Few fusion startups have been as closely watched as Helion. The 12-year-old company is backed by Sam Altman, rumored to be in talks with OpenAI, and has a deal to supply Microsoft with electricity by 2028 — years earlier than its competitors.
The company’s unorthodox approach to fusion power and relative secrecy has earned it plenty of fans — and critics. But don’t count its investors among the naysayers.
Helion announced Tuesday a $425 million Series F raise that pushed its valuation $5.245 billion. The startup also flipped the switch last month on its latest prototype, Polaris, which it anticipates will be the first fusion reactor to generate electricity.
Polaris, Helion’s seventh prototype, sits inside a 27,000 square-foot building in Everett, Washington. It took more than three years to build, which is quick by fusion industry standards. But to hit its ambitious 2028 deadline for Microsoft, the startup will have to move even faster on its commercial-scale power plant.
The difficulties Helion faces are in many ways similar to those in other leading-edge industries.
“In AI, what’s the big challenge? Getting the chips. In fusion, what’s the big challenge? Getting the chips,” CEO David Kirtley told TechCrunch in a recent interview. “Polaris is 50,000 of these large-scale, pulse-power semiconductors, and getting those set the timeline.”
The solutions it’s seeking are similar, too. The new investment will go toward bringing a significant amount of specialized manufacturing in-house. For example, the company had to order a type of short-term energy storage device known as capacitors three years in advance.
“Our goal is to go from waiting three years for a supplier to give us capacitors to us making our own capacitors but faster, so now we can make them in a year or less,” he said.
Despite having to build a supply chain from scratch, Kirtley remains optimistic that Helion can still deliver electrons to Microsoft in just a few years.
“We’ve been working on siting for the Microsoft facility for a few years already,” Kirtley said. He declined to name a location, but said the company has been working on permitting and grid interconnection, a process that can take years.
Part of Helion’s appeal — and part of the risk, critics would argue — is that its approach to fusion power differs from virtually every other startup in the sector.
Generally speaking, there are two main approaches: Magnetic confinement uses powerful magnets to squeeze plasma to get it hot and dense enough to spark fusion reactions, which are meant to burn continuously to generate steam to drive a turbine. Inertial confinement fires powerful lasers at fuel pellets, compressing them to the point where the fuel atoms fuse. To generate enough heat to feed a steam turbine, a reactor has to fire several times per second.
Helion is building something completely different, known as a field-reversed configuration reactor. The device looks like an hourglass with a bulge in the middle, and it’s ringed with powerful magnets, which guide and compress the plasma throughout the course of each reaction, which Helion calls a “pulse.”
At the beginning of a pulse, Helion injects a mix of deuterium and helium-3 into each end and heats it until it forms a plasma. Magnets then shape each plasma into a doughnut and propel them toward each other at more than 1 million miles per hour.
When the plasmas reach the fusion chamber — the bulge in the middle of the hourglass — they collide and are squeezed further by another set of magnets. This heats the plasma to more than 100 million degrees C, leading to a cascade of atoms fusing. Altogether, it’s similar to how a spark plug ignites fuel inside an internal combustion engine.
The energy added by the fusion reactions generates a surge in magnetic force, which pushes back on the reactor’s magnets. This extra magnetic force is then converted directly into electricity. If everything works as intended, Helion’s reactor will generate more electricity from that magnetic burst than it needed to power the magnets in the first place. And because the system harvests electricity from magnets instead of generating steam to spin a turbine, it should be more efficient, lowering the bar to break even.
The current design for a commercial-scale Helion reactor will pulse a few times per second, Kirtley said. A single reactor will generate 50 megawatts of electricity, and a power plant could contain multiple reactors.
In the lab, the company has small systems that can fire over 100 times per second, so it’s possible that future Helion reactors will be able to fire 60 pulses per second, the same frequency as electricity on the grid. “But there’s some big engineering challenges to get to those high repetition rates at the kind of big pulse powers where we talk about millions of amps flowing around,” Kirtley said.
Helion raised the new funding to speed work on the power plant, including expanding in-house machining capabilities and capacitor manufacturing. “One of the things that drove the Polaris timeline was actually making all the magnetic coils. And so I want to be able to make all of those in house,” he said.
The new round is smaller than the startup’s previous fundraise of $500 million. New investors in the round include Lightspeed Venture Partners, SoftBank Vision Fund 2, and a major university endowment. Existing investors Sam Altman, Capricorn Investment Group, Mithril Capital, Dustin Moskovitz, and Nucor also participated.