FEATURE
After decades of work to make practical fusion power a reality, there’s still no agreement on what a commercial fusion reactor would look like or how it would work.
Written by Michael Abrams
“FUSION IS HARD.” Talk to the engineers and scientists working to perfect commercial fusion power, and they tend to undersell the challenge before them. “Fusion is hard,” they’ll say. Some of the best minds on the planet have worked on the problem over the past 70 years, with billions of dollars spent, increasingly large and imposingly complex facilities built to contain and compress hopelessly twisted high-temperature plasmas, all without reaching a sustained—and sustainable—net output in energy.
But while hard is not impossible, the word does convey that reaching the long-sought goal of practical fusion power will require more than some billionaires writing checks.
“It’s almost like a Manhattan Project scale of effort that’s required,” said Morgan Pattison, an engineer and material scientist working as a special advisor to Blue Laser Fusion. “And I don’t think that’s fully been appreciated by the people who want this to happen. It’s just a huge, huge endeavor. And especially if we want to do it fast, which we do. If it’s worth doing, it’s worth doing fast.”
Helion’s biggest stumbling block toward commercialization is finding enough talent to poperly accelerate its technology forward. Source: Helion
What the recent influx of investor money into the fusion space has done is provide funding for concepts that might otherwise have been neglected. The mainstream approach to building a sustainable fusion reactor involves building a doughnut-shaped magnetic bottle such as a tokamak or a stellarator, to hold a twisting, hundred-million-degree plasma long enough for racing hydrogen ions to smash together, make a whole new element, and release excess energy. A torus is a great shape for this sort of bottle, since the plasma can race endlessly through the middle of the doughnut without reaching a front or a back.
However, magnetic bottles can come in all sorts of shapes. And if the goal is smashing electrically charged hydrogen isotopes together, a bottle isn’t always necessary. So the fusion field has split apart—fissioned, if you like—into a number of directions. The thinking seems to be, one of them will pan out, and the payoff will be enormous.
“If you do deuterium-deuterium fusion, there’s 60 billion years’ worth of energy,” said Steven Cowley, laboratory director of the Princeton Plasma Physics Laboratory in N.J. “The earth is going to be swallowed by the sun in 4.5 billion, right?”
Helion uses field-reversed configurations to create two closed rings of plasma, each with its own private magnetic fields, at both ends of a linear system of magnets. The system then accelerates those rings toward one another to merge in a smaller chamber. Image: Helion
Rings of Power
Escaping from the tokamak mindset doesn’t always mean leaving tori behind. Helion, a company out of Everett, Wash., uses field-reversed configurations to create two closed rings of plasma, each with its own private magnetic fields, at both ends of a linear system of magnets. The system then accelerates those rings toward one another—at about half a million miles an hour—where they merge in a smaller chamber.
“We want to form them large so they last a long time, so we have a long time to do fusion, but if we tried to bring those large volumes to high field that would end up with a very expensive system. So we form them large, and then we merge them together,” said Chris Pihl, Helion’s co-founder and chief technical officer. “The whole purpose of that is really just to get them from the large radius section to the small radius section, so we don’t have to spend as much energy to get them to high field, which translates to dollars.”
Pihl compares the system to injecting a fuel air mixture into a diesel engine. Just as a piston compresses that mixture until it ignites, expands, and pushes back on the piston, in Helion’s smaller chamber, where the rings merge, magnets compress the plasma till it ignites and fusion occurs. And as they heat and expand, they push back on the magnetic field.
Helion’s sixth prototype, called Trenta, managed 10,000 pulses, achieved plasma temperatures of 100 million degrees Celsius, and worked for 16 months before it retired in 2023. According to a press statement from the company, the Trenta prototype proved it could “produce commercially relevant fusion conditions.” The current prototype, Polaris, has stronger magnets, will pulse faster, and, if all goes well, will make electricity.
Lifting a scrubber into place. Photo: Helion
Getting real-world data to feed into their models is key, Pihl said, since high-temperature plasma behavior is devilishly difficult to predict.
“Plasma behavior is governed by 11 independent partial differential equations. And the consequence of that is, it’s very difficult to model with high fidelity. So that’s part of the reason why fusion is experimental, why we build machines and operate them to get data,” Pihl said. “Trenta is very, very close to a full-scale device. And our models are very predictive at these sorts of incremental steps along the way.”
Pihl continued, “We’ve seen very good alignment with the steps we’re taking from one machine to the next. And the incremental step we are taking between Trenta and Polaris is no larger than any step we’ve taken in the past.”
Fusion may be hard, but from Helion’s perspective, the biggest stumbling block to a commercial reactor isn’t physics.
“Our biggest pacer right now is finding talent,” Pihl said. “You might think of a fusion company being science heavy, but the company is made up of about 25 percent scientists, about 25 percent engineers, and the balance is mostly technicians. Talented engineers and technicians would really accelerate us the most at this point.”
Polaris has already made a plasma and may soon be fully operational.
“What we’re shooting for is achieving what this machine was built for,” Pihl said, “which is demonstration of first electricity from fusion.”
Mirror, Mirror
One of the most mind-bending challenges for torus-based fusion is modeling the topography of the plasma as it courses through the donut. But the whole issue of how to twist a plasma in a torus disappears if you don’t use a torus in the first place. That’s the thinking behind Realta Fusion’s “magnetic mirror” approach. In short, instead of sending a plasma around a donut, they send it sloshing back and forth in a cylinder.
“The main principle behind it is that you use two or more very beefy, strong magnetic field magnets on either end of the ends of the cylinder,” said Derek Sutherland, Realta’s vice president of research and development. “What you’re trying to do is create a gradient in the magnetic field that’s strongest where the magnets are, and weakest the furthest away from them, along the cylinder.”
Charged particles in the cylinder bounce back and forth between the high magnetic field regions at speeds of 10,000 to 100,000 meters per second. (“They’re booking it,” Sutherland said). The simple tube is intended to keep the size, number of magnets, and, ultimately, the cost, to a minimum.
Instead of a torus, Realta Fusion’s solution sends plasma back and forth inside a cylinder. Video: Realta Fusion
The magnetic mirror concept has been around since the beginning of fusion research, but the inability to produce strong enough magnetic fields was a hard limit on the concept. Today’s high-temperature superconducting (HTS) magnets have made the concept feasible.
“As someone who has never worked on a superconducting device before, I still find it amazing that you just walk into the lab and the magnets are just always on—and they’re just sitting there making a magnetic field with no resistance in them. It’s just incredible,” Sutherland said.
The company is partnered with the University of Wisconsin and runs experiments at their magnetic mirror, the WHAM (for Wisconsin HTS Axisymmetric Mirror). Together. they sustained their first plasma in July of 2024. The data from experiments on WHAM feeds into a digital twin—RealTwin, Realta calls it—of their coming Anvil machine. This play between experiment and simulation is helping them to solve some of the thornier problems associated with plasma wrangling.
Realta Fusion and University of Wisconsin researchers, operating the Wisconsin High Temperature Superconductor (HTS) Axisymmetric Mirror (WHAM) experiment, formed and held a plasma with an applied magnetic field strength of 17 Tesla on the confined plasma, achieving a series of firsts for fusion energy. This demonstration was the first use of an HTS magnet in a magnetic mirror configuration. Photo: Mason Yu, University of Wisconsin-Madison
Plasma resists confinement, and bubbles of it easily fly off when shuttling from one magnetic field to the other. Putting a spin on the plasma holds it all together. Sutherland likens it to a dipper laden with honey.
“If you just hold it stationary, the honey will just drip off and fall back into the pot. But if you start spinning it in your fingers, you can keep the honey stuck to it, because as soon as it’s about to drop off the bottom of the honey dipper, you’ve rotated it such a way that the honey is back on the top. You can just keep doing that indefinitely, to keep the honey stuck to it,” he said. “When we simulated this, we actually saw this effect. And those simulations guided us in how fast we should spin the plasma in order to have this stabilizing effect.”
All that new knowledge is feeding how they’ll construct Anvil, an interim device that will help Realta iron out remaining physics questions before building a true fusion power plant.
“As you scale up closer and closer to commercial plasma conditions—meaning very high temperatures and densities—there’s basically more energy in the system available that could drive instabilities,” Sutherland said. The company thinks it will be ready to make a smaller, cheaper, fusion power plant by the end of the decade. “I think what you’re going to see, as time goes on in the fusion field, is that it’s not going to be about who gets the gain first,” he said. “It’s going to be about who makes the best product.”
“Plasma behavior is governed by 11 independent partial differential equations. And the consequence of that is, it’s very difficult to model with high fidelity.”
—Chris Pihl, co-founder and chief technical officer at Helion
Flash and Bang
Maintaining a stable plasma for minutes at a time while forcing it to make fusion is, obviously, the enormous challenge faced by tokamaks, stellarators, and other architectures. “It’s like those squeezy balls that kids have—you squeeze it in your fist and bits and pieces leak out through your fingers,” Blue Laser Fusion’s Pattison said.
One branch of fusion research has aimed at not maintaining a plasma at all and instead has pursued so-called inertial confinement fusion. That’s the concept employed by the National Ignition Facility, which first produced breakeven energy production from a fusion reaction in 2022.
“Shooting something with lasers just may be a little simpler and commercially feasible than trying to maintain control of the plasma,” Pattison said.
Blue Laser is doing just that. Located just outside Santa Barbara, Calif., the company was started in 2022 by a group of three founders, including Shuji Nakamura, who won a Nobel Prize for inventing the blue light-emitting diode. (That breakthrough resulted in the now ubiquitous white LED and a revolution in the lighting industry.) “I tell people he’s already put more energy on the grid than any other person living or in history,” Pattison said.
A concept image of Blue Laser Fusion’s core technology that combines a coherent beam combining (CBC) laser with an optical enhancement cavity (OEC). Image: Blue Laser Fusion
To put even more energy on the grid, he and his colleagues drew inspiration from Caltech’s LIGO gravitational wave detectors, the miles long vacuum tubes with laser interferometers used to detect ripples in spacetime. These cavities essentially amplify lasers. The Blue Laser founders realized that a similar design could be used to amplify lasers to heat and crush a deuterium-tritium fuel pellet.
Where NIF fires off a single enormous pulse (some two million joules in just billionths of a second) for a flash of high output in the name of science, Blue Laser wants to do it continuously to make commercial energy.
“The pain points might be the capacitor,” said Pattison of the NIF approach. “You’re dealing with the really high intensity pulse throughout your entire system. With our concept, we’re amplifying at a much higher rep rate and lower peak powers. The cavity effectively converts lower power but higher frequency pulses to lower frequency, higher power pulses. That’s where we can build up a pulse and then deliver it to the target.”
Blue Laser Fusion also differs in terms of its target. At NIF, the lasers blast a cylindrical case that instantly heats up and emits high-intensity X-rays, which both heat up and crush the fuel. Blue Laser skips a step and fires its beams directly at a fuel pellet.
Also, while it is taking design cues from LIGO, the Blue Laser vacuum cavities won’t be kilometers long but rather 150 meters; the company is working with Caltech to create them. “We have sub-scale prototypes that we’re working on, and we’re in the process of going from 1.5 meters to 15 meters to 150 meters,” Pattison said. “We’re not at the full energy levels that we hope to achieve, but we actually think we’re pretty close.”
Century is Zap’s fusion engineering test platform. Source: Zap Energy
Lightning Bolt
A lot of what makes fusion difficult—the magnets, the cryogenic cooling, the amplified lasers—gets side-stepped by using a Z-pinch machine. The Z-pinch effect has been known since the early 20th century when the magnetic field from a bolt of lightning was discovered to have crushed a lightning rod.
At Sandia National Laboratories Inertial Confinement Fusion, they have used the same effect for decades, firing 100-nanosecond shots of two million joules of energy out of massive banks of capacitors several hundred times a year. The long-term goal of the program is to produce high fusion yields of 200 million joules of output or more, primarily for the purpose of “replicating the kinds of temperatures, densities, and radiation fields that exist during operation of a nuclear weapon,” said SNL Director Gregory Rochau.
Understanding the conditions that occur within a thermonuclear detonation can also help advance the knowledge needed for fusion power.
To get closer to those conditions, the researchers at SNL have recently added tritium to their fuel, boosting the overall yield by a factor of 100. “That’s important because it allows us to get better diagnosis of the types of conditions and reactions that are taking place, which improves the science and improves our path forward,” Rochau said.
The Z machine is located in Albuquerque. Photo: Randy Montoya/Sandia National Laboratories
SNL has partnered with several private companies to further its fusion research. But Sandia’s Z-machine, or one like it, will never become a power plant.
“There are additional things you’d have to do in order to be able to create viable energy, beyond just what we do on Z, even if we produced a high yield fusion output with net facility gain,” Rochau said. “Once a day is sufficient to meet the needs of the weapons program. That’s not sufficient to put energy on the grid.”
Using a Z-pinch to put energy on the grid is just what Zap Energy aims to do, and with a lot more than a single shot a day. One of Zap’s founders, Uri Shumlak, a professor at the University of Washington, showed how, with a Z-pinch, a plasma can last 100 microseconds instead of 100 nanoseconds—a factor of 1,000 increase.
Under a Z-pinch, a plasma, like lightning, quickly becomes unstable, flying off in all directions. What Zap does is create a column of plasma with shear flow to keep it self-contained.
“Imagine you are driving in the middle of an eight-lane freeway at 65 miles an hour,” said Alex Cheung, Zap’s director of design integration and systems engineering. “If you have cars next to you that are only driving 10 or 20 miles an hour, you can easily change lanes. But if all the traffic next to you is going at 60 or 55, if you want to change lanes, it’s pretty hard.”
Zap’s Century machine has shown the ability to create that sheer flow plasma shot after shot after shot. It’s managed to create 1,000 plasmas, one after the other, over a course of three hours.
Sandia National Laboratories principal investigator Dean Rovang checks out the Z Machine’s tritium gas transfer system. Photo: Randy Montoya/Sandia National Laboratories
An overhead view of Zap’s Century machine. Photo: Zap Energy
The next step will require greater power and faster repetition rates. And, right now, Century is non-reactive. Once it’s running on real fuel, it will require shielding and radioactive containment. Zap’s engineers are already at work on the designs for that next step, as well as the liquid metal that would harvest the energy to operate a power plant.
“I will not say that it’s easy,” Cheung said.
Fusion is hard. And maybe these alternative paths won’t matter in the long run. Perhaps the conventional wisdom that a torus-shaped magnetic bottle is the right approach is actually correct. Or maybe photovoltaics matched with battery storage will make electricity so inexpensively that no other technology will be able to compete. But listening to Cheung, it’s easy to see why engineers and physicists keep pursuing these alternatives.
“From an engineering point of view there are a ton of challenges,” Cheung said. “On the plasma front, there will be a lot more instabilities. There will be other unknowns. By building one, we can find out.”
Michael Abrams is a technology writer in Westfield, N.J. He last wrote about fusion in the August 2025 issue of Mechanical Engineering.
© 2026 The American Society of Mechanical Engineers. All rights reserved.