Low-Energy Fridays: Electricity co-location is controversial—and common sense
One odd feature of the electric grid is that power plants that produce electricity tend to be located far from people and businesses that use electricity. There is a logic to this, of course. Some types of energy generation are geographically restricted; for example, wind farms need to go where there’s a lot of wind, which is not where many people typically live. Even where an energy source is not so limited, environmental and other factors tend to favor building power plants away from population centers. Logical or not, the distance between electric supply and demand requires many miles of power lines to link the two, adding to the cost of electricity.
Increasingly, certain special types of electric consumers, such as data centers, have decided that it might be better to be near their power source. Data centers use a lot of power, and some of them can be located pretty much anywhere. Therefore, it has made sense for those companies to “co-locate” their data centers near existing power plants. By doing so, companies can lower costs and increase reliability without worrying about red tape and the increasing length of time needed to build out the central grid infrastructure required for them to locate elsewhere.
Co-location also ties in with another trend: the growing preference of data centers for nuclear power. Constellation Energy announced last week that its Three Mile Island nuclear power plant would reopen to provide power for Microsoft data centers. Many companies operating data centers have sustainability goals that commit them to using clean power while requiring a steady stream of power 24 hours a day, making nuclear power seem like a natural fit. Data centers can be co-located with any fuel source, but the marriage of these two trends has increasingly led companies to enter into agreements with nuclear power plants to deliver power directly and on-site. This provides a mechanism for companies to meet their needs while giving an economic boost to America’s largest source of carbon-free electricity.
Co-location is innovative, which (predictably) means it is also controversial. Critics have raised a variety of concerns about co-location, most of them unconvincing. One frequently cited concern is that because a co-located data center uses electricity from a power plant, there will be less electricity from the plant available to the rest of us. But this ignores the fact that the data center would use the same amount of electricity if it were located somewhere else. What’s more, it makes economic sense for data centers to co-locate with plants that have a surplus of power because they can get more favorable terms. This means that co-location could result in a more efficient use of electricity than if the same center were located elsewhere. The additional revenue for generators can also help keep existing plants online or spur new entrants. In short, concerns about co-location taking electricity away from the system are misplaced.
A more sophisticated set of considerations has to do with the way co-locating entities interact with the broader grid. In theory, a co-locating facility could be completely divorced from the wider grid, receiving all its electricity directly from the co-located plant. If the generator stops producing power, the load doesn’t take anything from the grid. In practice, however, many companies will still want to receive some of their power from the grid and might also want to sell excess power to the grid. This can trigger various forms of regulation, but co-location is hardly the only case of this sort of thing happening in the electric system. The process is analogous to some forms of net metering (where a utility pays households with rooftop solar for adding power to the grid), and it has long existed at larger scale with certain types of factories and their associated on-site generation. Because co-location arrangement details can vary, so should the regulatory response. The important point is that these arrangements are not fundamentally different from what already exists on the grid, which means the existing system can accommodate them.
Technological advances are transforming the way the grid operates, and the rise of co-location is an example of that. Regulators should welcome changes that can make the grid more efficient.