How close is nuclear fusion power?

Posted: December 20, 2024. Updated: March 17, 2026

Jump to:

• Kinds of nuclear fusion devices
• The first 100 years of fusion
• The major players
• 2026 state of the industry

Fusion researchers are all too aware that many people have become disillusioned by what seem to be perennial claims that practical fusion power is just 10 or 20 years away. If scientists can control nuclear fusion, it could produce four times as much energy per kilogram as nuclear fission, and nearly four million times as much as coal or oil. In principle, a few grams of hydrogen could produce all the energy a person living in a developed country would use in sixty years.^[1] Better still, it does not produce the radioactive waste or risk of meltdowns that come with nuclear fission: Its main byproduct is inert helium.

Kinds of nuclear fusion devices

But controlling nuclear fusion is difficult. To get the protons and neutrons within a hydrogen atom to fuse together, you have to get them close enough for the strong nuclear force to overcome the electrostatic forces that normally push them apart. The sun uses its immense gravity to do that. But, to push hydrogen atoms together in the lab, you have to heat them to extremely high temperatures—about 100 million degrees C (180 million F)—some ten times the temperature of the inside of the sun. At those temperatures, hydrogen becomes plasma, a state of matter in which the electrons separate from the protons and neutrons of the atoms’ nucleus.

Plasma at these temperatures will vaporize any material we might try to use to contain it. One of the main challenges to producing controlled nuclear fusion is figuring out some way to confine this super-hot plasma for long enough that its atoms can start to fuse and we can harness the energy that fusion gives off.

Two main methods of containing plasma show the most promise as candidates for creating a viable commercial fusion reactor: magnetic confinement and inertial confinement.

Magnetic confinement devices

Magnetic confinement devices use magnetic fields to confine plasma within a “magnetic bottle”—a difficult task some scientists liken to trying to hold together Jell-O with rubber bands. Researchers have developed four main kinds of magnetic confinement devices:

• Pinch devices contain plasma with magnetic fields generated by running electrical current through the plasma itself.
• Stellarators use powerful magnets to create magnetic fields that confine plasma in the shape of a complex doughnut, or torus.
• Tokamaks combine the technologies of pinch and stellarator devices, confining plasma in a torus shape using both external magnets and fields generated by current within the plasma.
• Magnetic mirrors bounce plasma back and forth between two high-density magnetic fields.

Of these four designs, tokamaks have become the leading magnetic confinement devices, though recently, private startups have revived both pinch devices and stellarators. In 2021, the JET tokamak in Oxford produced a record 59 megajoules (MJ) of energy from nuclear fusion—enough to boil about 60 kettles of water.^[2]

The major players in fusion power

Perhaps the most ambitious fusion project is currently ITER, a plan to build the world’s largest and most powerful tokamak in southern France. A consortium of countries including China, the EU, UK, U.S., India, Japan, Korea and Russia are collaborating to use the tokamak to produce 10 times the amount of energy needed to initiate fusion—enough for a viable power plant. Though the project was initially scheduled to be operational by 2025, it now expects not to produce fusion until 2039.

While the National Ignition Facility has been so far the only facility to achieve energy breakeven—doing so several times since 2022—its IFC technology needs to resolve many engineering problems before it’s able to generate power on a commercial scale. Right now, it can fire its lasers every few hours, but a viable IFC power plant would need to fire its lasers 10 times every second.^[3]

A number of private fusion startups have also sprung up in the last couple years. The Fusion Industry Association estimates that private fusion firms have raised $7.1 billion in total investment as of July 2024, with nearly a billion of that coming in over the last year.^[4]

• Commonwealth Fusion Systems, which came out of MIT, has raised more funding than any other startup: $2 billion since its founding in 2018. It has developed the world’s most powerful high-temperature superconducting magnets, which it will use to build a tokamak named SPARC.^[5] It recently announced plans to build the world’s first grid-scale commercial fusion power plant in Virginia.^[6]
• Pacific Fusion is using a novel technology called “pulsed magnetic fusion,” which uses magnetic pulses instead of lasers to inertially contain plasma.
• A number of other startups are reviving older fusion technologies. Zap Energy is building a pinch device, Type One Energy and Thea Energy are building stellarators, and Realta Fusion is building magnetic mirror devices. The ENN Group, one of China’s largest private companies, is working on tokamak variations called “compact toroid” magnetic confinement devices.
• Several companies are developing innovative new twists on traditional fusion devices. OpenStar is creating a reverse tokamak: instead of surrounding plasma with magnets, it levitates a magnet within the plasma itself. General Fusion plans to use magnets to contain plasma while it compresses it with steam-driven pistons.

OpenStar CEO Ratu Mataira remains clear-eyed but optimistic, telling CNN, “Not all of the fusion companies will be successful. OpenStar might be one of those, but we as a society will learn faster.”

The state of the nuclear fusion industry in 2026

The Lawrence Livermore National Laboratories National Ignition Facility (NIF) remains the only facility to have achieved fusion “breakeven”: getting more energy out of a fusion reaction than went into it.  But the amount of power it can produce is increasing. Its first breakeven reaction output 1.5X more energy than it used to power the reaction. The latest reaction, in April of 2025, resulted in over 4X the energy used to power it.

While the accomplishment is impressive, commercially viable fusion power will need to generate 50-100 times as much energy as that used to power the reaction says Debbie Callahan, a long-time physicist at NIF, now Chief Strategy Officer at fusion startup, Focused Energy.

Meanwhile, the number of private startups pursuing nuclear fusion continues to grow. The number went from 31 in 2022 to 45 in 2024 and has now grown to 53 as of 2025. Private investors have been putting over $2B per year on average into the industry over the last five years.

Many expect that the proliferation of private fusion enterprises will speed up solutions to the practical engineering challenges that now seem to be the biggest obstacles to commercial-scale fusion power. Over the last year, teams have made real progress on some of these challenges.

In January, China's EAST (Experimental Advanced Superconducting Tokamak) fusion reactor announced that it had controlled the interactions between plasma and reactor walls to keep plasma stable at densities beyond what’s called the Greenwald limit, at which plasma usually becomes unstable.

In October, American corporation, General Atomics, partnered with NVIDIA to create a digital twin of the DIII-D tokamak in San Diego, a kind of fusion device that uses powerful magnets to contain super-heated plasma within a large doughnut-shaped chamber. This virtual reactor combines sensor data, engineering models and physics-based simulations so that an international team of 700 scientists can test out “what-if” scenarios without risking damage to the physical reactor.

Sam Altman-chaired Helion announced in February that it had achieved plasma temperatures 10 times the heat of the sun. It has yet to initiate a fusion reaction, but it has an ambitious goal to supply power for Microsoft data centers by 2028.

Many experts are skeptical about Helion’s timeline, but other fusion initiatives claim to be on track to initiate fusion in the 2030s.

The considered consensus of energy scientists, however, seems to be that technological advancements have made fusion much more likely, but that it won’t be ready for widespread power generation until at least after 2050 or 2060.^[7]

^{[1] https://www.iaea.org/newscenter/news/what-is-nuclear-fusion
[2] https://www.bbc.com/news/science-environment-60312633
[3] https://lasers.llnl.gov/news/fire-powers-universe-harnessing-inertial-fusion-energy
[4] https://www.fusionindustryassociation.org/fia-launches-2024-global-fusion-industry-report/
[5] https://www.nytimes.com/interactive/2024/11/15/climate/fusion-energy.html
[6] https://virginiamercury.com/2024/12/18/virginia-to-host-worlds-first-fusion-power-plant/
[7] https://www.scientificamerican.com/article/what-is-the-future-of-fusion-energy/}