Since December 2022 the National Ignition Facility has reached fusion ignition eleven times, peaking at a record 8.6 megajoules in April 2025. But that milestone counts energy on target — not the roughly 300–400 megajoules the lasers draw from the grid. On that measure, the best shot reached about 2–3% of break-even.
On 5 December 2022 the National Ignition Facility (NIF) at Lawrence Livermore National Laboratory became the first place on Earth to get more energy out of a controlled fusion reaction than the laser energy delivered to its fuel: 3.15 megajoules out, from 2.05 in.1 What matters almost as much is what came after. The result was not a one-off — by 20 June 2026, NIF had reached ignition eleven times.1
The yields have climbed unevenly. The eighth shot, on 7 April 2025, set the records that still stand: 8.6 megajoules of fusion energy (give or take 0.45) from 2.08 megajoules on target — a target gain of 4.13.1 That one shot produced more than six times the 1.35-megajoule yield that counted as a record as recently as 2021.1 The most recent shot, on 20 June 2026, reached 7.9 megajoules at a gain of about 3.8 — high, but below the April 2025 peak.1
The shots are not a smooth upward line, and reading them as one would mislead. They are separate experiments with different aims: some at lower yield by design, one led by Los Alamos, one built to test how nuclear materials behave under extreme conditions.1 The honest summary is narrower than a rising curve — ignition has been demonstrated, and then demonstrated again.
“More energy out than in” is true — but only for one definition of “in.” Target gain compares the fusion yield with the laser energy that reaches the fuel capsule, about two megajoules.1 It does not count the electricity the laser system pulls from the grid to fire in the first place. By the laboratory’s own accounting, that draw is roughly 300 to 400 megajoules per shot.2
Set the record shot against that figure and the picture changes. 8.6 megajoules out, against 300–400 in from the wall, is about 2 to 3% of “wall-plug” break-even — the grid draw was some 35 to 45 times the fusion energy produced.2 Independent accounts of the facility give the same order of magnitude.34
This is not something the laboratory hides. NIF was built for nuclear-weapons physics, not electricity, and its lasers were designed to deliver as much energy as possible in a single shot rather than to be efficient.2 Livermore staff put the lasers’ wall-plug efficiency at roughly 1%, and say a power-relevant laser would need to reach 10–20% and fire around ten times a second.2
Closing a gap of that size is an engineering problem as much as a physics one, and it runs through different machines. NIF fires only a handful of times a year, with long resets between shots;4 a power plant would need to hold that rhythm continuously. The nearer-term energy bets sit elsewhere: Commonwealth Fusion Systems is assembling a compact tokamak, SPARC, that it aims to push to net energy around 2027, while the international ITER project targets its first plasma in the mid-2030s.5 Both timelines have a long record of slipping.5
The exhibit below keeps what has been measured apart from what is merely targeted. The solid bar is the record of NIF’s eleven ignitions. The hatched bars are stated goals — scenarios, not events — drawn so the reader can see exactly where measurement ends.
Whether wall-plug or engineering break-even is reachable at all, and on what timescale, is unsettled — NIF was not designed to answer it.2 The repetition leap, from a few shots a year to roughly ten a second, is unproven at scale.2 Some figures carry real uncertainty: the April 2025 yield is 8.6 ± 0.45 MJ, the February 2024 shot was an estimate, and the laboratory did not publish the laser energy for the June 2025 shot.1 And cost is entirely open — no public figure establishes what fusion electricity would cost, and none is claimed here.
Two threads this piece set aside to hold length.