Insurance Dimes

Research and insights, backed by data

Research and insights, backed by data

,

Every fusion startup that has raised over $100M

Private fusion companies have raised more than 9.7 billion dollars since 2021, a stunning bet on a technology that’s been “twenty years away” for seventy years. Here’s who’s raised what, what they’ve actually built, and the deadlines worth putting on your calendar.

$0Private fusion funding before 2015. The field was government only.
$5B+Raised in 2021 to 2022, after the HTS magnet breakthrough
$9.7B+Raised since 2023. Zero net electricity delivered so far.

That jump from zero to billions happened in less than a decade. The company-by-company breakdown below shows where the money went, and how each one’s claims stack up against what it’s actually built.

This is an interactive visualization. Tap or click on each company to reveal further information.

Every fusion startup that has raised over $100M

Building its Sparc reactor in Massachusetts, targeting operation in late 2026 or early 2027. Its planned commercial plant, Arc, would be built near Richmond, Virginia, with Google already agreeing to buy half its output.

In December 2025, TAE announced it would merge with Trump Media & Technology Group in an all-stock deal valuing the combined company at $6B, with TAE’s CEO becoming co-CEO of the merged entity.

The most aggressive timeline in the industry: Helion says it will deliver electricity to its first customer, Microsoft, by 2028. Backers include Sam Altman, SoftBank, BlackRock, and Peter Thiel.

Rather than chase net electricity right away, Shine earns money today from neutron testing, medical isotopes, and recycling radioactive waste while it develops fusion capability.

Its Series A alone topped $1 billion. Funds pay out in milestone-based tranches, an approach more common in biotech than energy. Led by former Human Genome Project chief Eric Lander.

Ran short of cash in 2025 and cut 25% of staff, then patched things together with smaller emergency rounds. It’s now planning to go public via a reverse merger with a SPAC.

Founded by the former chief scientist of the only fusion experiment ever to beat scientific breakeven. Emerged from stealth with a $450M Series A led by Bessemer Venture Partners.

Plans to mass-produce fuel targets at nearly a million per day. Also received $200M in grants and gained access to a decommissioned nuclear plant through utility partner RWE.

Now supplying magnets for the UK government’s STEP Fusion program, in addition to developing its own reactor.

Recently pivoted to pursue nuclear fission alongside fusion, and even a hybrid fission-fusion plant, aiming to bring in revenue sooner.

Plans to sell its technology to utilities like the Tennessee Valley Authority rather than build and operate plants itself, closer to how fossil fuel plants get built today.

One of the few well-funded startups betting on stellarators rather than tokamaks, arguing the twisted magnetic shape gives more stable, longer-lasting plasma.

Betting that whichever reactor design eventually wins, the industry will still need suppliers for the “balance of plant”: heating systems, heat extraction, and grid integration.

Its silicon-based fuel target is designed to lean on decades of semiconductor manufacturing know-how. Building a demo facility with Colorado State University, targeting 2027.

Uses control software to make dozens of simple magnets act like the twisted magnetic field a traditional stellarator needs, aiming to cut manufacturing cost and complexity.

In 2025, First Light dropped plans to build its own power plant, choosing instead to license its core technology to other companies and pursue science and defense applications.

In June, Xcimer turned on Phoenix, a prototype it says is the most powerful privately owned laser in the world, en route to a planned system five times more powerful than the NIF’s setup.

Funding totals reflect amounts disclosed by each company as of late 2025 and 2026 and are approximate where exact totals weren’t public. Figures belong to the companies that stated them, not to independent verification.

Of everything on that timeline, one commitment stands apart where an actual signed contract exists with a delivery date attached.

In August 2023, Helion signed the first commercial fusion power purchase agreement anyone has disclosed, promising Microsoft electricity by 2028. Of the seven companies tracked in the broader funding timeline, it is the only one with a public commercial contract of this kind (ie:- with a delivery date attached).

The agreement includes financial penalties if Helion fails to deliver, according to public reporting on the deal. The exact penalty terms have not been made public.

A missed date would do more than cost Helion money. This was the first corporate fusion PPA anyone signed. If it fails, it becomes the reference point every future fusion contract gets measured against, and a harder case for the next company trying to sign one.

Helion's fuel cycle also needs conditions no experiment has reached yet, so the physics risk and the contract risk are tied together here in a way most energy PPAs never have to deal with.

Those aren't the only companies in the race, either, just the ones with clear, dated funding rounds. A handful of others have raised capital too, even without a public timeline to check them against.

Click on each company's tile to reveal more information

Elsewhere in fusion

These four raised real money too, but public sources behind this piece did not include dated funding rounds for them, so they sit outside the main funding timeline.

A spherical tokamak built around a compact high temperature superconductor design. Building the ST80-HTS machine in the UK, targeting commercial operation in the 2030s.

Uses mechanical plasma compression instead of magnets or lasers. Building a demonstration plant at Culham in the UK, targeting the late 2020s.

Uses a z pinch approach that needs no external magnets to hold the plasma. Its demonstration system began operating in 2024.

Fires a hypersonic projectile at a fuel target instead of using magnets or lasers directly. Demonstrated fusion in 2022 and is targeting the 2030s for further progress.

With so much money and momentum in play, it's worth knowing exactly what's still left to solve. Here are the five engineering puzzles standing between today's experiments and a working power plant.

Five problems no one has solved yet

Every fusion company still has to clear these, regardless of how much it has raised.

A single gigawatt deuterium tritium plant needs about 56 kilograms of tritium a year. The entire global supply sits around 25 kilograms and is shrinking, because tritium has a 12 year half life. Every plant has to breed its own tritium from a lithium blanket, and nobody has proven that at scale yet.
The neutrons a fusion reaction produces damage the metal around them. No material has been tested under fusion level neutron exposure for anywhere near the decades a real plant would run. The dedicated test facility for this, called IFMIF DONES, has not been built yet, and money cannot speed up how fast materials absorb radiation damage.
The plasma facing components need to handle heat flux comparable to a rocket nozzle, for years rather than seconds. Tungsten is the leading candidate material, but it is brittle and erodes under this kind of load.
A power plant needs to run around 90 percent of the time. No fusion device has run continuously for more than minutes. The jump from a physics experiment to an industrial plant is much bigger than most coverage suggests.
Tritium is radioactive and it moves through solid metal. A working plant would process kilograms of it every day. Containing that flow safely is still an open engineering problem, separate from the supply problem above.

So where does that leave the industry? Two honest cases can be made, and both are worth holding at once.

Can they actually deliver

  • The HTS magnet breakthrough behind Commonwealth Fusion Systems is real, and it changes the cost math for building a compact tokamak.
  • Private capital moves faster than government labs, and several different approaches are being tested in parallel rather than in sequence.
  • Machine learning is already improving plasma control. Google DeepMind published a real time tokamak control result in Nature back in 2022.
  • Microsoft, Google, and Nucor Steel are putting real money behind future fusion power. That kind of demand signal is hard to fake.
  • Every energy technology gets cheaper as it scales up. The first fusion plant will be expensive. The tenth one might not be.
  • Seventy plus years and billions of dollars of research have produced zero net electricity from fusion so far.
  • Company timelines are also fundraising documents. Nearly every fusion delivery date in history has slipped.
  • ITER is a $25B international project staffed by the best fusion engineers in the world, and it is still behind schedule. Startups face the same physics with a fraction of that budget.
  • The tritium supply problem has no financial shortcut. Money cannot speed up how fast materials absorb neutron damage.
  • Solar and battery costs keep falling every year. By the time fusion is ready, it may be competing against clean energy that is already extremely cheap.

Leave a Reply

Your email address will not be published. Required fields are marked *