Fusion power might never be cheap, and that’s a problem

Fusion power has been the great “almost here” technology for decades. The promise is seductive: limitless, zero-emissions electricity, no meltdowns, no long-lived radioactive waste. But a new study published in Nature Energy pours cold water on the economic side of that dream.

The core issue is something called the experience rate — the percentage by which a technology’s cost drops every time its installed capacity doubles. Solar modules have an experience rate of 23%. Lithium-ion batteries: 20%. Onshore wind: 12%. These are the poster children of clean energy cost declines.

Fusion? The researchers estimate it’ll land somewhere between 2% and 8%.

That’s better than fission, which sits at a dismal 2%. But it’s nowhere near the 8% to 20% that many energy modeling studies currently assume. And it means fusion electricity could stay stubbornly expensive for a very long time, even after we figure out how to build the damn things reliably.

The study’s methodology is interesting. The team, led by Lingxi Tang at ETH Zurich, interviewed fusion experts and asked them to rate future power plants on three characteristics that historically correlate with low experience rates: unit size, design complexity, and the need for customization.

Fusion plants will be big. They generate heat, like coal or fission plants, so economies of scale don’t work the same way they do for modular technologies like solar panels. They’ll probably need less customization than fission plants — safety regulations should be simpler — but more than solar. And on complexity? “There was almost unanimous agreement that fusion is incredibly complex,” Tang says. Some experts literally said it was off the scale the researchers gave them.

So the final range of 2% to 8% makes sense, but it’s also sobering. To get fusion costs down significantly, you need a lot of deployment. And “a lot” likely means decades, not years. The US government allocated over $1 billion to fusion in fiscal 2024. Private-sector funding hit $2.2 billion between July 2024 and July 2025. That’s real money.

Tang doesn’t mince words: “If you’re talking about decarbonization of the energy system, is this really the best use of public money?” It’s a fair question when solar and wind are already cheap and getting cheaper, and when grid-scale storage is finally becoming viable.

But I also think the historical analogy cuts both ways. Egemen Kolemen at Princeton Plasma Physics Lab points out that in 2000, analysts predicted solar would stay expensive. Then China went all in, production exploded, and prices crashed. “People weren’t exactly wrong then,” he says. “They were just extrapolating what they saw into the future.”

The same could happen with fusion. A geopolitical push, a manufacturing breakthrough, a regulatory shortcut — any of these could change the curve. The study acknowledges it only looked at magnetic confinement and laser inertial confinement, the two dominant approaches. Other designs might have different cost profiles.

Still, I find the study’s conclusion hard to dismiss. Fusion’s inherent complexity and scale are real constraints. You can’t mass-produce a tokamak the way you stamp out solar panels. The tolerances are insane, the materials exotic, the engineering unprecedented.

So where does that leave us? Fusion might eventually work. It might even become a meaningful part of the energy mix. But anyone banking on it being cheap should probably adjust their expectations. We’ve been waiting 70 years for fusion to arrive. We might be waiting a while longer for it to be affordable.