Has fusion ever produced more energy than it used?

Three ways to count

The confusion comes from the fact that “more energy out than in” can be measured at three different boundaries.

• At the fuel. How much fusion energy comes out compared with the heating energy that actually reaches the plasma or the fuel target. In magnetic fusion this ratio is called Q.
• At the facility. How much fusion energy comes out compared with all the electricity the machine draws – for lasers or heating systems, magnets, cooling, pumps and control.
• As electricity. Whether a plant turns that fusion energy into electricity and sends more to the grid than it takes. No fusion device does this today, and none of the experiments described here was built to.

Each record in the news belongs to one of these levels. Most of them belong to the first.

The laser record: ignition at the target

The National Ignition Facility at Lawrence Livermore National Laboratory uses powerful lasers to compress a small capsule of deuterium and tritium until it fuses. This is inertial fusion, a different route from the magnetic confinement used in tokamaks. In the experiment of 5 December 2022 the lasers delivered 2.05 megajoules to the target and the fusion reactions released 3.15 megajoules – a target gain of about 1.5. Livermore notes that producing more fusion energy than the laser energy delivered to the target is the definition of ignition used by the US National Academy of Sciences in a 1997 review of NIF.

The result has been repeated and improved. In 2023 Livermore reported further ignition shots, including 3.88 megajoules on 30 July. A record of 5.2 megajoules followed in February 2024, and on 7 April 2025 an experiment produced 8.6 megajoules from 2.08 megajoules of laser energy, a target gain of 4.13 (status as of 3 November 2026).

The same researchers are explicit about the limits. A target gain above one, they write, “does not imply net energy gain from a practical fusion energy perspective”, because the energy consumed by the NIF laser facility is typically 100 times larger than the laser energy delivered to the target. NIF was not optimised to produce net energy; it is a scientific facility.

Magnetic fusion: still below one

In tokamaks and stellarators the fuel is heated by microwaves, particle beams and electric current, and Q compares the fusion power with that heating power. According to ITER, the world record for fusion power in a magnetic confinement device is still held by the European tokamak JET: in 1997 it produced 16 megawatts of fusion power from 24 megawatts of heating, a Q of 0.67.

No magnetic confinement experiment has yet reached Q = 1. ITER, under construction in southern France, is designed to be the first to produce net energy gain across the plasma and, beyond that, to yield a ten-fold return: 500 megawatts of fusion power from 50 megawatts of input heating power. Even then, ITER says plainly that it will not convert the heat it produces into electricity. Its research programme begins with hydrogen and deuterium plasmas; the Q = 10 goal belongs to a later phase with deuterium and tritium.

Energy records that are not gain records

Some records measure something else entirely, and are easy to misread.

In its final deuterium–tritium experiments JET released 69 megajoules of fusion energy in a pulse of about five seconds, using only 0.2 milligrams of fuel; the UK government announced the result on 8 February 2024, after JET had ended its scientific operations at the end of December 2023. That is a record for the total amount of fusion energy, and it showed that high fusion power can be sustained for seconds. It is not a statement that JET produced more energy than it was fed: its best ratio remains the 0.67 of 1997. The figures also differ slightly between sources – the Dutch institute DIFFER, a EUROfusion partner, gives 69.26 megajoules and 0.21 milligrams.

The stellarator Wendelstein 7-X reported an energy turnover of 1.8 gigajoules in 2025. IPP defines this as the injected heating power multiplied by the plasma duration – energy put in, not fusion energy out. The record shows that the machine can keep a hot plasma going for a long time, here 360 seconds, which is a different and also necessary step (see Tokamak or stellarator: what is the difference?).

Where machines like TCV fit in

Most fusion experiments are not trying to set gain records at all. Medium-sized tokamaks such as TCV in Lausanne test how to shape, control and cool a plasma – questions that have to be answered before a machine like ITER can work as planned. EUROfusion describes them as complementing and confirming the results of larger machines and broadening the experimental database. TCV’s best-known results, such as new plasma shapes and control by artificial intelligence, are described on TCV.

ITER itself notes that the cost of electricity from fusion cannot yet be estimated. The arguments for why fusion is pursued at all are summarised on How a tokamak works.

How to read the next headline

When a new fusion record is announced, three questions sort it out quickly. Which energy is being compared with which – fusion output with the heating that reached the fuel, or with everything the facility consumed? Is it magnetic fusion (Q) or laser fusion (target gain)? And is it a record for gain, for total energy, or for duration? The answers rarely make the result less impressive, but they show how far it is from a power plant.

Polish laboratories take part in this work too, including on JET and TCV: see How does Poland take part in European fusion research?