A 2009 MIT study concluded that nuclear power plants could be built for $4 per watt, and produce electricity for 6¢ per kWh. Reactors under construction in Finland and Sweden cost about $7.50 per watt; ones in China cost $1.50 per watt. Delays due to lawsuits, difficulty certifying a new design, and licensing in an ever-changing regulatory environment add significant cost, especially interest on capital. It would be helpful if the Nuclear Regulatory Commission were to adopt the French system of licensing reactor designs, instead of individual reactors.
The operating cost of a reactor is quite low because fuel cost is low. Raw uranium contributes 0.001¢ per kWh to the cost of electricity from light-water reactors. Enriching the concentration of the fissile isotope (235U) from the natural state of 0.7%, to 5%, and fabricating fuel assemblies, contribute 0.5¢ per kilowatt hour (kWh). The lowest-cost electricity in California, 5¢ per kWh, is produced by the Diablo Canyon Nuclear Generating Station. Fixed cost amortization over the life of the facility contributes 74%, or 3.7¢ per kWh. Labor and other non-fuel recurring costs are 0.8¢ per kWh. The average California utility price is 15¢ per kWh.
The 3.3 GWe Palo Verde nuclear generating station in Arizona was constructed for $1.79 per watt. Its delivered price for electricity is 4.3¢ per kWh. It is the most profitable electric utility in the U.S.
Waste disposal is incorrectly cited as a social cost not internalized in the pricing structure. Since 1981, utilities have been paying 0.1¢ per kWh into the Federal Waste Disposal Fund for this purpose; it is included in the rate customers pay.
Another factor sometimes cited is subsidies. Federal subsidies for light-water reactors are larger than subsidies for coal, gas or hydro generation, but substantially less than for wind or solar photovoltaic (PV).
|2013 Federal subsidies for electricity generation,|
|cents per kWh.|
State and local subsidies vary. The additional California solar PV subsidy is 40% of the Federal subsidy.
The first full-scale instance of any new system is always expensive, but both construction and operating costs always decrease with experience. A 300 MWe pilot-scale IFR-type reactor could be built for less than $8 per watt. A GE/Hitachi consortium estimates they could build 380 MWe modular instances called S-PRISM (Super Power Reactor Inherently Safe Modular) for less than $2 per watt, if they were to have a stream of orders that is sufficiently secure to justify a factory to construct essentially identical ones, instead of building each one, subtly different from any other, on site.
Argonne National Laboratory, Merrick and Company, and the Landmark foundation proposed a forty hectare $500 million pilot-scale electrorefining facility to process 100 tonnes per year of any type of used fuel, a small fraction of the cost of a PUREX facility. Because utilities have been paying into the Federal Nuclear Waste Disposal Fund, and because Yucca Mountain has been canceled, this facility and similar larger-scale facilities ought to be constructed using those funds, not funded as part of the construction of new reactors, and not from the general fund of the Federal treasury.
If the goal of modernizing the energy sector is to reduce or eliminate carbon dioxide (CO2) emissions, comparison to fossil fuels is irrelevant. Several scientists calculated the only renewable source that can provide all current energy usage is solar. Wind cannot provide more than about 15% of current total energy usage, which will surely increase (and wind won't). Conservation and all other schemes, alone or together, are inadequate to close the gap between wind supply and energy demand.
Solar PV panels cost about $1.80 per peak installed watt. Setting aside their inability to destroy used nuclear fuel, it seems attention ought to focus on them instead of new designs of nuclear reactors. The amount of electricity produced in a year, divided by the amount that would be produced if the system ran continuously at full power, is the capacity factor. The Sacramento Municipal Utility District observed 15% for their solar installations. Nuclear generating stations average more than 90%. With a 15% capacity factor, the cost of a solar panel, at $1.80 per peak watt, is $12 per average watt, six times the expected cost of S-PRISM modules.
A nuclear power plant should last fifty years. Solar panel manufacturers offer twenty-five year guarantees, so one should expect them to last a bit longer, but probably not fifty years, and a solar panel must operate more than four years to repay the energy invested in its fabrication, deployment, and recycling. The $12 per average watt capital cost amortized over twenty-five years at 5%, deducting the four-plus year energy payback period, with batteries and doubled panel capacity to charge batteries, is 50-60¢ per kWh -- twelve times greater than the delivered cost from Palo Verde.
The capital cost for solar PV panels does not include operating costs, the cost of energy storage for use during nights or cloudy days, significant distribution grid changes necessary to exploit diffuse sources, and recycling. The storage problem is quite serious. It has not received significant attention. If all the batteries ever produced were to be fully charged (which is impossible because most have been recycled), they could not satisfy one night's demand. Batteries in electric automobiles are sufficient for five hours at full power. A fully charged entirely electric American automobile fleet, which consumes 20% of the nation's energy, connected to the grid at sundown, could supply 20% of current energy demand for five hours. In the morning no one could go to work because the batteries would be discharged.
When all factors are considered, the unsubsidized cost of electricity from mature-market solar panels is greater than the expected cost from a first-of-a-kind fast-spectrum nuclear reactor, and much greater than the expected cost from mature-market reactors.