A fusion gain figure with no stated denominator is not a weak number. It is not a number at all. This page explains why, and what this site records instead.
One shot · National Ignition Facility · N221204 · 2022-12-05
1.5
Target gain
fusion yield ÷ laser energy delivered to the target
195×
apart
0.0077
Wall-plug gain
fusion yield ÷ energy the facility drew to fire it
Same shot. Same day. Same 3.1 MJ of fusion energy. The two figures differ by a factor of 195 and neither is wrong — they answer different questions. 2.05 MJ of laser light reached the target; roughly 400 MJ was stored in the capacitor bank to produce it. Report the first ratio and you have a scientific milestone. Report the second and you have a machine consuming vastly more energy than it releases. Both are true, which is exactly why the number needs its denominator attached.
Most directories store a single gain value. That field cannot be populated
honestly, because the question “what is this device’s gain?” has no single answer. Once one
exists, every figure poured into it loses the only context that made it meaningful, and a
target gain from one machine sits in the same column as a plasma gain from another.
So there is no metric:gain here. An undenominated Q is not merely rejected by
validation — it is unrepresentable. There is nowhere to put it. Gain exists
only in the 6 forms below, each carrying its numerator and denominator in
the registry.
| Form | Numerator | Denominator | Notes |
|---|---|---|---|
| Target gain (Q_target) Q_target | Total fusion energy yield of the implosion | Laser energy delivered to the target, including hohlraum losses | Fusion energy released divided by laser energy delivered to the target assembly. |
| Capsule gain | Total fusion energy yield | Energy absorbed by the capsule | Fusion energy divided by energy absorbed by the fuel capsule. Always larger than target gain — quoting it as 'the' gain overstates performance. |
| Plasma gain (Q_plasma) Q_plasma | Fusion power produced | External heating power coupled into the plasma | Fusion power divided by external heating power coupled into the plasma. The magnetic-confinement counterpart of target gain. |
| Equivalent D-T gain (extrapolated) | Modelled D-T fusion power | External heating power coupled into the plasma | Gain a D-D or D-³He shot WOULD have reached had it run on D-T. A calculation, never a measurement. Requires conditions.isSimulated = true and is excluded from achieved leaderboards. |
| Engineering gain (Q_eng) Q_eng | Gross electrical output | Total electrical input including all plant recirculating power | Net electrical output divided by total electrical input, recirculating power included. No device has exceeded zero. |
| Wall-plug gain | Total fusion energy yield | Total facility grid energy drawn for the shot | Fusion energy released divided by total facility energy drawn from the grid for the shot. The number that is almost never quoted, and typically two to three orders of magnitude below target gain. |
Ordered roughly from the most flattering denominator to the least. A press release will reach for the top of this table; an engineer assessing whether a plant can power itself cares about the bottom of it.
9 in total, 2 verified against a primary source. Targets are shown but are permanently ineligible for any ranking — a roadmap number is not a result.
Total fusion energy yield of the implosion ÷ Laser energy delivered to the target, including hohlraum losses
Scientific gain Q_sci for shot N250406, from Table 4 of Wurzel & Hsu. Recorded as TARGET gain: an inertial-confinement shot, so the denominator is driver energy delivered to the target. Driver energy 2100000 J, yield 8600000 J.
Continuing progress toward fusion energy breakeven and gain as measured against the Lawson criteria (independent)
Total fusion energy yield of the implosion ÷ Laser energy delivered to the target, including hohlraum losses
Verified against the peer-reviewed paper. Quoted: "achieved a target gain Gtarget of 1.5".
Achievement of Target Gain Larger than Unity in an Inertial Fusion Experiment (independent) · Achieving Fusion Ignition | National Ignition Facility & Photon Science (operator)
Fusion power produced ÷ External heating power coupled into the plasma
Scientific gain Q_sci for shot 42976, from Table 4 of Wurzel & Hsu. Recorded as PLASMA gain: this is a magnetic-confinement device, so the denominator is external heating power coupled into the plasma, not laser energy on a target.
Continuing progress toward fusion energy breakeven and gain as measured against the Lawson criteria (independent)
Fusion power produced ÷ External heating power coupled into the plasma
Scientific gain Q_sci for shot 76778, from Table 4 of Wurzel & Hsu. Recorded as PLASMA gain: this is a magnetic-confinement device, so the denominator is external heating power coupled into the plasma, not laser energy on a target.
Continuing progress toward fusion energy breakeven and gain as measured against the Lawson criteria (independent)
Modelled D-T fusion power ÷ External heating power coupled into the plasma
DT-EQUIVALENT gain: what this deuterium plasma WOULD have produced running on D-T. A calculation about a shot that never happened, so it is marked simulated and is excluded from achieved leaderboards by the registry definition.
Overview of JT-60U Results toward High Integrated Performance in Reactor-Relevant Regime (operator)
Total fusion energy yield of the implosion ÷ Laser energy delivered to the target, including hohlraum losses
Scientific gain Q_sci for shot 103952, from Table 4 of Wurzel & Hsu. Recorded as TARGET gain: an inertial-confinement shot, so the denominator is driver energy delivered to the target. Driver energy 30000 J, yield 869.9999999999999 J.
Continuing progress toward fusion energy breakeven and gain as measured against the Lawson criteria (independent)
Total fusion energy yield ÷ Total facility grid energy drawn for the shot
Computed as 3.1 MJ yield (peer-reviewed) over 400 MJ stored per shot, the latter quoted verbatim from LLNL: "the power conditioning system stores approximately 400 megajoules of electrical energy for each NIF shot". Previously 0.01, derived from a ~300 MJ figure that traced only to secondary reporting. IMPORTANT — read the denominator carefully. 400 MJ is the LASER CAPACITOR BANK, not total facility grid draw: it excludes cooling, controls, cryogenics and target fabrication. True facility draw is therefore HIGHER, which makes this value an UPPER BOUND on wall-plug gain. The real figure is lower than 0.0078. Held at medium confidence and NOT promoted to achieved: it is a derived ratio across two sources rather than a reported measurement, and its denominator is an acknowledged approximation.
Power Conditioning System (operator) · Achievement of Target Gain Larger than Unity in an Inertial Fusion Experiment (independent)
Fusion power produced ÷ External heating power coupled into the plasma
ITER design point: 500 MW fusion from 50 MW of injected heating. A TARGET. Permanently ineligible for any achieved leaderboard.
The machine (operator)
Fusion power produced ÷ External heating power coupled into the plasma
CFS states SPARC is designed to exceed Q=1, with higher figures discussed. Recorded conservatively as the stated threshold rather than the optimistic case — a planned value should not flatter.
SPARC: Proving commercial fusion energy is possible | Commonwealth Fusion Systems (operator)
The wall-plug figure above is an upper bound, not a measurement. Its denominator is the laser capacitor bank, which excludes cooling, controls, cryogenics and target fabrication. Total facility draw is higher, so true wall-plug gain is lower than the number shown. It is held at medium confidence and has deliberately not been promoted to achieved, because it is a ratio derived across two sources rather than a reported result.
Saying so costs nothing and is the whole point. A benchmark that never shows its own soft spots is not more accurate — only less observable.
This page is generated from the metric registry and the measurement set, so it cannot disagree with the data it describes. See methodology for the full standard, the benchmark for every measurement, and the changelog for every correction — including the day the wall-plug figure moved from 0.01 to 0.0077.