Nuclear fusion timeline: how close are we to fusion power, and what do breakthrough headlines mean?
A fusion experiment can make real progress and still be far from a power plant. The trick is knowing which milestone the headline is talking about.
The short answer: closer on physics than on electricity. On December 5, 2022, the lasers at the National Ignition Facility got more fusion energy out of a fuel target than they put into it, and in April 2025 the gain reached 4.13, yet no fusion machine has sent a single kilowatt-hour to the grid. The first attempt to sell fusion electricity, Helion’s plant for Microsoft, is scheduled for 2028. Grid-scale plants such as ARC in Virginia are early-2030s targets, and ITER plans to start deuterium-tritium fuel in 2039. The dated milestones are further down.
Fusion has a strange public rhythm. Every few years, a headline declares a breakthrough. Every few years, the familiar joke returns: fusion is always decades away. Both reactions flatten the story. Fusion research has made real progress in plasma control, superconducting magnets, high-power lasers, materials and diagnostics. At the same time, a moment of experimental gain is not the same as a power station sending electricity to the grid. The field is difficult because it has to do many hard things at once: create extreme conditions, hold them long enough, protect the machine, extract heat, breed fuel in some designs and run reliably. Reading fusion news well means asking which layer moved forward.
- No fusion electricity has reached the grid yet
- Do not confuse fusion with fission
- The temperature is not the only hard part
- Tokamaks try to hold plasma with magnetic fields
- Laser fusion compresses fuel in tiny bursts
- Net energy depends on what boundary you draw
- A plant has to do boring industrial work
- Climate urgency changes the question
- A ladder of milestones, and where the field is on it
- How close are we to fusion power? Dated milestones to watch
- Latest fusion breakthrough headline: which gain is it?
- How to read a fusion breakthrough headline
- Overreacting in both directions
- FAQ
This is for you if
- You see fusion breakthrough headlines and want to know what they really mean.
- You want to understand tokamaks, laser fusion and the gap to electricity without equations.
- You care about climate technology but dislike miracle language.
Skip this if
- You need a plasma physics course.
- You are looking for investment guidance on fusion startups.
- You want a yes-or-no date for commercial fusion.
No fusion electricity has reached the grid yet
Fusion is the process that powers stars: light atomic nuclei combine and release energy. On Earth, the challenge is not whether fusion can happen. It can. The challenge is making it happen in a controlled machine often enough, efficiently enough and safely enough to produce useful power.
That distinction matters. A lab shot, a plasma record and a power plant are different achievements. Headlines often blur them because fusion works sounds simpler than one subsystem improved.
Do not confuse fusion with fission
Fission splits heavy atoms. Fusion combines light ones. Fission plants already generate electricity. Fusion plants are still being developed. The waste profile, fuel cycle, safety behavior and engineering problems differ.
Fusion is attractive because it promises abundant fuel options and no runaway chain reaction like a fission reactor. It still produces activation in materials, requires complex fuel handling and faces huge engineering barriers. Cleaner does not mean simple.
The temperature is not the only hard part
People often hear that fusion requires temperatures hotter than the Sun. That is true, but temperature is only one variable. The plasma also has to be dense enough and confined long enough for reactions to matter.
The famous triple product combines temperature, density and confinement time. A reader does not need the formula to understand the point: fusion is a balancing act, not a single thermostat setting.
Tokamaks try to hold plasma with magnetic fields
A tokamak uses powerful magnetic fields to confine a ring of plasma. This is the most developed path toward magnetic confinement fusion. The machine is a careful negotiation among plasma stability, heating, magnets, wall materials and maintenance.
Progress can come from better magnets, better plasma control, better materials or better modeling. A headline about one of those is meaningful, but it is not automatically a complete plant.
Laser fusion compresses fuel in tiny bursts
Inertial confinement fusion uses lasers or other drivers to squeeze a tiny fuel capsule so quickly that fusion occurs before the fuel flies apart. This route is famous for dramatic ignition headlines. It proves important physics under extreme conditions.
A power plant version would need repeated shots, durable targets, efficient lasers, heat capture and an industrial supply chain for the fuel capsules. The experiment and the power station are related, but the engineering path is very different.
When you see the phrase net energy gain, always ask which boundary it uses. Energy released by the fusion against energy delivered to the fuel, or against the total electricity the whole machine pulls from the wall? The first has been done. The second is still far off, and many headlines quietly swap one for the other.
Net energy depends on what boundary you draw
Some milestones report more energy out of the fuel than the energy delivered to the fuel. That is important. It is not the same as more electricity out of the building than electricity consumed by the whole facility.
This is not trickery. Scientists define boundaries carefully. Headlines often do not. When you see net gain, ask whether it refers to the fuel capsule, the plasma, the laser, the magnets or the entire plant.
A plant has to do boring industrial work
A commercial fusion plant must run steadily, handle heat, maintain components damaged by neutrons, breed or supply fuel, convert heat to electricity and do it at a cost that competes with other low-carbon energy.
The phrase engineering problem can sound dismissive. It should not. Engineering is where laboratory physics becomes infrastructure. Fusion will be real for the grid only when these systems work together.
Climate urgency changes the question
Fusion could be a powerful clean-energy source, but climate decisions cannot wait for a technology that is not yet commercial. Solar, wind, storage, grids, nuclear fission, efficiency and demand management are available tools now.
That does not make fusion research a distraction. Long-horizon research is how future options are created. The mistake is treating future fusion as an excuse to delay present deployment.
A ladder of milestones, and where the field is on it
“How long until fusion” is hard to answer because it is not really a question about time. It is a question about a series of gates, and they have not all been passed. Lay the gates out in order and you can grade a headline yourself, by how far up the ladder it pushed things. What follows is an order, not a schedule; anyone handing you a precise year deserves a raised eyebrow.
| Rung | What passing it means | How you would know it happened |
|---|---|---|
| 1. Light the fuel | Plasma reaches the temperature and density where fusion actually occurs | Cleared long ago. Treat it as the starting line, not as news |
| 2. Target gain | One reaction releases more energy than was delivered onto the fuel itself | Reports carry a qualifier such as “scientific energy gain”. Check which layer is being counted |
| 3. Whole-machine gain | The machine releases more energy than it draws from the grid | Whether anyone dares publish the wall-plug figure |
| 4. Sustained operation | Not a few seconds, but running long enough to matter industrially | The vocabulary shifts from peak values to duration and repetition rate |
| 5. Materials survive | Chamber walls hold up under long neutron bombardment and can be replaced | People start discussing replacement cycles, downtime and irradiation data |
| 6. Fuel self-sufficiency | The plant breeds its own tritium instead of drawing down scarce stockpiles | Terms like tritium breeding ratio appear with measured figures attached |
| 7. Power on the grid | Heat becomes electricity, reaches the grid, and the economics hold up | Look for tariffs, grid agreements and kilowatt-hours actually delivered |
Roughly where things stand: rung one was cleared long ago, rung two has a widely reported result on the inertial confinement route, and everything from rung three upward is still experimental and engineering work. That is why one week can carry both “major breakthrough” and “still decades away” without either being dishonest. They are usually describing different rungs. Next time a timeline lands in front of you, ask which rung it crosses. The further up, the more the number is worth taking seriously.
How close are we to fusion power? Dated milestones to watch
The ladder above gives the order. These are the dates attached to it, as the projects themselves stated them in September 2026. The future ones are targets, and the last column says what each would prove if it arrives on time.
| When | Who | What happens | What it proves, and what it does not |
|---|---|---|---|
| 5 Dec 2022, done | National Ignition Facility (NIF), Lawrence Livermore, US | 3.15 MJ of fusion energy from 2.05 MJ of laser light on the target | Target gain above 1, rung two. The lasers draw about 300 MJ of electricity to fire a 2 MJ shot, so the facility as a whole was nowhere near break-even. |
| 7 Apr 2025, done | NIF | Record 8.6 MJ from 2.08 MJ, a target gain of 4.13; ignition reached for the 11th time on 20 June 2026 | Rung two can be repeated. Still one shot at a time. |
| 20 Jan 2025, done | EAST tokamak, Hefei, China | High-confinement plasma above 100 million degrees held for 1,066 seconds | Progress on duration, rung four. No net energy. |
| 2027, target | SPARC, Commonwealth Fusion Systems, US | Q above 1: more fusion power out of the plasma than the heating power put in | Plasma gain in a magnetic machine. SPARC is not built to make electricity. |
| 2027, target | BEST, Hefei, China | Machine completed, then burning-plasma experiments aiming at 20 to 200 MW of fusion power | A burning plasma in a compact tokamak. Demonstrating electricity is a later goal. |
| 2028, target | Helion, Orion plant, Chelan County, Washington, US | At least 50 MW for Microsoft under the first fusion power purchase agreement, initial operations in 2028; the generator building has been under construction since 2026 | The first real test of rung seven. Watch for kilowatt-hours actually delivered, not a ribbon-cutting. |
| Early 2030s, target | ARC, Commonwealth Fusion Systems, Chesterfield County, Virginia, US | A plant of about 400 MW; Google has agreed to buy half its power | Rung seven at commercial scale. |
| 2034, 2036, 2039, targets | ITER, France | Research operation from 2034, full magnetic energy in 2036, deuterium-tritium fuel from 2039 under its 2024 re-baseline; goal Q of 10 or more, 500 MW of fusion heat from 50 MW of heating | Plasma gain in a very large machine. ITER does not generate electricity. |
So, how close are we to fusion power? Gains inside the target and the plasma are being shown now. The first attempt to sell fusion electricity to a customer is scheduled for 2028 and has not happened yet. Grid-scale plants are early-2030s targets. A commercial fusion energy timeline you can trust will show up as delivered kilowatt-hours, not as another record shot.
Latest fusion breakthrough headline: which gain is it?
Most headlines that make fusion sound almost ready report one of three different ratios. Name the ratio and you know what the headline is worth.
| Ratio | What is divided by what | Best result so far | What it leaves out |
|---|---|---|---|
| Target gain | Fusion energy out ÷ laser energy that reached the fuel | 4.13 at NIF, April 2025 | The electricity the lasers pulled from the grid |
| Plasma gain, Q | Fusion power out ÷ heating power put into the plasma | 0.67 at JET in 1997 for a magnetic machine; SPARC aims above 1 in 2027, ITER aims for 10 | Magnets, cooling and turning heat into electricity |
| Wall-plug gain | Electricity the plant sends out ÷ electricity the plant draws | No device has passed 1. NIF draws about 300 MJ for a 2 MJ shot | Nothing. This is the one that matters for a power bill |
- Find which ratio the story uses. If it does not say, assume it is not wall-plug gain.
- Compare the number with the table. A target gain above 4.13 would be a new record; any wall-plug gain above 1 would be historic.
- Look for electricity: megawatts of electric output, a grid connection, kilowatt-hours sold. Fusion power in MW on its own is heat.
- Check the date against the list above. Is this the milestone promised for this year, or a slip from last year’s?
How to read a fusion breakthrough headline
Ask what moved: plasma duration, temperature, magnet performance, target gain, laser efficiency, material endurance, fuel breeding or plant design. Then ask what remains between that milestone and grid electricity.
Good fusion news can be worth celebrating without pretending your home will run on it soon. Cautious optimism is not pessimism. It is the only way to keep the milestones in view.
One more practical habit helps: read the institution behind the claim. A national lab, a startup, a university team and an international project may all use the word milestone, but they are usually measuring different things. Compare the claim with the group's previous result, not with a commercial plant in your imagination. That keeps excitement attached to the right scale.
| Claim | What it may mean | Question to ask |
|---|---|---|
| Ignition achieved | Fusion reactions were self-heating in a defined experiment | Was the whole facility energy-positive? |
| Record plasma time | Better confinement or control | At what temperature and density? |
| New magnet milestone | Stronger or more practical fields | Does it survive plant conditions? |
| Pilot plant announced | A development plan exists | Is it funded and what systems are demonstrated? |
| Commercial timeline | A company target | What assumptions does the date depend on? |
- Find the boundary behind any net-energy claim.
- Separate plasma physics progress from power plant engineering.
- Look for heat extraction, materials and fuel cycle details.
- Treat timelines as scenarios, not promises.
Overreacting in both directions
The physics is real and progress is measurable, but commercialization is hard.
A shot is a physics milestone, not an industrial power system.
It may help in the future. Current climate work needs current tools.
FAQ
Why is fusion so hard if stars do it?
Stars use enormous gravity. Earth machines must replace that confinement with magnets or rapid compression.
Is fusion safer than fission?
It has different safety characteristics and no runaway chain reaction, but it still requires careful fuel, radiation and materials management.
What is ITER?
ITER is a large international tokamak project in France intended to demonstrate key fusion plasma capabilities, not a commercial power plant. Under its 2024 re-baseline, research operation starts in 2034 and deuterium-tritium operation in 2039, aiming for ten times more fusion power than heating power. It will not generate electricity.
Should we fund fusion research?
That is a policy question, but the technical case is long-term optionality. It should not replace near-term clean energy deployment.
Which rung has fusion actually cleared?
Lighting the fuel was done long ago and target gain has a widely reported result on the inertial confinement route. Everything from whole-machine gain upward is still experimental and engineering work.
How close are we to fusion power?
Closer on physics than on electricity. Target gain above 1 was shown at NIF in December 2022 and reached 4.13 in April 2025, but no machine has yet sent out more electricity than it draws. Helion aims to start supplying Microsoft in 2028, Commonwealth Fusion Systems aims to run its ARC plant in Virginia in the early 2030s, and ITER plans deuterium-tritium operation from 2039.
Sources & further reading
Updated: June 2, 2026. A milestone ladder was added on September 8, 2026, breaking “how long until fusion” into seven checkable gates. On September 29, 2026 we added the dated milestones to watch through 2039, from NIF’s record shots to Helion’s 2028 target and ITER’s schedule, and a table of the three kinds of gain a headline can mean. On October 19, 2026 we put a short dated answer to “how close are we to fusion power” at the top of the page.
Fang Yu is the editor of FutureLens, turning published papers, official materials and public explanations into plain-language notes. He is most interested in the gap between a technology's public pitch and the evidence a careful reader can actually check. More about the author