A New Fusion Prototype Floats Into Action

Many fusion energy startups talk about trying to replicate the awesome power of the sun. But only one is also trying to replicate the sun’s structure, by having the reactor’s most important component—a powerful magnet—float in a vacuum, surrounded by a ball of thermonuclear glowing gas.

Two weeks ago, OpenStar Technologies achieved “first plasma”—a cloud of ionized helium contained by a superconducting magnet suspended at the center of a prototype device called Junior, in New Zealand.

OpenStar is still years away from producing fusion, let alone net power gain, but its founder and CEO Ratu Mataira told IEEE Spectrum that its design, which will eventually include that magnet hovering meters above the ground, might be humanity’s best shot at achieving commercial-scale fusion energy.

“It’s the only fusion configuration that nature doesn’t want to destroy immediately at all times,” he said.

Making a Miniature Sun

The sun undergoes fusion by virtue of its immense mass, with gravity forcing hydrogen nuclei to fuse into other particles and release a burst of energy. To replicate that process on Earth, some researchers and startups are using magnetic fields to squeeze hydrogen nuclei together.

A traditional doughnut-shaped tokamak reactor uses magnets outside the reactor to shape the plasma inside, but this plasma configuration is notoriously difficult to create and maintain. OpenStar’s idea is to use a dipole magnet at the center of the reactor instead. Dipole magnets are common in nature: Earth’s own magnetic field that protects us from stellar radiation is an example. A dipole’s big advantage is that they are naturally stable, potentially opening the door to sustained fusion reactions.

“We solve a lot of complication by moving to the dipole,” said Mataira. “The one catch that it makes the magnet look significantly harder to do.”

Placing a superconducting magnet that requires chilling to a few tens of degrees above absolute zero in the midst of a fusion reaction at 175 million °C might sound insane. But there is some physics working in OpenStar’s favor: Mataira points out that convection in the plasma will drive heat outward toward the reactor walls, rather than inward toward the magnet. “It doesn’t kill the concept of having a magnet inside,” he said.

“It’s the only fusion configuration that nature doesn’t want to destroy immediately at all times.” —Ratu Mataira, OpenStar Technologies

In the experiment last week, the magnet was precooled to about 30 kelvins (-240 °C), which gave OpenStar an 80-minute window until it warmed up and could no longer provide the necessary magnetic field. In future iterations, the magnet will have onboard cooling from liquid helium to extend that duration, although it will still need to be shut down periodically to be cooled again.

There are other limits to Junior’s operation. Although the high-temperature superconductors in the magnet have extremely low electrical resistance, it is not zero. As the reactor operates, the magnet will gradually lose energy. Superconducting magnets in tokamaks have the assistance of a wired power supply, but running cabling through the middle of a fusion reaction would be impossible.

Instead, future versions of the central unit will contain batteries to extend the life of magnet. And then there is Junior’s most sci-fi feature of all: The entire magnet unit will levitate in the center of the reactor, thanks to a permanent magnet that will be installed above the vacuum chamber.

“This is one of the least intimidating aspects of the build,” said Mataira. “That’s because this machine is benchmarked off the LDX levitated dipole experiment at MIT. They levitated for about six years, and we have the former chief experimentalist of that program working with us.”

LDX shut down in 2014, having proven some of the basic science and engineering behind levitated dipole reactors but never achieving fusion.

Mataira expects Junior will tread a similar path. A successor reactor will have a magnet that is the same size of Junior’s 1.2-meter-wide dipole but with four times the strength. That will help OpenStar study radio-frequency technologies to heat the plasma. However, Mataira is not expecting OpenStar to achieve fusion until at least its third generation of devices, possibly as soon as 2027.

The Long Road to Commercial Fusion

Commercial OpenStar reactors, when they arrive perhaps next decade, promise to be smaller while still generating net power, compared to enormous tokamaks. “We think that dipoles will actually be quite small in terms of megawattage to be power positive,” said Mataira. “That would be a dream scenario because we can imagine deploying 25- to 50-megawatt units for data centers in remote locations, while we develop the multigigawatt units that are going to be really important for the overall climate transition.”

There are a host of engineering challenges between firing up Junior and selling power, notes Andrea Di Vita, a plasma physicist who worked on the Joint European Torus (JET) fusion program in the United Kingdom. “From the engineering point of view, the main issue is protection from thermal loads,“ he said. “But dipole can be a game changer. The plasma pressure for a given magnetic field is 13 times higher in a dipole machine than in a tokamak. And the power from the fusion reaction, if you can get there, is more than 150 times larger.”

Mataira is realistic about the road ahead. “OpenStar is significantly less mature than other fusion programs in a lot of ways, and there might be some big, bad problem around the corner that’s going to kill the idea,” he said. “But if you look for plasmas that you can produce in a lab at a steady state that also exist in nature, there is actually only one candidate. And that’s a levitated dipole plasma.”

This article appears in the February 2025 print issue as “A New Fusion Prototype Achieves First Plasma.”