Gross, net and capacity factor

Three capacities, one reactor

Take Taishan nuclear power plant 1-1 in China, which states all three:

The gap between thermal and gross is thermodynamics: this unit converts 38% of its heat into electricity. A water-cooled reactor sits in the low to high thirties, and the ceiling is set by how hot the coolant can run before it has to be pressurised into impracticality; gas-cooled plants do better for exactly that reason. The gap between gross and net is the plant running itself. Reactor coolant pumps, feedwater pumps, cooling tower fans and the rest draw 90 MW here, 5.1% of gross output.

Across the 417 operating units that publish both figures, net output is 5.6% below gross: 403.3 GW gross against 380.9 GW net. That is the size of the error you make by comparing one source's gross figure with another's net.

Why this site publishes gross

Not because gross is the better number. Because it is the field the source populates most completely: 2 of the 419 operating units on this site have no net capacity on record at all, and a total assembled from a mixture of gross and net figures would be a number with no definition. Every capacity column here says which one it is, and the reactor pages carry all three where the source has them.

Capacity factor is a different kind of number

Capacity and capacity factor get conflated constantly. Capacity is a rate. Capacity factor is a ratio over a period: the energy a unit actually generated, divided by the energy it would have generated running flat out for the whole period.

Running Taishan nuclear power plant 1-1 at 1,750 MW for every hour of a year would produce 15,330 GWh. No plant does that. It has to be shut down and refuelled, typically for a few weeks every 18 to 24 months, and it will lose days to equipment faults and grid conditions. A capacity factor of 90% means those interruptions cost a tenth of the theoretical output.

This is where nuclear looks different from most generation. The fleet spends most of its hours at full power, which the daily US status feed shows directly: the availability figure on that page is the same idea measured over a single day rather than a year, and a unit sitting at 0% is almost always refuelling rather than broken.

What no capacity figure tells you

• Whether the unit ran. A retired reactor keeps its capacity in the record. The fleet table separates capacity from status for exactly this reason, and the totals on this site count only units with an operating status.
• What it cost. Capacity says nothing about capital cost per kilowatt, which varies by a factor of several between projects and countries.
• What the fuel cost. That comes out of thermal output and burnup, not electrical capacity. See fuel market prices for the worked breakdown.