The Grid We Chose

Between 1960 and 1980, the U.S. population grew from 179 million to 227 million, an increase of approximately 26%. During that same period, installed electrical generating capacity expanded from 167 GW to 579 GW, an increase of 246%. We expanded generating capacity at nearly 10 times the rate of population growth. More importantly, approximately 365 GW of that expansion was conventional fossil and nuclear generation, capable of delivering electricity independent of weather conditions.

Now consider 2005 to 2025. The population increased from 295 million to 342 million, adding almost exactly the same number of people, while generating capacity increased just 31%. But the disparity goes well beyond the rate of expansion. During those 20 years, the U.S. retired approximately 145 GW of coal capacity, added 130 GW of natural gas, and experienced a net reduction of approximately 47 GW in conventional fossil and nuclear capacity. Meanwhile, we added roughly 300 GW of intermittent wind and solar generation. We expanded installed generating capacity at just twice the rate of population growth, while actually reducing our conventional dispatchable generating capability.

The economic consequences are equally important. We replaced substantial amounts of established dispatchable generation with intermittent resources that produce inexpensive electricity when available, but cannot independently satisfy continuous demand. When those resources aren't producing, the remaining dispatchable fleet must fill the gap, increasingly relying on fuel-price-sensitive natural gas and higher-cost peaking resources. In marginal-price electricity markets, the cost of the last generating resource needed to satisfy demand establishes the energy clearing price. Adding hundreds of gigawatts of inexpensive intermittent generation does not necessarily improve the marginal economics during the constrained hours when dependable capacity matters most.

And all of this occurred before accounting for the accelerating electrical demand associated with AI, data centers, and other technology-driven infrastructure. We spent two decades replacing conventional generating capacity without expanding its net capability, while fundamentally changing the economics and dispatchability of the electrical system. Now we're attempting to accommodate a new era of continuous, industrial-scale electrical demand and debating whether large loads are responsible for the resulting capacity constraints and increasing electricity prices. Perhaps the more consequential question is why we spent 20 years developing an electrical system increasingly incapable of economically accommodating the next generation of demand.