Balancing Supply and Demand Every Second

The grid has almost no storage compared with its size. Batteries are growing fast, but at any given second the power going into the wires must equal the power coming out, plus losses. If it does not, frequency drifts, and if it drifts far enough, protective equipment starts cutting things off. Somebody's job, every second of every day, is to keep that balance. That somebody is a balancing authority, and EIA's 2016 tally counted 37 of them in the Western Interconnection alone.

The shape of a day

Demand follows people. EIA's analysis of hourly data shows the lowest demand around 5 a.m., when most people sleep, and on summer weekdays a climb through the afternoon to a peak around 5 to 6 p.m. as air conditioners work hardest. Winter days are flatter, often with one bump in the morning and another in the evening. Weekends and holidays run lower.

Baseload, intermediate, peaking

The old way to describe the fleet is by the slice of the curve each plant fills. It still works as a first picture, though wind, solar and batteries have scrambled it.

Role	What it does	Typical plants	Runs
Baseload	Serves the floor of demand that never goes away	Nuclear, many coal units, some hydro, geothermal	Most hours of the year, commonly at a 70 percent capacity factor or higher. U.S. nuclear averaged 91 percent in 2025, per EIA.
Intermediate (load following)	Ramps up in the morning and down at night	Gas combined cycle, hydro, some coal	Many hours, at varying output
Peaking	Covers the highest few hundred hours	Gas combustion turbines, older oil units, batteries, hydro with storage	Usually under 15 percent capacity factor. EIA's 2025 average for gas combustion turbines was 14.1 percent.
Variable	Produces when the wind blows or sun shines	Wind, solar	Whenever weather allows. Other plants fill in around it.

Why not just build peakers for everything? Because the economics run in opposite directions. A nuclear plant costs a fortune to build and very little to run, so you want it running every hour to spread the building cost out. A gas turbine is cheap to build and costly to run, so it makes sense only for the few hours when nothing else will do.

Dispatch: climbing the stack

Operators line up available resources from cheapest to most expensive to run one more megawatt-hour, called marginal cost. Wind, solar, hydro and nuclear sit at the bottom because their fuel is free or cheap. Gas combined cycle comes next, then older steam units, then peakers. The operator runs up the stack until supply meets the demand forecast for that interval. This list is called the merit order or supply stack.

Real dispatch adds constraints. A transmission line might be full, forcing a pricier plant near the load to run instead of a cheap one far away. Coal and nuclear units cannot swing quickly. Hydro operators must respect fish flows and irrigation obligations. And forecasts miss, so operators re-dispatch every few minutes. In the West, the Western Energy Imbalance Market does that re-dispatch across many utilities every five minutes, covered in lesson 5.

Reserves: the spare tire

The biggest single thing that can go wrong is losing your largest plant or line in an instant. Operators keep operating reserves to cover it.

• Regulation: plants on automatic control that nudge output up and down second to second to hold frequency at 60 Hz.
• Spinning reserve: online, synchronized capacity that can add output within minutes.
• Non-spinning reserve: offline units, such as fast gas turbines, that can start and reach output within about 10 minutes.
• Demand response: customers paid to cut use on request, such as an irrigation pump or a factory line.

Specific reserve amounts are set by NERC reliability standards and regional rules, so they vary. The principle does not: hold enough ready capacity to survive the worst single loss.

Inertia: the first few seconds

When a 1,000 MW plant trips offline, reserves take minutes to respond. Something has to hold the line for the first few seconds. That something is inertia: the momentum stored in thousands of tons of spinning steel in generators and turbines across the interconnection. As frequency starts to fall, those machines give up a little of their rotational energy and slow gradually instead of suddenly. Automatic governors then add fuel or water, and reserves come in behind them.

Wind turbines, solar panels and batteries connect through inverters, which do not provide inertia the same way by default. As they replace spinning plants, some grids see frequency fall faster after a trip. Engineers answer with "grid-forming" inverters, batteries that respond in a fraction of a second, and synchronous condensers (spinning machines that make no power but add inertia). How much of a problem this is depends heavily on the grid. Engineers are still working it out, and neither side of the argument has a settled verdict.