It is no longer a hypothetical goal in the commercial fusion machine of the private sector to reach temperatures of 150 million degrees Celsius. Helion Energy reports that its seventh-generation prototype, Polaris, attained that temperature and was also able to achieve a demonstration of deuterium-tritium (D-T) fusion as well as two engineering indicators, which turns the company 2028 grid-electricity plan into an ambition, to a set of testable constraints.
The amount is significant due to the fact that on Earth, the fusion systems have to accelerate plasmas that are much hotter than the core of the Sun as a means of offsetting the absence of the pull of the star. In most of the fusion community 100 million degC is considered a realistic limit of commercially significant plasma conditions not because it represents a net electricity, but because it is the point at which the disparity between laboratory physics and consistent operation in hardware capable of being serviced, inspected and rebuilt vanishes.
In February 2021, an old pattern of criticism of venture-backed fusion by Helion was re-introduced: there are too few scientific gateways between the money-raising and a declaration of grid power. Under its own description, the first privately made fusion energy machine was 1964 with Polaris, which it said showed a demonstration of measurable D-T fusion, but that the system has since run with a variety of fuels. The company has also re-iterated that it was found to have regulatory authorization to own and utilize tritium in fusion experimentation- a practical fact that tends to be overwhelmed by temperature news but is capable of determining test rate, diagnostics and facility design.
The key engineering bet languishing at the bottom of those milestones: the way electricity is tapped. A variety of mainstream reactor designs transform fusion energy into heat, which is converted to steam which is in turn converted to electricity by turbines. The design of Helion is intended to produce electricity in a more direct manner by extracting energy due to variations in magnetic field of plasma in a pulsed, magneto-inertial style compression, and avoidance of thermal balance-of-plant equipment. Should such a power plant be experimentally successful, it would be possible to reduce the complexity of certain components and redistribute complexity elsewhere: pulsed power device, high-reliability switching, component wear in the presence of repeated electromagnetic loads, and the control systems to maintain cycles constant in plant-scale repetition rates.
The odd parallel build strategy at Helion can be explained using that systems view. Polaris, which is located in the Seattle area, is the high-speed machine of the plasma campaigns and diagnostics, and the company has also initiated construction project of its first commercial machine location site in Malaga, Washington. The risk of adding hardware, but closing physics and reliability questions, is building “next” hardware, but it is also a manufacturing philosophy: make prototyping a series of stepping-stones, instead of one-off monuments.
The wider industry is branching out on various methods of putting the same physics to use. Commonwealth Fusion Systems is optimistic that high-field tokamaks based on REBCO high temperature superconductor magnets will reduce the size of devices as well as boost confinement performance, and that SPARC will serve as a net-energy plasma demonstration on the way to ARC, a grid-oriented system. In contrast, General Fusion focuses on magnetized target fusion using a liquid-metal wall and a staged program around its LM26 machine, outlining a path which includes validating power-plant subsystems (seals, valves, and heat exchange) as part of its commercialization path.
The next technical bridge is the fuel and repeatability that Helion has. The company has defined D-T operation as a move towards deuterium-helium-3 operation service with Polaris to drive the temperatures and operating consistency to an extent where that transition can be believable. The milestone, so to speak, is not so much a finish line as a new constraint: the machine is now hot enough to reveal which parts of the design fail first, which diagnostics work under real campaigns, and what “rapid iteration” means when parts wear out predictably, not hypothetically.
