The Future Is Here
We may earn a commission from links on this page

The Marathon Race Toward Fusion Power Could Be Reaching a Sprint

Nuclear fusion is the holy grail of energy research. Two experiments in France and Massachusetts do fusion differently, but both are making progress.

We may earn a commission from links on this page.
Technicians work on the bioshield inside the Tokamak Building at the International Thermonuclear Experimental Reactor (ITER) in Saint-Paul-les-Durance, southeastern France, on July 28, 2020.
Technicians work on the bioshield inside the Tokamak Building at the International Thermonuclear Experimental Reactor (ITER) in Saint-Paul-les-Durance, southeastern France, on July 28, 2020.

In September, a very large, incredibly powerful magnet was unloaded in Saint-Paul-lez-Durance in Southern France to be incorporated into the International Thermonuclear Experimental Reactor (ITER), a major international collaboration that is attempting to prove the technical feasibility of nuclear fusion.

Across the pond, within days of that massive magnet’s arrival, a team of researchers at the Massachusetts Institute of Technology, in collaboration with the private company Commonwealth Fusion Systems, announced their latest achievement in the race toward economical nuclear fusion: a successful test of their SPARC experiment, which runs on a relatively small, high-temperature superconducting magnet. These two magnet-driven experiments represent two approaches to fusion power, a holy grail of energy research.


Nuclear fusion has already been achieved. The record holder for controlled fusion power is held by a machine affectionately called JET, which produced 16 megawatts of fusion power in the late 1990s. The difficulty that physicists and engineers face now—and have been facing since fusion was achieved—is managing to get more power out of nuclear fusion reactors than is used by the machines to run the reactions.

Nuclear fusion is a reaction that produces huge amounts of energy, but it doesn’t occur naturally on Earth. If humans could safely and economically produce more energy from fusion reactions than it takes to power the reactions (and it takes a lot of energy to do that), we would no longer depend on carbon-based energy sources like coal, oil, and natural gas. But we’re getting ahead of ourselves.

A magnet.
The 110-ton magnet that recently arrived at ITER, one of an eventual six destined to be a part of the reactor’s central solenoid magnet. Sept. 9, 2021
Photo: Nicolas TUCAT / AFP (Getty Images)

Nuclear fusion describes a reaction that occurs when the light nuclei of two atoms fuse to form a single nucleus. In that process, an immense amount of energy is released. (This is Einstein’s E=mc2 in practice.) Things need to be extraordinarily energized for fusion to occur, which means they need to be really, really, hot, at 100 million degrees or more. Nuclear fusion is what makes the Sun shine, as hydrogen atoms combine to form helium, releasing energy in the process. If scientists could make that process work on Earth—and make the process work at scale—it would make energy a whole lot cleaner by cutting fossil fuels out of the equation.

“You can tell it’s it’s a hard problem, because people have been working on it for decades, with, you know, with really serious efforts, smart people, lots of money, big machines,” Martin Greenwald, a physicist at MIT’s Plasma Science and Fusion Center and a member of the MIT-CFS collaboration, said in a video call.

None of this is to be confused with nuclear fission, which is what drives today’s nuclear power plants and produces energy by splitting apart heavy nuclei. Nuclear fission produces less energy than nuclear fusion and generates radioactive waste products, which fusion does not.

A tokamak at Princeton University in 1975.
The Princeton Large Torus, photographed in 1975, which demonstrated that fusion was technically possible.
Photo: Wikimedia Commons

ITER and SPARC both rely on machines called tokamaks, first invented in the 1950s, which confine superheated plasma made up of particles that can interact to produce fusion reactions. Tokamaks are built in toruses, which is just a geometrist’s way of saying doughnut shapes. Tokamaks aren’t the only machines built for fusion: There are also stellarators, which are like tokamaks but more twisty. If a tokamak is a doughnut, a stellarator is a cruller.


The devices are built to generate magnetic fields, in order to contain the plasma that makes fusion possible. The magnets recently in the news (ITER’s super-large one and SPARC’s relatively small one) are part of the tokamaks and are used to confine the plasma, keeping it out of contact with ordinary matter. Inside a tokamak, the plasma is reminiscent of cotton candy being whipped into shape; over time, a delightful array of blues, purples, and pinks are what we can see of the ongoing physics. (This normally cannot be seen, but one tokamak—the COMPASS tokamak in Prague—has a camera installed inside.)

ITER’s magnet is one 110-ton module of the eventual six-module central solenoid magnet; when completed, the central solenoid will be the largest superconducting magnet ever built, with a field nearly 300,000 times as powerful as Earth’s magnetic field, according to the Department of Energy. The entire tokamak will weigh 23,000 tons. ITER’s goal is to produce 10 times as much fusion power as the power the machine needed to make it, but that’s just the fusion breakeven the machine is trying for—it doesn’t account for the amount of electricity put into the system to generate the initial plasma. So an entire operation that generates net power (what’s called engineering breakeven) is a loftier goal, one that ITER isn’t striving for.


With twice as strong a magnetic field, “you can have twice as small a device for the same performance,” said Ana Koller, a physicist specializing in nuclear fusion at the Max Planck Institute for Plasma Physics, in a video call. “But that track was pretty much until recently a dead end, waiting for a technological push from the superconductor side.”

As one might expect from extremely complex machines that take a long time to build, operate, and update, fusion experiments require something of a “constant duct taping,” according to Koller. As a big, international collaboration, ITER—first conceived 40 years ago—has had some delays along the way. ITER and the MIT-CFS teams are both racing toward the goals of running fusion reactions; ITER is currently expecting to run its first plasma in 2025, the same year that MIT-CFS expects SPARC will be complete. In turn, SPARC is laying the scientific ground for a pilot fusion plant called ARC, which could be operational by early 2030. Neither of these projects will actually put electricity on the grid; they are experiments that are attempting to prove the principle of fusion itself, that more power can be produced than was used to generate the reaction.


“This is not that kind of a race where the end goal is to humiliate and completely obliterate your opponent,” Koller said. “This is a race where we get diversity in fusion research that we can work with in the future—so not the one where the winner takes it all.”

Greenwald has been working on nuclear fusion for the better part of 50 years, but the recent technological innovation on the MIT team’s side is something of a watershed moment. “The idea of using high-temperature superconductors to get to higher field magnets had sort of been in our DNA,” he said, but until the recent engineering breakthrough, the team didn’t know how they’d manage it.


Now that they have, the MIT-CFS team is moving full-steam ahead with SPARC, the technology demonstrator for the eventual ARC reactor. ARC will be built the same way, with layers of flat superconducting material stacked on top of each other and chilled to 20 kelvin to generate a magnetic field. ARC hinges on SPARC’s ability to prove the concept. “The next step is to go a little bit larger,” Greenwald said, “to a whole facility, which is making net power.”

While ARC will be about twice as large as SPARC, it’s still way smaller than ITER, which was built with a larger vessel to hold more plasma and thus increase the likelihood of fusion reactions. The MIT-CFS team’s new magnets make it possible for devices to perform similar amounts of fusion to a machine 40 times larger in volume, according to an MIT release.

This is all easier said than done, of course, which the many years of research is testimony to. And if fusion is ever to become a useful path toward a cleaner energy future, it’ll need to be relatively cheap and scalable.


“We are physicists, we have to be skeptical in absolutely everything,” Koller said, “but we need that dose of optimism” in order to do the work.

When it comes to fusion, there are evangelists and skeptics and purported realists, though it can be hard to draw a bead on whose expectations most closely meet reality. The running joke about fusion power is that it’s always 30 years away, somewhere just beyond the scientific horizon. Whether you believe fusion is a pipe dream or nearly upon us, the progress that’s been made is undeniable. And maybe—just maybe—we’ll see some of that progress realized within the next decade.


Update: This post has been updated to clarify that the tests of the new reactors could prove the technological feasibility, not necessarily the economic viability, of fusion power, and to emphasize that neither SPARC nor ITER will be putting electricity back into the grid; both are experiments hoping to demonstrate fusion power is possible.