Fusion reactors: Efficiency
Two kinds of gain, and why neither yet pays
What this tab covers How much of the energy available is actually captured or extracted, in per cent: theoretical limits, laboratory records, typical new equipment and in-service averages, each with exactly what is being measured.No fusion machine has yet returned more electricity than it draws from the grid. The record laser shot (April 2025) released about four times the laser energy hitting the target, but only about 3% of the electricity used to fire the lasers (calculated).
Chart: Efficiency ranges
Record CommercialEfficiency by measure
| Measure | Range | Type | What is measured |
|---|---|---|---|
| Target gain, laser fusion (NIF record, April 2025) | 4.13 x (gain)1 | Record | Fusion energy released (8.6 MJ) divided by laser energy delivered to the target (2.08 MJ) |
| Plasma gain Q, tokamak sustained for 5 seconds (JET, 2021) | 0.33 x (gain)5 | Record | Fusion power out divided by heating power put into the plasma |
| Wall-plug gain, NIF ignition shots (December 2022 to April 2025) | 1.05 to 2.9%12 | Record | Fusion energy (3.15 MJ in 2022, 8.6 MJ in 2025) divided by the about 300 MJ the laser draws from the grid per shot; the 2025 value assumes a similar grid draw (calculated) |
| Heat to electricity assumed for future tokamak plants | 35 to 38%18 | Commercial | Thermal-to-electric efficiency range used in UKAEA cost modelling; a design assumption, not a measurement |
Where the energy is lost
| Step | Loss |
|---|---|
| Grid electricity to laser light at NIF | over 99% lost (about 2 MJ of light from about 300 MJ drawn) (calculated)2 |
| Fusion heat to electricity in a steam or gas turbine cycle | about 62-65% rejected as low-grade heat (calculated from 35-38% conversion)18 |
Two gains matter. Scientific gain (Q) compares fusion energy with the heating or laser energy reaching the fuel. Wall-plug gain compares electricity out with all electricity in, including magnets, lasers, cryogenics and pumps. ITER, the international tokamak in France, has a design target of Q=10 (500 MW of fusion power from 50 MW of heating); this has not been demonstrated, and ITER will not generate electricity.
A power plant needs Q well above 10 because its heat passes through a turbine at 35-38% efficiency and its heating systems are themselves inefficient. Laser fusion needs a gain of 10 to 20 on the laser energy and lasers firing about 10 times a second; NIF fires about three times a day.
Sources
- National Ignition Facility (FY2025 annual report) Lawrence Livermore National Laboratory, 2025
- Ignition Accomplished: NIF Looks to the Future Optica, Optics & Photonics News, 2022
- Nuclear-fusion reactor smashes energy record Nature, 2022
- Extrapolating Costs to Commercial Fusion Power Plants (Foster et al., IEEE Transactions on Plasma Science 52) UKAEA, 2024