Some media and the usual cohort of environmentalists have once again decided to disengage their brains and embrace some bullshit to bolster their narrative, as is their wont. The latest is a story that solar power is cheaper than nuclear, based on a report that some stories are calling “a new study by two researchers at Duke University”. Despite that sciencey description, this is not a peer-reviewed paper and it’s not from independent researchers. It’s a position paper published by NCWarn, an environmentalist organization whose primary goal is the elimination of nuclear power.

The paper is at heart nothing but rhetoric arguing in general terms that nuclear is bad and solar is good. The only hard numbers to support the contention that solar is cheaper are relegated to an appendix, where it is revealed that completely different methods were used to calculate the costs for solar and nuclear. They took the costs of nuclear from a single hand-picked study (and uprated them, assuming that nuclear would get more expensive as time went on), but they came up with their own formula for capital cost per kilowatt for solar.

Their formula incorporates the project cost, an “amortization factor” which is effectively how much of this project cost is attributable to any given year (they estimated about 7.8% based on a 25-year lifetime), the generating capacity of the project, and the capacity that could actually be used—solar only generates 18% of what it could if it were high noon on a clear day every second of the year. Then they included federal and state subsidies to knock the price down by another 65%. Their sample calculation shows an $18,000 system generating 3 kW at 85% efficiency and 18% utilization, for a real cost of 35 cents per kilowatt-hour, or 16 cents per kilowatt-hour after subsidies.

As I said, they don’t apply this formula to nuclear plants, however they do include an appendix of recent nuclear plants, their capacities, their original estimated cost, and the upward cost revisions, in an attempt to show how the cost of these plants is always underestimated. Of course, this means we can plug these numbers into their formula and compare solar and nuclear apples-to-apples. With no subsidies (so comparable to the 35 cents per kilowatt-hour number for solar), here’s what we get for their list, using only the higher revised costs (after overruns exceeding 300% in some cases):

  • Florida Power & Light Turkey Point Reactor: 7.2 cents per kilowatt-hour
  • Progress Energy Shearon Harris 2 & 3: 3.7 cents per kilowatt-hour
  • Progress Energy Levy: 9.1 cents per kilowatt-hour
  • CPS South Texas Project: 6.0 cents per kilowatt-hour
  • S. Carolina Elec. & Gass V.C. Summer: 5.0 cents per kilowatt-hour
  • Duke Energy William Lee: 4.5 cents per kilowatt-hour
  • PPL Bell Bend: 8.3 cents per kilowatt-hour
  • TVA Bellefonte: 7.1 cents per kilowatt-hour
  • Atomic Energy of Canada, Ltd. Darlington (cancelled): 9.6 cents per kilowatt-hour
  • Constellation Energy Calvert Cliffs: 5.3 cents per kilowatt-hour

Even using this list of the most egregious nuclear plant cost overruns, the authors’ own formula concludes that completely unsubsidized nuclear power is half the price of heavily subsidized solar. If a nuclear plant approaches even a quarter the cost per kilowatt of a solar project it is cancelled as uneconomic. And there’s a reason for this: residential US electricity prices average around 10–12 cents per kilowatt hour. Even the subsidized 16-cent price given for solar is way more than people are currently paying for power.

I should say for the record that I consider their formula to be absurdly naive, and that I’m a big fan of solar technology for a lot of reasons (distributed infrastructure chief among them). But I’m extremely concerned about the willingness of environmentalists to embrace lies and brand them as science. There’s a reason I’ve learned to question any numbers reported by environmentalists: in most cases where I’ve checked, they’ve turned out to be fraudulent.