The Fusion Race Is Real Now: Four Countries, Four Breakthroughs, One Year

The Punchline That Stopped Being Funny

Fusion energy has been "30 years away" since the 1950s. It's the oldest running gag in science. Physicists roll their eyes when you bring it up. Investors used to laugh.

Nobody's laughing in 2026.

In the first three months of this year, four separate countries hit fusion milestones that would've been front-page news a decade ago. China broke a plasma limit scientists thought was impossible. France kept a fusion reaction running for over 22 minutes. An MIT spinoff in Massachusetts started assembling an actual reactor. And South Korea is gunning for a record that could change everything.

Something shifted. And most people missed it.

China: Breaking the "Unbreakable" Limit

On January 1, researchers at China's EAST tokamak — yes, their "artificial sun" — published a paper in Science Advances that made plasma physicists sit up straight.

They'd exceeded something called the Greenwald density limit. In plain terms: there's a theoretical ceiling on how tightly you can pack plasma particles inside a fusion reactor. Go past it, and the plasma destabilizes. The reaction dies.

EAST went past it. The plasma held.

The technique involved a method called ECRH-assisted ohmic start-up — using microwave heating to stabilize the plasma at densities that should've caused a collapse. The team plans to push further in follow-up experiments.

Why it matters: denser plasma means more fusion reactions per cubic meter. That's the difference between a science experiment and a power plant.

France: 22 Minutes and Counting

On February 12, the WEST tokamak in southern France held a plasma for 1,337 seconds. That's 22 minutes and 17 seconds — smashing the previous record for sustained plasma confinement in a tokamak.

This isn't about temperature. WEST wasn't trying to be the hottest reactor on the planet. It was testing endurance. Can you keep a fusion reaction stable long enough to actually generate useful energy?

Twenty-two minutes says: getting there.

The WEST team, part of the EUROfusion consortium, built their reactor specifically as a test bed for ITER — the massive international fusion project being constructed nearby in Cadarache. Every second of sustained plasma at WEST feeds directly into ITER's design and operational plans.

The catch: ITER itself is years behind schedule and billions over budget. Originally estimated at 5 billion euros, costs have ballooned to somewhere between 18 and 20 billion. Manufacturing faults, pandemic delays, and the sheer complexity of building a first-of-its-kind machine have all taken their toll.

But WEST's record shows the underlying physics works. The engineering just needs to catch up.

Massachusetts: Magnets Going In

While government-backed projects grind through bureaucracy, a private company outside Boston is building fast.

Commonwealth Fusion Systems — spun out of MIT and backed by over $2 billion in private funding — announced at CES 2026 that it had installed the first of 18 massive superconducting magnets in its SPARC reactor.

Each magnet weighs 24 tons. They're D-shaped, made from high-temperature superconducting tape, and they're the core technology that makes CFS's approach different. Stronger magnets mean you can build a smaller reactor. Smaller reactors are cheaper, faster to construct, and easier to iterate on.

CFS aims to turn SPARC on next year. If it works — if it achieves net energy gain — the company plans to build ARC, a commercial fusion power plant, shortly after.

They're not alone in the private sector. Helion Energy broke ground on its own fusion plant last year, with a power purchase agreement to supply Microsoft with electricity by 2028. Private fusion companies have collectively raised over 13 billion euros globally. The Fusion Industry Association is asking the US government for another $10 billion.

The money is real. The hardware is real. The timelines are aggressive but not fictional.

South Korea: The 300-Second Target

South Korea's KSTAR reactor — another "artificial sun" — set a record in 2024 by sustaining plasma at 100 million degrees Celsius for 48 seconds. Their target for 2026: 300 seconds.

Five full minutes of plasma hotter than the core of the sun.

KSTAR's research director Si-Woo Yoon has called 300 seconds a "critical point" for scaling up fusion operations. The logic: if you can maintain stable plasma for five minutes, you've proven the physics work at a duration that matters for continuous power generation. Everything shorter is a sprint. Five minutes starts to look like a marathon pace.

Whether they hit it this year remains to be seen. But the trajectory — from 30 seconds in 2021 to 48 seconds in 2024 — shows consistent progress, not hype.

The Bigger Picture: Why 2026 Is Different

Previous fusion "breakthroughs" were isolated achievements. One lab would set a temperature record. Another would sustain plasma for a few seconds longer. Interesting, but easy to dismiss as incremental.

2026 is different because the breakthroughs are happening everywhere at once — and they're addressing different problems simultaneously.

China is solving density. France is solving duration. The US private sector is solving engineering and economics. South Korea is solving sustained high-temperature operation.

These aren't competing approaches. They're complementary pieces of the same puzzle. A working fusion power plant needs all of them: dense, hot, stable plasma maintained long enough inside a reactor that's affordable to build and operate.

The investment picture has changed too. Private fusion funding hit $13 billion in 2025 — up from essentially zero a decade ago. The US Department of Energy released a formal Fusion Science and Technology Roadmap. China has made fusion a national priority. The EU is funding it through ITER and a network of supporting experiments.

What Could Go Wrong

Plenty. ITER's delays are a warning that big fusion projects are brutally hard to execute. CFS's SPARC has never been turned on — there's a gap between installing magnets and achieving net energy. Helion's 2028 deadline for delivering power to Microsoft is ambitious by any measure. And KSTAR's 300-second target is exactly that: a target, not a result.

The realistic timeline for commercial fusion power — electricity in the grid, powering homes — is still the 2030s at the earliest, more likely the 2040s or 2050s for widespread deployment.

But that timeline has been compressing. A decade ago, serious scientists said 2060. Now credible estimates cluster around 2040. That's still far away. It's also moving in the right direction, faster than it used to.

The Question That Matters

Fusion won't save us from climate change in time. The math doesn't work — we need massive renewable deployment right now, and fusion can't contribute for at least another decade.

But fusion could define the second half of this century. Near-limitless clean energy would reshape everything: water desalination, industrial manufacturing, transportation, agriculture. The countries and companies that crack it first won't just generate electricity. They'll generate power in every sense of the word.

For 70 years, fusion was a punchline. In 2026, four countries are building the setup for a very different kind of delivery.

The joke might finally be over.