Nuclear fusion.
The tape comes first.

Helion promised Microsoft electricity from a fusion plant by 2028. In September, it quietly changed its website to say "initial operations" in 2028, and a spokesperson added that full power would come in 2029 or 2030.

That's fusion right now: more money than ever, the first public stocks, and the first quiet edits to the timelines.

The timelines are the thing to watch, because the physics isn't really in doubt. Fusion energy comes from the reaction that powers the sun. Heat hydrogen past 100 million degrees, squeeze it hard enough, and the atoms merge and release a huge amount of energy, with no meltdown risk and no long-lived waste. The engineering is the problem. There are three finish lines: a plasma that gives back more energy than it took to heat it, a whole machine that comes out ahead, and electricity on a wire. Only the first has been crossed, at a US government lab.

None of that has slowed the money. Fusion companies have raised more than $14 billion, about a third of it in the past year, and Google, Eni, and Microsoft have signed contracts for power that doesn't exist yet. You can even buy fusion on a US exchange for the first time, though the first public nuclear fusion stocks have had a rough start. General Fusion listed in July and then restated its SPAC predecessor's results to fix a $411 million error. TAE Technologies is merging into Trump Media, which is lending it hundreds of millions just to get to the closing.

No fusion plant will sell a kilowatt-hour in 2027, but the companies building them are already placing real orders. So for nuclear fusion stocks, the list runs in the order the money arrives:

• Suppliers: the superconducting tape and parts every reactor needs. Paid on delivery.
• Buyers: big companies that invested in fusion or signed up for its power. Fusion is a free option on top.
• Reactors: the two public pure-plays. Most upside, most ways to lose.
• Private: the three companies whose milestones move everything above.

Almost every magnet-based fusion machine runs on high-temperature superconducting (HTS) tape, thousands of kilometers of it per reactor. Only a few companies make it, and they get paid whether or not the plasma cooperates.

Fujikura TYO: 5803

On September 30, Commonwealth Fusion Systems ordered more than 10,000 kilometers of superconducting tape from Fujikura for its first power plants, and called it the largest such order ever placed. That's fusion revenue that exists today, on a purchase order.

The catch is that Fujikura's stock already trades on AI, because it also makes the optical fiber going into data centers. You're paying an AI multiple and getting fusion as the bonus.

Furukawa Electric TYO: 5801

Furukawa owns SuperPower in New York, one of the few HTS tape plants outside Japan, and supplies the tape for Tokamak Energy's machines in the UK. In June it agreed with Tokamak Energy, the magnet partner on Britain's STEP fusion program, to study building tape capacity in the UK. When an order like CFS's starts eating the world's supply, Furukawa is the second call every reactor builder makes.

Bruker NASDAQ: BRKR

Most people know Bruker for lab instruments. Its superconductor unit is one of the world's leading makers of low-temperature superconducting wire, and a subsidiary builds parts for ITER, the giant international fusion project in France. Bruker says fusion helped lift its superconductor revenue this year, and orders from energy research are growing. It's the least exciting fusion stock you can own, which is the point.

These companies either invested in fusion or signed up to buy its power. Fusion won't move these stocks for years, but if a plant works, they already own the first electrons.

Alphabet NASDAQ: GOOGL

Alphabet has a seat at two fusion tables. It's an investor in both CFS and TAE, it agreed in 2025 to buy 200 megawatts from CFS's first commercial plant in Virginia, and DeepMind works with CFS on AI that controls the plasma. You don't buy Alphabet for fusion, but if you own it, you own a real stake in the outcome.

Eni NYSE: E

Eni has backed CFS since 2018, and in 2025 it signed a deal worth more than $1 billion for power from that same Virginia plant. Between Eni and Google, more than half the plant's output is sold before it exists. You get an oil major's dividend while you wait on one of the best tickets in fusion.

Constellation Energy NASDAQ: CEG

When Helion signed the first fusion power contract with Microsoft in 2023, Constellation was named the company that would market and deliver the electricity. It already runs the biggest nuclear fleet in the US, and it's who the hyperscalers call when they want carbon-free power in bulk. If fusion power ever gets sold, someone has to sell it. Constellation has the first contract.

The only two ways to own a fusion developer on a US exchange. Both carry baggage that has nothing to do with physics.

General Fusion NASDAQ: GFUZ

General Fusion has the most down-to-earth design in fusion: compressed-gas pistons squeeze a plasma inside a wall of liquid metal, with no exotic magnets or giant lasers. It listed in July with about $150 million in cash, which it says funds its LM26 machine through its milestones to the end of 2028.

The problem is the books. In September it restated its SPAC predecessor's first-quarter results to fix a $411 million non-cash error and disclosed a material weakness in its controls. The SPAC's auditor resigned and General Fusion's own auditor took over. None of that touched the cash. But a moonshot only works if investors trust the numbers, and right now that's the milestone that matters most.

Trump Media & Technology Group / TAE Technologies NASDAQ: DJT

TAE Technologies is a serious fusion company. It has built five generations of prototypes, raised more than $1.6 billion from investors including Google and Chevron, and aims to demonstrate fusion energy at its first plant in 2031. The only way to own it is through Trump Media, which is merging with TAE in an all-stock deal that splits the combined company roughly in half.

The merger paperwork was filed on September 30, and Trump Media has already lent TAE $200 million, with up to $100 million more available. Until the deal closes, you're buying a social media company with a bitcoin pile. After it closes, you get all of that plus a fusion company.

The three companies below set the mood for the whole sector. When they hit a milestone, the nuclear fusion stocks above rally. When they slip, everything sells off.

Commonwealth Fusion Systems

If you follow one fusion company, make it CFS. It has raised about $4 billion, nearly 30% of all private fusion money, to build SPARC, a compact tokamak in Devens, Massachusetts. In September SPARC's cooling plant reached its operating temperature, and CFS is aiming for its first plasma that gives back more energy than it took to heat it in 2027.

The Fujikura order tells you where its head is. CFS is already buying tape for ARC, its first power plant, before SPARC has proved anything.

Helion Energy

Helion made fusion's boldest promise: 50 megawatts to Microsoft from its Orion plant in Washington state. Its Polaris prototype became the first private machine to fuse deuterium and tritium in February, and it raised $465 million in June at a $15.5 billion valuation, later topped up to $500 million.

It's also the first to blink. The 2028 date now means "initial operations," with full power in 2029 or 2030. Helion publishes very little data, so outsiders can't check its numbers. Polaris producing net electricity is the result that would settle the argument.

Pacific Fusion

Pacific Fusion launched in 2024 with a $900 million first round, now past $1 billion raised, and broke ground on a $1 billion campus in Albuquerque on August 25. Its machine dumps a massive electrical pulse into a fuel target the size of a pencil eraser. It's aiming for a whole machine that comes out ahead by 2030.

It also signed a memorandum with the National Nuclear Security Administration, which wants those fusion bursts for weapons research. That's a government partner before Pacific Fusion sells a single kilowatt-hour.

Fusion's first big product is going to be cheap superconductor tape.

Every magnet-based reactor needs thousands of kilometers of superconducting tape, and until recently almost nobody made it at scale. Fusion changed that. Orders like CFS's are paying for the factories that will drive the cost of tape down, years before any reactor needs to pay for itself.

Cheap superconductors are useful far beyond fusion. Superconducting grid cables carry several times the power of copper in the same trench. They also make lighter motors for ships and aircraft and cheaper MRI machines. None of that needs a plasma to behave.

So even if every fusion timeline slips another decade, the tape industry fusion built is already real and already getting cheaper. For nuclear fusion stocks in 2027, the most bankable thing fusion produces is the supply chain, not the power.