How a Power Grid Works
A power grid is a machine that produces electricity and delivers it in the same instant. Power stations, wind farms and solar farms feed electricity into high voltage transmission lines, local distribution networks step the voltage down and carry it to homes and businesses, and a control room adjusts generation continuously so that output matches consumption. Because there is almost no storage in the middle, that balancing act never stops, not for a single second.
Generation: many machines, one system
Most electricity is still made by spinning a magnet inside a coil of wire. A gas turbine, a coal boiler, a nuclear reactor, a hydro turbine and a wind turbine all do the same job by different means: they turn a shaft, and the shaft turns a generator. Solar panels are the exception, producing direct current that an inverter converts into alternating current that the grid can accept.
What matters for the grid is not only how much each plant can produce but how quickly it can change. A hydro plant can go from nothing to full output in minutes. A gas turbine takes minutes to hours depending on its design. A nuclear reactor prefers to run steadily. Wind and solar produce what the weather allows, and can be turned down but not up.
Every country ends up with a different mixture of these machines, shaped by geography, history and policy. You can see the mixture for each fuel on our fuel pages, and the pattern is striking: Ember’s statistics for 2025 show nuclear supplying about 69% of generation in France, a share that no other large country comes close to.
Transmission: the long distance network
Transmission lines are the motorways of the system. They run at very high voltages, typically between 100,000 and 800,000 volts, because losses in a wire depend on the square of the current flowing through it. Raising the voltage lowers the current needed to carry the same power, which is why a transformer sits at almost every junction between one part of the system and the next.
Transmission networks connect generation to the cities and industries that use it, and they connect regions to each other, so a shortage in one place can be covered by a surplus somewhere else. They are meshed rather than linear: electricity does not travel along a chosen route but spreads across every available path in proportion to the electrical resistance of each one. This is one reason grid operation is harder than it looks. You cannot simply dispatch a flow from A to B and expect it to stay there.
Distribution: the last few miles
Distribution networks take power off the transmission system, step it down through a series of substations, and deliver it at a few hundred volts to a socket in a house. Historically this was a one way street. It is now increasingly two way, because rooftop solar, small wind turbines and batteries connected at the distribution level push power back up the network when they produce more than the local area is using.
That change has consequences. Equipment designed for flow in one direction has to be re-rated, and the operator of a local network needs visibility of thousands of small generators it does not control.
The balance that never stops
Here is the part that surprises most people. The grid holds essentially no buffer. If a city switches on its kettles, the extra power comes from generators producing more in that same moment, not from a tank somewhere.
So the operator forecasts demand for the day ahead, hour by hour, using weather, the day of the week, holidays and history. Generators are scheduled to meet that forecast. Then, as the day unfolds and the forecast turns out to be a little wrong, the operator adjusts: a plant is asked to lift output, another to drop, a battery discharges, a large industrial user reduces its load under a contract.
The errors being corrected are small in percentage terms but large in absolute terms. A 1% forecast error on a system serving 40 million people is hundreds of megawatts, roughly the output of a mid sized power station.
Frequency: the grid’s pulse
The way the system tells you whether the balance is holding is frequency. Every generator connected to an alternating current grid spins in step with all the others, at 50 cycles per second in most of the world and 60 in the Americas, Japan’s western half, Korea and a few other places.
When demand exceeds generation, the spinning machines give up some of their rotational energy to cover the gap and slow down slightly, so frequency falls. When generation exceeds demand, they speed up and frequency rises. Frequency is therefore a real time measure of balance that every point on the system can read at once, and it is the signal that automatic controls respond to.
Operators hold frequency within a narrow band, usually a fraction of a hertz. Layers of response are contracted in advance: some generators change output automatically within seconds, others within minutes, and a final layer exists to disconnect blocks of demand automatically if frequency falls far enough to threaten the whole system. That last step is deliberate and controlled. It exists to prevent a cascade in which one machine after another trips off to protect itself.
Who runs all this
Somebody has to be responsible for the balance in each area, and the arrangements differ by country. In the United States the job is done by balancing authorities, dozens of them, each responsible for matching supply and demand in its own footprint and for settling the flows across its borders. We publish an hourly picture of these areas from the US Energy Information Administration on our grid zone pages, including large ones such as ERCOT in Texas, and a US overview at our United States page.
In Europe, transmission system operators run their own control areas and cooperate through a synchronous network that stretches from Portugal to Turkey. In most of the world there is a single national operator, sometimes part of a vertically integrated utility that also owns the power stations.
The titles differ. The physics does not. Somewhere, every second, a computer and a small team of people are keeping the amount produced equal to the amount consumed.