Load following is the practice of adjusting power plant output up and down throughout the day to match the constantly changing demand for electricity. It sits between two other modes of power generation: baseload plants that run at a steady output around the clock, and peaker plants that only fire up during the highest-demand hours. Load-following plants fill the gap, ramping production up during the morning when people wake up and businesses open, easing back in the early afternoon, and ramping again in the evening.
How Load Following Fits Into the Grid
A power grid has to balance supply and demand moment by moment. If generation falls behind consumption, the grid’s frequency drops; if generation exceeds consumption, the frequency rises. Either scenario can damage equipment or cause blackouts. To prevent this, grid operators rely on three tiers of generation.
Baseload plants, typically large nuclear and coal facilities, run continuously and produce the cheapest electricity per unit because they operate at a steady, efficient output. They cover the minimum level of demand that exists even at 3 a.m. Peaker plants, often small natural gas turbines, only switch on during the highest-demand windows and are expensive to run but can start quickly. Load-following plants occupy the middle ground. They adjust output as demand rises and falls through normal daily patterns, absorbing the routine swings that baseload plants are too sluggish to handle and that don’t justify firing up a peaker.
Natural gas combined-cycle plants are the most common load-following resources today because they can ramp relatively quickly and operate efficiently across a wide output range. Hydroelectric dams also excel at load following since water flow through turbines can be increased or decreased in minutes.
The Automated Systems Behind It
Grid operators don’t call power plants on the phone and ask them to turn up. The process is handled by a computerized system called Automatic Generation Control (AGC), which monitors grid frequency and adjusts the output of multiple generators across different plants in real time. AGC sends signals every few seconds, nudging generators up or down to keep supply and demand in balance.
Load following operates on a timescale of roughly 10 minutes to several hours, covering the predictable rises and dips in electricity use. It’s distinct from two related services. Frequency regulation is faster, correcting second-to-second fluctuations. Spinning reserves are generators already running and synchronized to the grid that can ramp to full output within 10 minutes if another plant trips offline unexpectedly. Together, these services form what grid operators call “operating reserves,” though the exact definitions vary by region.
Why Plants Can’t Just Turn On and Off
Thermal power plants, those that burn fuel to create steam, face real physical limits when changing output. Every plant has a minimum stable level, the lowest output it can sustain without shutting down. Below that threshold, combustion becomes unstable, emissions spike, and equipment can be damaged. For a coal plant, that floor might be 40% of its maximum capacity. For a nuclear plant, the range is even narrower, which is why nuclear is almost always treated as a “must-run” baseload resource in the United States.
Combined heat and power (CHP) plants, which generate electricity and useful heat simultaneously, also have high minimum output levels because their heat customers need a steady supply. In Texas, the overall minimum generation floor from CHP plants alone is roughly 6 gigawatts, about 40% of the state’s total CHP capacity. California’s CHP fleet shows slightly more flexibility, with output dropping as low as 30% of maximum during overnight hours.
These minimum generation constraints matter because they determine how much room the grid has to absorb variable renewable energy. If thermal plants can’t reduce output further and wind or solar keeps producing, something has to give.
The Cost of Constantly Ramping
Cycling a power plant up and down is hard on equipment. A comprehensive analysis of more than 150 coal-fired units found that the financial costs of cycling operations are substantial. Damage shows up as increased maintenance expenses, forced outages, shorter equipment life, and degraded efficiency. When a plant operates at partial load or ramps frequently, its heat rate (how much fuel it burns per unit of electricity) worsens significantly. That means more fuel burned for less output.
The specific cost categories include higher maintenance and overhaul spending, replacement energy costs during forced outages, extra fuel and staff needed for startups, and long-term capacity losses from shortened unit lifespans. One study found that correcting excessive temperature swings during startups at a single plant could reduce costs by at least 20% per start. Multiply that across hundreds of plants cycling daily, and the system-wide expense is considerable.
How Renewables Changed the Game
Solar and wind generation can’t be dispatched on command. When the sun is shining, solar panels produce power whether the grid needs it or not. This has fundamentally changed what load following looks like for grid operators. The relevant metric is no longer just total electricity demand but “net load,” the demand remaining after subtracting wind and solar output.
In California, the growth of solar capacity has created what’s known as the duck curve. During midday, solar floods the grid, pushing net load to very low levels. Then in the evening, as the sun sets and people come home, solar output drops rapidly while demand climbs. The result is an extreme ramp: conventional power plants must increase output very quickly over just a few hours. That swing is getting steeper every year as more solar comes online, and it’s no longer unique to California. The pattern is appearing across the country and globally wherever solar penetration is growing.
This creates two problems. The first is operational. Ramping gas plants fast enough to cover the evening surge stresses equipment and makes it harder for operators to match supply and demand in real time. The second is economic. Conventional plants that once ran most of the day now sit idle during sunny hours, earning less revenue. If those reduced earnings make a plant uneconomical to maintain, it may retire. Losing dispatchable capacity makes the ramping problem worse, since there are fewer plants available to handle the evening surge.
Battery storage is increasingly being deployed to smooth these transitions, absorbing excess solar during the day and discharging it during the evening ramp. This effectively shifts load-following duty from gas turbines to batteries for a portion of the cycle.
How Load Following Is Paid For
In restructured electricity markets, load following can be compensated as an ancillary service, a support function that keeps the grid stable beyond simply producing energy. However, pricing for load following has historically been underdeveloped. Many markets still bundle load-following costs into flat tariffs rather than charging customers based on how much their usage patterns actually contribute to the need for ramping.
This creates a fairness problem. A factory with steady, predictable consumption imposes far less load-following burden on the grid than a facility with wildly fluctuating demand, yet both may pay similar rates. Generators providing load-following service often aren’t adequately compensated for the wear and tear on their equipment. Some regions have moved toward more granular pricing, but developing competitive markets for load following remains a work in progress. Until pricing better reflects the true cost of ramping, generators have limited financial incentive to invest in the flexibility the grid increasingly needs.