The panel has become cheap. The system around it has not.
That does not mean we should replace large power plants with millions of tiny ones. Utility-scale solar and wind still have important economies of scale, and transmission is indispensable. It lets New England use Quebec hydro, allows a windy region to support a calm one, and gives one part of the country a way to help another during periods of stress. With data centers, factories and electrification pushing demand upward again, the United States will need more generating capacity and more transmission.
What we should question is the assumption that every increase in demand requires an equivalent increase in infrastructure.
Europe is beginning to experiment with another approach through flexible grid connections. A factory or data center that might otherwise wait years for the network to be enlarged can sometimes connect earlier if it agrees to cut its demand during the relatively few hours when the local grid is constrained. Instead of simply asking, “How much electricity do you need?” the utility begins asking a more useful question: When do you actually need it?
That turns the traditional utility bargain on its head. Electricity delivered without regard to system conditions is not costless, and customers that can be flexible may be cheaper to serve than customers that require maximum power without interruption. A factory that can pause one process for an hour, a warehouse that can pre-cool or a data center that can delay some computing has something valuable to sell the grid: flexibility.
The same principle can work across thousands of homes. Batteries, electric cars, thermostats and other devices can be coordinated into what utilities call a virtual power plant. When the system approaches its peak, it can call on those customers to use less electricity or release some they stored earlier. Instead of maintaining another peaking plant for a few extreme hours, a utility may be able to buy a little flexibility from a great many customers.
None of this makes the grid obsolete. Quite the opposite. A large interconnected grid provides reliability and geographic diversity that a battery in the garage cannot match, and large-scale generation will often produce electricity more cheaply than thousands of individual household installations.
The question is how much work we really need to ask the grid to do.
A house that produces most of its own electricity does not need the utility to supply 30 kilowatt-hours every day. A battery that carries that house through the evening may eliminate much of its demand at precisely the hour when capacity is most expensive. A factory willing to shed part of its load for a few critical hours does not necessarily need the same grid connection as one demanding maximum power without interruption.
For the ordinary household, the point is not that dinner should wait until electricity gets cheaper at midnight. It is that people should not have to pay for vastly more power plants, transformers and wires than are really necessary simply because millions of flexible devices happen to turn on during the same few hours. Let the car charge later. Let the water heater run earlier. Let the battery cover the peak. Keep the lights on and cook dinner when you want.
The economics should recognize that difference. Customers who help during scarce hours should be paid for providing capacity, while utilities should have to compare another transformer or feeder with storage and flexible demand rather than assuming construction is always the answer. Flexibility has value, but reliable household electricity still has to remain basic service.
The underlying choice is not centralized power versus decentralized power. We will need both. Build large solar and wind projects where scale makes them cheapest. Build transmission where geographic diversity makes it valuable. Keep dependable generation available for periods when weather and storage cannot carry the load. But use batteries, electric vehicles, controllable buildings and some local generation to reduce the amount of infrastructure that must exist solely because everybody might want maximum power at the same moment.