Fusion's Year of Firsts: What Helion, Realta and Thea Have Actually Shown

Fusion energy has spent decades as the technology that arrives in ten years. In 2026, private companies started publishing results that are narrower than that promise but more concrete than anything the sector has put on paper before.

Helion's newsroom statement on 13 February says its seventh-generation Polaris prototype became the first privately developed fusion machine to run on deuterium-tritium fuel and to reach plasma temperatures of 150 million degrees Celsius. The company notes it broke its own industry record of 100 million degrees Celsius, set by its sixth-generation Trenta device. Within the industry, 100 million degrees Celsius is treated as the threshold for a commercially relevant machine. That number matters because it is the one independent specialists were invited to comment on.

Ryan McBride, an inertial confinement fusion and pulsed power expert who has worked as a department manager at Sandia National Laboratories and taught at the University of Michigan, said he had reviewed diagnostic data from Helion. "It is exciting to see evidence of D-T fusion and temperatures exceeding 13 keV or 150 million degrees Celsius, and I look forward to seeing more progress," he said in the Helion release. Jean Paul Allain, associate director for fusion energy sciences at the Department of Energy's Office of Science, said the data indicated "strong progress" and that the results reflected "the growing capability of the U.S. fusion ecosystem."

Two caveats sit inside Helion's own text. The D-T campaign is one step in Polaris' testing programme, and the company says it will keep raising temperatures to reach conditions for deuterium-helium-3, the fuel it intends to use commercially. Polaris itself is a prototype, not a power plant.

The plant in Malaga is the harder test

Helion began construction in July 2025 on the site of Orion, its first commercial machine, in Malaga, Washington. The stated customer is Microsoft. TechCrunch reported on 28 September 2026 that Orion is a 50-megawatt plant intended to supply electricity to Microsoft beginning in 2028.

A temperature reading and a grid connection are different categories of achievement. The first is a physics experiment under controlled conditions; the second requires licensing, fuel handling at scale, uptime and a customer willing to take the power. Helion's own release describes the 2026 results as "one step," and the company's funding history suggests investors are pricing the later steps, not the earlier ones. A company news page referenced by Helion lists a $465 million Series G round and a later report of a $15.5 billion valuation, though those items appear as headlines rather than as statements in the 13 February release itself.

TechCrunch also reported that Helion's Series G, initially announced at $465 million in June, closed at $500 million in September. The same piece notes that Commonwealth Fusion Systems raised another $1 billion in July, bringing its total funding to $4 billion, and that in April CFS became the first fusion company to apply to PJM Interconnection, the largest U.S. wholesale electricity market. CFS is building SPARC as a demonstration machine on the way to its ARC plant.

So the sector now has two companies with announced grid plans, one of them naming a delivery year. That is a shift in kind, not just degree. It also means the 2028 date is now the number to watch, because it is falsifiable.

Realta's conversion claim is narrower, and it says so

On 19 June 2026, Realta Fusion and the University of Wisconsin-Madison demonstrated direct energy conversion on the Wisconsin HTS Axisymmetric Mirror device. Realta describes this as the first time a commercial fusion company has demonstrated DEC applied to a fusion plasma.

DEC means capturing the kinetic energy of charged particles leaving a plasma as electric current, rather than routing everything through a thermal cycle. Realta's first prototype is a single-stage converter with three meshed grids: a grounded grid, an electron repulsion grid and an ion collector grid. During mirror plasma operations, the assembly drew multiple amps of current at around 100 volts, which Realta says is enough to illuminate a few lightbulbs.

The company is explicit about the limits. In the same post it states that this is "neither a demonstration of net-electricity production nor large-scale conversion of fusion-born power directly into electricity." WHAM runs on deuterium only, so most of the directly converted energy is input power the team supplied to heat and sustain the plasma. Realta says it will scale the converter's voltage to draw more current in the weeks after the demonstration, and aims for multi-kilowatt and eventually multi-megawatt capability on future devices.

That kind of self-limitation is worth noting. Fusion press releases have a long history of sliding from a laboratory measurement to an implied power plant. Realta's post does the opposite, listing what has not been shown before listing what has.

Thea bets on magnets and software

Thea Energy announced a $100 million Series B on 27 May 2026, led by US Innovative Technology Fund with participation from General Innovation Capital Partners and Linse Capital, among a longer list of investors. The company is commercialising the stellarator, a magnetic confinement architecture that spun out of Princeton University and Princeton Plasma Physics Laboratory in 2022.

Thea says the money will expand magnet manufacturing, including a second facility in northern New Jersey, and support siting and construction of Eos, a large-scale integrated stellarator. The company says it will select a site for Eos later this year and double its team. It also says it holds DOE certification of its Helios preconceptual design milestone, described as a first among awardees of the Milestone-Based Fusion Development Program, and that it has built and operated what it calls the world's first superconducting magnet array capable of producing the complex fields commercial stellarators need.

Gaetano Crupi, a managing director at USIT, framed the investment in terms of manufacturing. Prior three-dimensional stellarator magnets made the system impractical to build, he said, and Thea's approach shifts complexity "from precision mechanical fabrication to software-defined controls." That is a claim about engineering economics, and it will be tested at Eos scale, not in a press release.

ITER is still the benchmark, and still delayed

Against all of this, the public project remains the reference point. CNET reported in May 2026 that ITER, the International Thermonuclear Experimental Reactor in southern France, carries an estimated cost of $22 billion and is designed to contain plasma at 150 million degrees Celsius. Cracks in the piping of a heat shield were found in 2020, along with welding distortions, and the COVID-19 pandemic contributed to a years-long delay and an additional $5 billion for repairs.

ITER's chief strategic advisor, Laban Coblentz, told CNET that the collaboration between China, Russia, the US, Europe, Korea, India and Japan was "either genius or insane." Javier Artola, a scientist modelling ITER's plasma, described the project as publicly funded knowledge. More than 30 countries are party to the agreement, and all of them get access to the science.

The pitch for ITER inside the private boom is straightforward: every problem it solves is one fewer for startups. The counter-pitch is that ITER's delays are exactly why private money moved in. Both can be true. ITER's plasma target and Helion's reported Polaris temperature are the same figure, 150 million degrees Celsius, which is a useful reminder that a temperature milestone is not by itself a commercial milestone.

What the year actually established

Strip the announcements back and 2026 produced three things. A private machine ran on D-T fuel and hit a record temperature, confirmed in general terms by outside reviewers quoted by the company. A private company converted plasma energy to electricity at lightbulb scale, with explicit caveats about what that does not mean. And a stellarator developer raised $100 million and named a construction target for a large integrated device.

None of that is electricity on the grid at scale. The next hard number on the calendar is Helion's stated 2028 start of supply to Microsoft from a 50-megawatt plant in Malaga. If that slips, the sector's credibility slips with it, because it is the first private fusion commitment specific enough to be checked.