midday
ERCOT in Texas looks a bit different thanks to its proclivity toward open market and some changes in geographic area since 2000. It appears that even in 2000 the daily load profile for Texas had a slower daily ramp than in NYISO. What has changed is the net load profile and it has changed in a big way. Net load is basically the load on the grid minus what is being serviced by variable resources such as wind and solar. Grid Status has a nice blog about this if you want to learn more. Texas’s midday peak has basically been flattened to even below night time baseline. However, the sunrise and sunset times still see some peaking, though even that is muted by the mass deployment of energy storage resources to the ERCOT grid. Absolute peak load in ERCOT has nearly doubled since June 2000 and ERCOT routinely sets new load and renewable output records, with every marker related to those being eclipsed in the last 12 months. Texas is in a race to deploy as much electricity generation as possible, so it has used solar to great effect. Energy storage is catching up as well, so it is a matter of time before the net load profile of ERCOT starts to look more like a flat line. This is a much different sort of electricity market than Edison would’ve imagined I think.
Coal country’s narrowing peak
PJM has expanded multiple times since 2001. This explains the massive load expansion since 2001 (y-axis values are about 3x now). However, the story looks like a combination of NYISO and ERCOT when you start to dig into the load profile. Like NYISO, there is a significant difference from 2001 when load used to ramp in the morning, hold to a flat peak throughout the day, then begin to fall off at the end of the typical work day, then accelerate the fall when people typically go to bed. Like ERCOT, the net load is starting to show some spread with the load, indicating utility scale resources have been deployed and are being tracked. This speaks to frustration from utility scale developers with the well documented interconnection queue issues that PJM has had over the last decade. This has held up utility scale projects from entering the market and only last year were changes made to expedite the process of approving new solar projects in PJM. It looks like the slowness pushed quite a few projects behind the meter, but we’re now seeing some completed projects starting to make an impact. PJM’s record tracker looks more like ERCOT with new solar records just about every week as we head toward the summer solstice.
What is missing from PJM is an indicator that energy storage is being deployed. This is the largest electricity market in the country and the narrowing peak is a huge opportunity for storage to fill the gap. PJM isn’t releasing energy storage data yet that I can see, but the load shape indicates there isn’t much to speak of. This could be a result of the aforementioned connection queue issues, but if the mass solar deployment in PJM continues, I expect to see energy storage shaving that evening peak down faster than we have seen in ERCOT (and CAISO) where storage was a riskier bet than pure solar deployments.
Load shapes impact everyday people
I wrote about this after I remembered a conversation with a friend who interned at NYISO back in the 2000s. He mentioned the midday peak was always a concern back then and now it’s barely a blip. Investing in infrastructure to deploy cheap solar and energy storage shifts the market but also improves the reliability of the grid substantially because most electricity continues to be used when the sun is up. Electricity rates in NYISO have gone up in a huge way thanks to NYISO shutting down nuclear and ramping up volatile gas resources. If NYISO had gone in on solar in the same way that ERCOT and CAISO has, the utilities could’ve captured some of the value that is now held by behind the meter resources. Instead, utilities in NYISO are stuck serving that evening peak and grasping at an ever smaller piece of the pie as they raise rates to keep afloat. This makes it more likely that those with the means to put up behind the meter will do so, and the higher rates will most severely impact those who can’t afford their own solar (and eventually their electricity bills). It’s no longer experimental to build solar (and storage) at utility scale and see if it works. It’s the responsible thing to do from both a business and social standpoint. The data says so.