Does a grid need baseload power?

A grid needs enough capacity available at every hour, including the hardest hours. Whether that comes from plants running constantly or from a mix of flexible plants, storage and imports is an engineering and economic question, not a requirement of physics.

What words do engineers use instead?

Dispatchable, meaning output can be set on instruction; firm, meaning available when called regardless of weather; and flexible, meaning able to change output quickly. Each says something specific that baseload does not.

Is baseload demand still a real thing?

Yes. Demand still has a floor that is never crossed, and that floor is genuinely called baseload demand. The dispute is about the second meaning, where the word describes a category of power plant.

Baseload Power and Why the Term Is Disputed

Baseload originally described the part of electricity demand that never goes away, and then, by extension, the large power stations built to serve it by running continuously at full output. The term is disputed today because grids now need plants that can change output quickly more than they need plants that never change, and because the word is often used to argue that some technologies are necessary while others are not.

The original meaning

Look at a chart of a day’s electricity demand and you see peaks and troughs, but you also see a floor. Even at four in the morning, hospitals, street lighting, refrigeration, telecoms and industrial processes are running. That floor is baseload demand, and it is a genuine, measurable thing.

For most of the twentieth century, planners matched the shape of demand with three kinds of plant:

Baseload plants covered the floor. Large coal and nuclear stations, expensive to build, cheap to run, slow to start and stop. They ran flat out, all year, except for maintenance.
Mid merit plants followed the daily rise and fall, ramping up in the morning and down at night.
Peaking plants covered the last few hours of highest demand. Cheap to build, expensive to run, and used only briefly.

The logic was purely economic. If a plant has a high capital cost and a low fuel cost, you want it running as many hours as possible to spread the capital over more units of energy. If it has a low capital cost and a high fuel cost, you want it idle most of the year. Baseload was a way of operating a plant, chosen because it was the cheapest way to use one.

This is still how a lot of the world’s electricity is produced. Ember’s statistics show coal supplying about a third of world electricity in 2025, much of it from large stations designed to run continuously.

What changed

Three things unsettled the picture.

The middle of the day stopped being empty. Solar output now covers a large part of demand in sunny hours, and wind fleets can cover a large part of it whenever it is windy. The demand left over for everything else, the net load, is no longer a smooth shape with a stable floor. In some systems it dips near zero on mild, sunny, breezy afternoons and rises steeply a few hours later.

Capacity factors fell. Plants built to run flat out are increasingly asked to stop for part of the day and start again. That costs money, wears equipment and, for some designs, is simply not possible without shutting down for a long period. A plant whose economics assume it runs most hours of the year does badly when the market only wants it for some of them.

Flexibility became the scarce commodity. When the difficult thing was covering a steady floor, plants that never moved were exactly what was needed. Now the difficult things are the evening ramp, the still winter week and the sunny spring weekend, and all three reward the ability to move.

The case against the word

Critics of the term make several distinct points, and they are worth separating.

It confuses a job with a technology. Baseload is a way of running a plant, not a property the plant is born with. Some nuclear fleets follow load routinely, most notably in France, which shows that the link between the technology and the operating pattern is a choice rather than a law.

It implies a physical requirement that does not exist. The grid requires supply to match demand at every instant. It does not require any particular generator to run continuously. A combination of flexible plants, storage, interconnection and demand response can meet the same demand.

It is used to shut down debate. Saying a system needs baseload can slide into saying it needs a specific technology, skipping the question of whether that technology is the cheapest way to get the service. Critics argue that grids should procure the services they actually need, which are energy, capacity at peak, ramping, reserves, inertia and voltage control, and let any resource that can supply them compete.

It hides the real problem. The hard hours in a renewable heavy system are not the quiet ones at four in the morning, which are now easy. They are the evening peaks and the multi-day weather events. A plant that runs constantly helps with those, but so does one that runs only when needed, and the second may be cheaper.

The case for keeping it

Defenders of the concept, and of the plants usually described by it, respond in kind.

Firmness has real value. Capacity that is available regardless of weather reduces how much of everything else a system has to build. Wind and solar need backup, storage, transmission or all three, and those costs belong in any honest comparison.

High utilisation is not optional for some technologies. A nuclear plant has a very large capital cost and a small fuel cost, so its cost per unit of electricity depends heavily on how many hours it runs. Telling such a plant to run half the time roughly doubles the cost of its output. That is an argument for careful system design, not a flaw in the word.

Terminology should not be confused with engineering. Whatever you call it, a system still has to work through a week when the wind does not blow. Some who defend the term are really defending attention to that constraint, and object to it being waved away.

Better words

Most engineers now prefer terms that say something specific:

Dispatchable: output can be set on instruction. Gas, hydro and batteries are dispatchable. Wind and solar are not, though they can be turned down.
Firm: capacity that can be relied on at a given time, regardless of weather. Increasingly qualified as clean firm when the interest is in low carbon versions.
Flexible: able to change output quickly, in either direction, often measured in megawatts per minute.
Inertia and system strength: physical services traditionally supplied by large spinning machines, now also obtainable from properly configured power electronics.

These distinctions matter because a plant can be firm without being flexible, flexible without being firm, or dispatchable while being too slow to help with a fast ramp. Baseload blurs all of that into one word.

How to read the argument

When someone says a grid needs baseload, it is worth asking what they mean. If they mean demand never falls to zero, that is simply true. If they mean a system needs capacity available in its hardest hours, that is also true, and everyone agrees. If they mean only a particular technology can provide it, that is a claim about cost and engineering that should be argued on those terms rather than settled by a word inherited from a different era of grid design.