CyphrQ
Menu

Fusion Energy Progress: What NIF Ignition and ITER Actually Show

NIF has reached fusion ignition many times since 2022. Here is what that proves, what it doesn't prove about power plants, and what ITER's dates say.

By CyphrQ.ai

Written with AI assistance and human editorial review. Every figure linked to its source.

10 min read

Fusion energy progress is real and measurable: since December 2022, the US National Ignition Facility has repeatedly got more fusion energy out of a fuel target than the laser energy put into it, with a record of 8.6 MJ out for 2.08 MJ in on 7 April 2025, according to Lawrence Livermore National Laboratory (LLNL). That is a physics milestone. It is not yet a power plant, and the gap between the two is large.

This post sets out what has been shown, what has not, and what to watch next. Every figure comes from the lab or organisation that produced it.

What did the National Ignition Facility actually achieve?

The National Ignition Facility (NIF) at Lawrence Livermore National Laboratory (LLNL) fires powerful lasers at a tiny capsule of hydrogen fuel. The capsule is crushed and heated until its nuclei fuse and release energy. "Ignition" here means the fusion reactions released more energy than the laser light delivered to the target.

LLNL's ignition page gives the first result: on 5 December 2022 the experiment produced "3.15 megajoules (MJ) of fusion energy output from 2.05 MJ of laser energy delivered to the target."

It was not a one-off. LLNL lists later shots that also crossed the line:

Date Laser energy on target Fusion yield What LLNL notes
5 Dec 2022 2.05 MJ 3.15 MJ First ignition
30 Jul 2023 2.05 MJ 3.88 MJ Repeat
12 Feb 2024 2.2 MJ about 5.2 MJ (estimated) More than double the input
23 Feb 2025 2.05 MJ 5.0 MJ Target gain 2.44
7 Apr 2025 2.08 MJ 8.6 MJ (±0.45 MJ) Record yield, target gain 4.13
1 Oct 2025 2.065 MJ 3.5 MJ (±0.17 MJ) Tenth ignition
20 Jun 2026 not stated 7.9 MJ (±0.4 MJ) Eleventh ignition, gain about 3.8

LLNL also says the 8.6 MJ result was more than six times the 1.35 MJ record set in 2021. The table shows shots as listed on LLNL's page on 26 October 2026; it leaves out several, and the page is updated as new shots happen, so check it for anything later.

In plain terms: the result has been repeated, and the best yields have grown. That matters in science. A single result can be luck; a series is evidence.

What does "ignition" not mean?

Ignition is measured at the target. It compares the fusion energy released with the laser energy that reached the capsule. It does not count the electricity the facility used to make that laser light.

LLNL has been open about this. At an IAEA webinar reported by LLNL in March 2023, LLNL's Jean-Michel Di Nicola said that for the ignition shot "NIF drew about 300 to 400 megajoules out of the electrical grid." He also said: "NIF was not designed to be energy efficient; it was designed to demonstrate scientific break-even."

So the 2022 shot returned 3.15 MJ of fusion energy for a few hundred megajoules of electricity. My arithmetic from those two figures: the fusion output was roughly 1% of the grid draw. That is not a criticism of the experiment, which was never meant to be efficient. It shows why "ignition" and "power plant" are different claims.

Ignition also does not mean:

  • Continuous output. NIF fires single shots. A plant would need to fire many times a second, without stopping.
  • Electricity. No turbine or generator is attached. The energy is measured, not used.
  • A cost. Nothing in the results says what fusion electricity would cost.

How far is fusion from a working power plant?

The same LLNL article gives a sense of scale. LLNL's Abbas Nikroo said a laser-driven fusion plant would need to fire "at about 10 Hertz (10 times a second)" and "we would need to shoot at about a million times a day." He added: "I think we would need an energy gain of about 100 or so."

Compare that with the record target gain of 4.13. The gain has to rise by a large factor, the targets have to become cheap and mass-produced, and the lasers have to become far more efficient. Di Nicola said modern lasers can reach "10 to 20 percent wall-plug efficiency," which is far better than NIF's, but a plant would still need all of these at once.

None of the sources I read gives a date for laser-driven fusion electricity on a grid. I won't offer one.

Where does ITER fit in?

ITER is a different approach. Instead of crushing a capsule with lasers, it holds hot, electrically charged gas (plasma) in place with powerful magnets inside a ring-shaped machine called a tokamak. It is being built in southern France. ITER's overview lists the ITER Members as "China, the European Union, India, Japan, Korea, Russia and the United States."

Two European countries outside the EU are worth a note. The page says that "following Brexit, and the subsequent withdrawal from Euratom," the United Kingdom "discontinued its participation in the ITER project." Switzerland renewed its association with EU programmes, including Euratom Research & Training, as of 1 January 2026, and became a member of Fusion for Energy, the EU's body for ITER.

ITER's stated goal is "500 MW of fusion power from 50 MW of input heating power", a tenfold return written as Q=10. The page is also clear on one point people often miss: "ITER will not convert the heating power it produces as electricity."

ITER is an experiment to show the physics and engineering at power-plant scale. Electricity is for the machines after it.

When will ITER start fusion experiments?

ITER's overview page lists this timeline:

  • 2034: Start of Research Operation
  • 2036: Full magnetic energy
  • 2039: Start of deuterium-tritium operation

Deuterium and tritium are the two forms of hydrogen that fuse most readily, so 2039 is when ITER plans to run with real fusion fuel. ITER describes the 2039 date as four years later than its 2016 plan.

ITER's July 2024 baseline article put the extra cost to the ITER Organization at "EUR 5 billion," still under review by members at the time, and called the new schedule a proposal being used as a "working reference."

Since then, the ITER Council has moved it forward. At its November 2024 meeting, the Council said its delegations "endorse the overall approach proposed for the new project baseline (Baseline 2024)," while "requesting continued efforts to reduce risks and optimize costs." That release does not say the cost was approved. A year later, the November 2025 Council release said that "for 2024 and 2025, the project's Schedule Performance Index and Cost Performance Index have been greater than 1.0," which ITER reads as running slightly ahead of schedule and under the Baseline 2024 spending plan. As of 26 October 2026, I found no later Council release. Treat the dates as ITER's endorsed plan, not a guarantee.

What has fusion research not shown yet?

Based on the sources above:

  1. Net electricity. No fusion device in these sources has produced more electricity than it used.
  2. Repetition. NIF fires single shots. A plant needs about ten per second, per LLNL's own researchers.
  3. Gain at plant scale. The record target gain is 4.13. LLNL's estimate of what a plant needs is about 100.
  4. ITER's fuel phase. Deuterium-tritium operation is planned for 2039.
  5. Fuel supply. A plant running on deuterium-tritium needs a steady tritium supply. ITER plans to test "mockup in-vessel tritium breeding blankets," which are designs for making tritium from lithium inside the machine. That test has not happened yet.

What fusion energy progress should you watch next?

  • NIF's shot list. LLNL updates its ignition page. Watch whether yields above 8 MJ become routine and whether gain keeps rising.
  • ITER's milestones against its own dates. A schedule is a promise. The 2034 start of research operation is the next big one.
  • Laser efficiency and target cost. These decide whether laser fusion can become a plant. Look for figures from labs, not slogans.
  • Private company claims. Many companies are working on fusion. Read their claims as "the company says" until a lab, regulator or peer-reviewed paper confirms them. Not investment advice.

Going deeper: for security and IT teams

This section is written for technical readers and decision-makers, including IT and infrastructure leads who get asked whether fusion should feature in energy or data-centre plans. The short answer is no, not for any planning horizon these sources support. The longer answer is in how the gains are defined.

Three different gains, often confused

Term What it divides Best published figure here
Target gain Fusion energy out ÷ laser energy on the target 4.13 (NIF, 7 Apr 2025)
Plasma gain, Q Fusion power ÷ heating power into the plasma ITER's goal is Q≥10 (500 MW from 50 MW)
Facility or "engineering" gain Electricity out ÷ electricity in for the whole plant Not achieved in any source read here

Headlines usually report the first or second and readers assume the third. When someone says "net energy," ask which line of this table they mean.

Worked example from LLNL's own figures

For the December 2022 shot: 300 to 400 MJ drawn from the grid, 2.05 MJ of laser light delivered, 3.15 MJ of fusion energy released. Dividing those figures, the laser light was about 0.5% to 0.7% of the grid draw, and the fusion yield about 0.8% to 1%. That is my calculation from LLNL's published numbers, not a figure LLNL published. LLNL did not publish a grid draw for the 8.6 MJ shot, so I have not computed one.

What a laser-fusion plant needs, in LLNL's words

  • Repetition at about 10 Hz, or about a million shots a day.
  • Target gain of about 100.
  • Lasers at 10% to 20% wall-plug efficiency.
  • Mass-produced, low-cost targets, implied by the shot rate.

Each is a separate engineering programme. Progress on one does not carry the others.

What ITER's Q=10 does and does not cover

Q compares fusion power with the heating power put into the plasma. It leaves out the electricity used by magnets, cooling and other plant systems, and ITER will not generate electricity at all. ITER's value is in proving burning plasma physics, long pulses at full magnetic energy, and tritium breeding blanket tests at scale.

How to read a fusion claim (questions to ask)

  1. Which gain is this: target, plasma or facility?
  2. Was it measured, with an uncertainty, or modelled?
  3. Has it been repeated, and is the shot list public?
  4. What fuel was used? Deuterium-tritium results are the ones that matter for plants.
  5. Who confirmed it: the company, a national lab, a regulator or a peer-reviewed paper?

Planning implication

If you plan power for data centres or large IT estates, plan on today's generation, storage and grid connections. For how fast that load is growing, see the IEA figures in AI and data-centre power. ITER's own schedule puts fusion-fuel experiments in 2039, and electricity comes after that.

What I did not verify. I did not read peer-reviewed papers on individual NIF shots or the US Department of Energy's fusion roadmap for this post. Figures are as published by LLNL and ITER.

What should you do when a fusion headline lands?

  1. Use the right word. Say "ignition" or "target gain" for NIF results, and keep "net electricity" for the day a plant produces it.
  2. Check the source page. LLNL's shot list and ITER's timeline are public and current.
  3. Ask the five questions above before passing on a fusion headline to your board or team.
  4. Keep fusion out of near-term energy plans. Treat it as long-term research to follow, not capacity to count on.
  5. Revisit yearly. Set a reminder to check NIF's yields and ITER's milestones against this post.

Key takeaways

  • NIF first reached ignition on 5 December 2022: 3.15 MJ out from 2.05 MJ of laser energy on target, per LLNL.
  • The record is 8.6 MJ from 2.08 MJ on 7 April 2025, a target gain of 4.13.
  • The 2022 shot drew about 300 to 400 MJ from the grid, according to LLNL's Jean-Michel Di Nicola. Ignition is not net electricity.
  • LLNL researchers estimate a plant needs about 10 shots a second and a gain of about 100.
  • ITER aims for 500 MW of fusion power from 50 MW of heating, will not make electricity, and plans deuterium-tritium operation from 2039.

Sources