Is Fusion Commercialization in Sight? Not Yet, Says John Holdren

Scientists at Lawrence Livermore National Laboratory made history when they achieved fusion ignition in December 2022, producing more energy from a nuclear fusion reaction than was used to trigger the reaction. Since then, they have repeated the feat four times.

Scientists, governments, and companies have been trying for decades to develop fusion power, which could potentially serve as a source of cheap, carbon-free electricity without much of the long-lasting radioactive waste produced by traditional nuclear fission. 

The recent advances at LLNL set off a wave of investment in companies working on fusion as a power source, and the United States and Japan announced a new partnership on April 10 to accelerate the technology. But some experts are skeptical that commercialization is around the corner. 

"I don’t want to be the skunk at the garden party, but I would say that predictions of commercial fusion by 2030 or 2035 really are hype at this point," said John Holdren, Co-Director of the Belfer Center's Science, Technology, and Public Policy Program, during a United States Energy Association virtual press briefing on Wednesday, April 3. 

While tremendous progress has been made on the physics side of the fusion puzzle, according to Holdren, there are still major obstacles to be overcome before fusion reactors can contribute electricity to the power grid.

Watch a recording of the press briefing and read a transcript of selected remarks by Holdren below. 

Selected Remarks by John Holdren

How close are we to achieving commercialization of fusion?

I don’t want to be the skunk at the garden party, but I would say that predictions of commercial fusion by 2030 or 2035 really are hype at this point. We haven’t even yet seen true energy breakeven, where true energy breakeven means the fusion reactions are generating more energy than had to be supplied to the facility. It is not true energy breakeven simply to get more energy out than what reached the fusion fuel, for example. 

And after we get true breakeven, we need to get sustained operation at, in inertial confinement cases, repetition rates that are meaningful for practical fusion or, in the case of magnetic fusion, sustained operation for more than a few minutes, more than a few hours. Ultimately, these plants are going to have to run for thousands of hours per year if they are to be economical. 

When we get past those thresholds, which, again, we have not yet really approached in many cases, we’re going to have to deal with a large constellation of technical challenges: dealing with the neutron fluxes and their effects on the structural materials in the reactor, dealing with tritium recycling with extremely high containment rates so that you don’t exceed plant boundary restrictions on how much tritium you let out. These are going to be tough questions. They are perhaps comparable in difficulty to the physics questions that have now taken about seventy years and probably a hundred billion dollars to solve and we’re not even yet there on the physics.

I worry about this because being over-enthusiastic about early commercial applications of fusion is not only bound to generate disappointment but it feeds the public’s belief in technological miracles that will save us from the difficult task of dealing with climate change and other energy challenges with the options that are closer to practical reality.

What will the commercialization of fusion look like? How much will fusion power cost?

I think part of the hype has not just been about how quickly we’ll have commercialization, but what a miracle it will be once we do. 

I think fusion is well worth striving for. I began my career in it in 1965 as a graduate student at MIT working on fusion. I still think it is a spectacularly important goal, but it is not a holy grail. And the claims that it will be dirt cheap, the claims that it will create absolutely no radioactive materials – this is nonsense. The fuel will be cheap, but as with fission, where the fuel is also cheap, the key to the cost is going to be how much it costs you to build and operate the facility, not how much it costs to fuel it. 

I’ve been involved, as Steve Dean has been, over many years, in sharing and participating in studies of the future of fusion, and all of those studies have led to the result that it would be a great challenge to make fusion cheaper than fission. It’s possible. It cannot be ruled out. We should strive for that. But the fact that fuel is abundant and cheap is not a prescription for cheap energy, and the fact that fusion is cleaner than fission doesn’t mean that it’s absolutely clean. The first generation plants will have a lot of tritium and neutrons will create radioactivity in the structural materials. It’s less than the radioactivity involved in fission. That’s desirable. We should aim for it. But we should not mislead the public that this is the holy grail of absolutely clean, dirt cheap energy. Ain’t going to happen.

When will fusion be economically competitive with other fuels?

I don’t think we know what the time scale for competitiveness is going to be, and we don’t know, in part, because there is, as others have said, enormous potential for innovation in this space. Nobody has a clear crystal ball for what innovation will produce over the next twenty years. But I think when one looks at the obstacles that have to be overcome, it is, in my view, safe to say we will not see economically competitive fusion on the power grid before 2050. If I were a betting person and I expected to live to collect, I would place a bet on that.

Again, that’s not to stay that there hasn’t been tremendous progress. There has been, both in the inertial confinement fusion and in the magnetic fusion space. I would also say, by the way, in answer to something that was asked on that chat a while ago, magnetic fusion is actually closer to true breakeven than inertial confinement fusion is at this point. The best magnetic devices have come much closer to generating as much energy as was supplied to the system, and they’ve also come closer to the duration requirements. Tammy Ma talked about replicating the ignition phenomenon four times over the space of a year. A commercial fusion power plant would have to replicate that about ten times a second. They’re really a long way off.

If and when fusion is commercialized, will fusion power plants be connected to the power grid?

There is a variety of possibilities, as Diana Grandas has correctly said, but I think that the betting is that the largest contributions from fusion are most likely to come from grid-connected systems for a number of reasons. 

The requirements of a fusion power plant are such that there is going to be a considerable premium on centralized use of those power plants on a scale where-- economies of scale in how you deal with tritium, in how you deal with energy management, in how you deal with maintenance, are going to, I think, drive us in the case of the biggest applications to grid-connected systems. This is not to deny Diana’s correct point that there is an immense amount of diversity in the approaches that are being pursued now. Some of them might well be attractive in off-grid applications. But I think the probability is that the grid-connected ones are going to be the biggest contributors.