Like most regulated monopoly utilities in the Southeast, Duke Energy (Duke Energy Carolinas and Duke Energy Progress) intends to build 12 new gas plants, all but one in North Carolina.

*Duke has not announced a location for CCs 4 and 5, but at the moment, Davidson County is the location indicated by their most recent transmission study.

Proposing to add almost ten thousand megawatts (ten gigawatts) in five years is quite the growth spurt. Analysis of Energy Information Agency (EIA) data for Duke Energy and its preceding utilities (Progress Energy and Carolina Power & Light) puts this growth spurt in context. If all twelve plants are built, adding an average of two gigawatts per year for five years would be the largest period of new capacity addition in its history.

From Boom to Bust

But Duke has had ambitious plans before, particularly during the late 1970’s and early 1980’s — the nuclear boom — which then proved unsuccessful. Between 1971 and 1987, Duke added 11,240 MW of nuclear capacity (and eleven reactors) to the Carolinas grid, but this was over a period of 16 years. What is equally compelling about this time period is how many of Duke’s planned nuclear units were cancelled. It turns out that another eleven nuclear reactors totaling 13,525 MW — more than half of the planned fleet — were never built. Duke still has site control of all but one – the South River trio of reactors planned for the North Carolina Coast.

Duke’s nuclear capacity additions, its cancellations, and the proposed gas buildout are shown below.

Most of what Duke built in the 1970s and early 1980s was coal (Belews Creek, Roxboro, and Mayo, not included in the chart above) and nuclear (Robinson, Oconee, Brunswick, McGuire, Catawba, and Harris). Duke Energy (and preceding utilities) had planned to build almost 25,000 MW of nuclear generation, but the anticipated load never showed up.

Let that sink in. Duke and its predecessors in the Carolinas cancelled 13,525 MW of nuclear projects between 1979 and 1983 because the load forecasts were ultimately wrong as costs skyrocketed, load moderated, and public sentiment about the technology changed. If built as planned, the rate impact would have been unimaginable. Building large-capacity, expensive generation to meet load projections that can evaporate is risky.

Deja Vu, All Over Again

And that is the situation we are in again today. Duke wants to build almost 10,000 MW of new gas plants at a time when load growth is once again uncertain, costs are skyrocketing, and there are concerns about the technology. If the AI bubble rapidly deflates, Duke customers are left paying for stranded, underutilized assets. If data centers shift locations to the shale gas basins in Pennsylvania/Ohio/West Virginia or Texas/Arkansas/Louisiana — as we are increasingly seeing them do — Duke customers are left paying for stranded, underutilized assets. Customers are also tethered to rising methane gas prices thanks to LNG exports.

Duke earns a hefty profit when it builds things, and the bigger, the better. Like all of our Southern utilities, Duke is structurally incentivized to over-estimate load growth and to build more generating capacity than is needed. They are not incentivized to work with neighboring utilities to share capacity. The result is that all of our major Southern utilities are asking regulators to approve plans to allow them to build far more capacity than will be needed. A recent change in state law allows Duke to pass plant construction costs on to ratepayers even before the plant has been completed and deemed “used and useful.” This law was not in place during the nuclear boom/bust, and this is one reason ten of the eleven plants were canceled before construction began (the exception was in Cherokee County, SC, which was actually canceled twice).

Duke is back at the NC Utilities Commission, asking for another big building boom, looking to pass the costs onto ratepayers while shareholders pocket the profits, and they are hoping no one will remember that their load prediction track record is a dismal failure. But numbers don’t lie, and we have receipts.