Energy
Helion Energy: Powering the OpenAI & Microsoft AI Race
Nuclear fusion could become the long-awaited perfect energy source: no carbon emission, unlimited fuel supply, not producing any significant pollution, and massively powerful.
And contrary to classical nuclear power plants, using nuclear fission, it has zero risks of uncontrolled chain reaction or causing large-scale radioactive incidents like Chornobyl or Fukushima.
Nuclear fusion works by recreating on Earth the conditions inside the core of the Sun, with hydrogen (usually deuterium, a hydrogen isotope with an extra neutron) exposed to massive pressure and tens or even hundreds of millions of degrees so that they can fuse into helium atoms or other heavier elements.
As light elements’ nuclei contain more energy than heavier ones, this releases an amount of energy per atom >10x more energetic than even the most powerful nuclear fission reactions.
Source: Nature
Scientists have been creating the conditions for nuclear fusion in experimental reactors since the late 1950s. However, due to the extreme conditions required to make it happen, nuclear fusion has so far failed to be commercially viable.
A key issue is that for nuclear fusion to be a cheap source of energy, plasma needs to be maintained for several minutes, ideally hours, so that the initial energy spent to create these temperatures and conditions is “paid back” by the sustained fusion.
So while creating the plasma is somewhat “easy, keeping the plasma contained and stable is the hard part, usually requiring massive magnetic fields produced by superconducting magnets cooled a mere few degrees above absolute zero.
(You can learn more about the fundamentals of nuclear fusion in our dedicated report “Nuclear Fusion – The Ultimate Clean Energy Solution on the Horizon.”).
Still, progress in AI & computing, plasma sciences, and advanced materials has made new fusion reactor designs more compact, cheaper, and energy efficient.
So, from academic experiments, the field has rapidly evolved in the past decade, with many private companies joining the race, convinced that nuclear fusion is now a mature enough technology to become commercially viable.
Many are moving beyond the more classic tokamak or even stellarator, looking for innovative options to avoid the struggle associated with maintaining ultra-high temperature plasmas stable in a donut-shaped reactor.
One of them is Helion Energy, whose “Pulsed Magnetic Compression” could not only create fusion more efficiently, but also include a way to directly drain energy from the plasma, instead of the more usual heat extraction -> steam -> electricity, which causes significant losses.
Helion Energy History
Helion Energy was founded in 2013, and in 2015 was awarded a contract by ARPA-E (Advanced Research Projects Agency–Energy), an agency within the United States Department of Energy charged with supporting the development of advanced energy technologies.
In 2023, Helion made headlines as it signed a promise to provide Microsoft with energy from nuclear fusion by 2028, a very tight deadline that surprised all observers, and contributed to the realization that the date for achieving commercial viability of fusion technology was maybe getting a lot sooner than expected.
Since then, the company has achieved further technical prowess and nuclear fusion accomplishments, with other AI companies now looking for the company to provide them with energy as well (see below).
As of January 2025, Helion is valued at $5.4 billion, with some key backers including Sam Altman (OpenAI), SoftBank Vision Fund 2, Lightspeed Venture Partners, and Peter Thiel’s Mithril Capital.
Helion Energy Technology Concept
Nuclear Pistons
Most nuclear fusion concepts are looking at creating long-lasting plasma and fusion reactions. While theoretically more productive, this is also a very energy-consuming process, so these designs have remained unable to produce net energy.
Instead, Helion Energy is on a pulsed magnetic fusion approach called magneto-inertial fusion (MIF), designed to run in short, high-energy pulses to continuously produce electricity.
In its most simplified form, the technology used by Helion Energy can be described as a high-tech piston not completely dissimilar to the pistons used in a combustion engine. Here, too, the compression of the fuel results in its ignition.
This pulse approach lets the reactor create a fusion reaction in a series of short moments, without needing to sustain it for a long period.
The way it works is that high-voltage magnets turn a gas of light elements into a plasma ring.
Two such rings are being created, each at the opposite end of the reactor, 40 feet apart. They are then accelerated with magnetic fields at astonishing speeds exceeding 1 million mph. The crash of the plasma rings into each other is so powerful that it is enough to force the nuclei of the atoms to merge, despite both being positively charged.
Source: Helion Energy
The colliding plasmas are then further compressed in a sudden pulse by another powerful magnetic field. This creates conditions with a temperature above 100 million Celsius degrees (180,000,000 °F), a level considered required for any commercially viable fusion reaction.
So overall, Helion Energy is trying to cause the conditions for fusion not by imitating stars (heat and pressure) but by using kinetic energy and collision, in a way more similar to a particle accelerator than a classical fusion reactor, with the addition of compression by a short magnetic pulse replacing the attempts to create stable long-duration compression.
Direct Electricity Capture
With the exception of photovoltaics, almost all electricity generation technologies use the heating of water into steam to turn a turbine as a way to convert heat or mechanical energy into usable power. This is a tried and tested method, but it also loses a significant portion of the initial energy.
For most fusion reactors, the concept stays the same, and the heat of the plasma and sustained fusion would be directed to a cooling system, capturing the energy with a turbine.
Helion Energy is planning to use a radically different method, using the fact that the fusion reaction pushes the reactor’s magnetic fields outward. Following Faraday’s Law, this movement induces an electric current directly into the coils surrounding the machine.
This electric current can be directly harnessed and utilized without any further equipment or energy conversion. This can be a much more efficient method of harnessing the generated energy, strongly improving the economics of the operation.
In addition, this reduces the complexity of the reactor and its overall size: no massive cooling tower, no need for water intake, no complex piping, no steam turbine, no supercritical steam needing containment, etc.
It would result in Helion Energy’s 50MW prototype, promised to Microsoft, fitting into a large industrial building, with a footprint of 30,000 to 100,000 square feet at maximum, roughly the same as a football field.
Lastly, no need for further equipment to capture the energy generated, which reduces the cost of the power plant, makes permitting easier, and reduces supply chain risks.
Helion Latest Achievements
Improving Reactors
Helion tested its concept with Trenta, its 6th fusion prototype, powered by deuterium-helium-3 reactions, but also working with deuterium-deuterium reactions. The reactor performed 10,000+ pulses, reached 100 million degrees, and overall proved the concept at a small scale.
It was followed by the much larger and more ambitious Polaris reactor, built in only 3 years.
“Our philosophy has always been to build, test, iterate, and repeat as fast as possible. And that’s exactly what we’re doing right now.”
David Kirtley, co-founder and CEO of Helion
This 19-meter (62-foot) reactor can do the same reactions as Trenta, but also deuterium-tritium reactions. This made Polaris the first privately funded fusion energy machine to operate with deuterium-tritium fuel.
This is an important change, as the reliance of Trenta on Helium-3, a very rare element that might be mined in space in the future. However, for now, the plan is still to ultimately use Helium-3.
The reaction reached new highs in temperatures, achieving 150,000,000 degrees Celsius (270,000,000 °F), and Helion will continue to increase plasma temperatures in Polaris in future tests.
Polaris is also being used to validate the direct electricity capture approach that will be used in Orion.
“Seeing the data from the Polaris test campaign, including record-setting temperatures and gains from the fuel mix in their system, indicates strong progress. Our ability to get fusion on the grid requires approaches that enable rapid turnaround in design and testing, and these results reflect the growing capability of the U.S. fusion ecosystem.”
Jean Paul Allain – Associate Director for Fusion Energy Sciences in the Department of Energy’s Office of Science.
In parallel, Helion began building in July 2025 on the site of Orion, its first commercial machine, in Malaga, Washington, which will deliver electricity from fusion to the grid for Microsoft.
Progress on permitting was reported in October 2025, and Washington state passed the House Bill 1018, which classified fusion as a clean energy source and legally distinguished it from traditional nuclear fission, making its regulatory pathway a lot simpler.
“The state of Washington is the world’s leading hub for fusion energy, which one day soon could provide vast amounts of the type of power we need to keep electricity prices down and increase America’s economic competitiveness,”
In-House Manufacturing
Building Polaris was also an exercise in ramping up in-house production, with Helion Energy manufacturing itself the required quartz tubes and high-voltage capacitors. 2,500 capacitors will be needed for building Orion, with both workers and robotics used for the production process.
This is crucial as Helion is planning to mass-produce its commercially-viable design, not dissimilar to how production of SMR fission power plants is being planned.
“If you want to scale quickly, and if you want to be able to build an intelligent manufacturing process, you have to have engineers with a really good understanding of how the thing works. And you have to have design engineers with a really good understanding of what’s hard about manufacturing.”
At the end of 2025, the company signed a lease near its Everett headquarters for a 166,000 square-foot space dubbed Omega, where the company will install an assembly line to build thousands of capacitors, putting Helion as a strictly power-generation-focused company, not a research and IP-driven project.
“Helion is a manufacturing company. It’s not an R&D company. It’s not a science experiment. It’s very much a manufacturing company.”
So for Helion Energy, the manufacturing capacity is a bet on its future success in building and making Orion profitable, and then immediately scaling up fusion power plants with mass production. This is maybe the most special aspect of the company, marking it as uniquely ambitious and optimistic about its timeline, even compared to other aggressive fusion companies.
“These high volume lines are not for our Orion machine, but for the next machine. A factory operating at 50% of its design capacity or less can spit out Orion, no problem. But we’re really looking beyond that into 2030.”
Helion Energy’s Deals
As mentioned before, Microsoft (MSFT ) was early in betting on Helion, securing its first 50MW prototype, Orion, for its facilities in Washington state, with construction already ongoing. The power plant will connect “just upstream of the Microsoft data centers”.
Another major partner hungry for energy is negotiating with Helion Energy: OpenAI. The deal under discussion would have OpenAI receiving as much as 5 gigawatts of power by 2030, or 100x larger than the initial deal with Microsoft. This is almost as much power generation capacity as Washington state and the USA’s largest hydropower dam, Grand Coulee Dam.
This might seem overly ambitious, but so did the 2023 deal with Microsoft. And as Sam Altman is directly a shareholder in Helion Energy, he likely knows a lot about the company’s actual capacity to deliver or not.
“Sam has played an integral role in Helion’s development, helping us focus on the thing that matters most: deploying fusion for customers as quickly as possible to fully satisfy the world’s need for clean and abundant energy. We look forward to continuing to work with him in this new capacity.”
David Kirtley – Helion CEO
Can Helion Energy Make Fusion Work?
Helion is doing fusion in its own unique way, and this carries both the chance that it might work quicker than anyone expected and serious risks.
In addition to its unique method for fusion, direct power generation is a rather bold move that could increase the yield of future power plants by 2-3 fold, as heat-to-steam-to-power conversion usually has very low efficiency.
This also makes the reactor a lot less capital-intensive, a potential crucial point for quick ramping up of the company’s production to 5 GW in negotiation with OpenAI and even greater capacity post-2030.
In addition to competitive capital cost, Helion’s fusion power plant is projected to have negligible fuel cost, low operating cost, high up-time, and is able to deliver on-demand and quickly variable amounts of power, something that renewables cannot do and even nuclear fission struggles to achieve.
However, a few challenges are still down the road:
- Producing 50 MW will require several pulses per second, which could accumulate mechanical and thermal stress on components, especially with years or decades of continuous operations planned.
- Maintaining the shape of the plasma rings is notoriously difficult, and doing it at scale might be even more so.
- Helium-3 is very rare in nature, so Helion plans to “breed” helium by fusing deuterium (D-D fusion) in their own reactors and allowing the resulting tritium to decay into helium. This method is yet to be demonstrated at scale.
- Reactors could sustain damage from radiation over time, which could impact the economics of their operation.
- The regulatory framework for nuclear fusion does not exist yet and depends on the famously slow and cautious Nuclear Regulatory Commission (NRC), and might be slower to be finalized than Helion’s scheduled ramp-up in production.
With a very compressed deadline for the launch of Orion, it will become clear rather quickly if these issues can be tackled successfully.
Investing In Helion Energy’s Success
Microsoft
(MSFT )
As the AI race intensified, it became apparent that human resources of GPUs might not be the most critical resource. Instead, energy supply might be the biggest and hardest to solve chokepoint in deploying and operating ever larger and more numerous AI data centers.
This is why Microsoft was early in restarting fission power plants for its data center exclusive use. And why did it sign a deal with Helion when no one believed the timeline promised by the company?
“While the path to commercial fusion is still unfolding, we’re proud to support Helion’s pioneering work here in Washington state as part of our broader commitment to investing in sustainable energy.”
Melanie Nakagawa – Microsoft’s chief sustainability officer
This could prove a definitive advantage for Microsoft against other hyperscalers constrained by energy shortages.
Of course, the same can be said of OpenAI, but as both Helion and OpenAI are still privately listed, Microsoft is the main way to get exposure to this energy revolution in the making.
Besides a potentially fusion-powered AI data center, Microsoft is also betting on traditional nuclear, SMRs, and other technologies to bring it a definitive advantage in AI, in addition to software and model development.
The company is also a leader in quantum computing, cloud computing, video games, and a key provider of B2B digital services. You can read details about these segments of Microsoft in our investment report dedicated to the company.
Latest Microsoft (MSFT) Stock News and Developments
