The Year Fusion Stopped Being a Joke

For decades, fusion power has been “30 years away” — a punchline for anyone who’s followed energy technology. The promise of unlimited clean energy from the same reaction that powers the sun, always tantalizingly close, never quite arriving.

That joke doesn’t land anymore.

What Changed

Two things happened that weren’t supposed to:

First, the magnets got better. High-temperature superconducting (HTS) magnets — a technology once confined to research papers — became manufacturable. These magnets create stronger magnetic fields, which means you can confine fusion plasma in a smaller device. Smaller devices iterate faster. Faster iteration means you actually learn things and make progress.

Commonwealth Fusion Systems, an MIT spinout, built a factory in Massachusetts that produces these magnets at scale. This isn’t a lab demonstration. It’s industrial production.

Second, private capital showed up. Over $10 billion has now flowed into private fusion companies. Not government research grants that drip out over decades, but venture capital and corporate investment that expects results on a startup timeline.

When you have billions of dollars and HTS magnets, the engineering timeline compresses dramatically.

The Actual Timeline

SPARC, CFS’s demonstration reactor, is expected to produce its first plasma this year (2026). Net energy — more power out than in — is expected by 2027. Their first commercial plant, ARC, will deliver 400 megawatts to the Virginia grid in the early 2030s.

Helion Energy has already signed a power purchase agreement with Microsoft: 50 megawatts of fusion power, contracted for delivery by 2028. Microsoft isn’t known for signing contracts for vaporware.

TAE Technologies plans their commercial plant, Da Vinci, to be grid-ready in the early 2030s.

These aren’t hopes. They’re business plans with signed contracts.

The Record Book

In case you think this is still theoretical physics:

  • February 2025: WEST (France) sustained plasma for 1,337 seconds — over 22 minutes — at 50 million degrees Celsius. World record.
  • 2024: JET (UK) generated 69 megajoules from 0.2 milligrams of fuel. Record energy output.
  • 2024: KSTAR (Korea) sustained plasma at 100 million degrees for 48 seconds. They’re targeting 300 seconds this year.
  • December 2022: National Ignition Facility achieved ignition — more energy out than the laser put in. The physics is settled.

Why Now Matters

The timing is almost eerie. AI is scaling to consume entire power grids. Data centers are being built in jurisdictions based primarily on energy availability. Tech companies are signing deals with nuclear plants and funding fusion research specifically because they need clean, reliable, massive power sources.

Fusion and AI may be locked in a symbiotic relationship: AI helps optimize plasma control (DeepMind demonstrated this in 2022); fusion provides the clean energy AI’s growth demands.

The Real Question

Fusion is no longer a question of “if.” It’s a question of “when,” and the answer is: this decade for demonstrations, next decade for scale.

The better question is whether it arrives fast enough. Climate targets. AI energy demand. Grid reliability. The 2030s are when fusion starts contributing, but the 2040s are when it might actually dominate.

The joke about fusion being 30 years away is finally outdated. Replace it with: “Fusion is 5-10 years away, and the contracts are signed.”

That’s not a punchline. That’s a schedule.

Sources: IAEA World Fusion Outlook 2025, MIT News, Commonwealth Fusion Systems, industry reports