did fusion really achieve net gain, and if so, what happens next?

It’s a common joke in the popular science world that fusion is always just 20 years away. Ever since we’ve been able to create thermonuclear weapons, which use fusion to create massive blasts capable of leveling cities, physicists have been obsessed with harnessing that energy to revolutionize the world. Unfortunately, containing the kind of explosions that power stars in human made devices has been far more difficult than they ever anticipated, but researchers kept drifting closer and closer to sustainable fusion reactors.

This brings us to the National Ignition Facility in California, which was created to figure out how to start a viable fusion reaction and learn how it could become a useful power source. For the last decade, shot after shot of its powerful laser array came tantalizingly close to reactions that at least returned as much energy as was put into them, but never quite crossed the line. That all changed on December 5th, when a 2.05 megajoule pulse compressed a tiny fuel pellet which exploded with 3.15 megajoules of energy. For the first time ever, fusion produced net gain.

Of course, there is a significant caveat to this achievement. While the lasers compressing the fuel of deuterium and tritium gas in a trillionth of a second put out around two megajoules of energy, they required well over 400 megajoules to charge and fire. The reaction saw significant net gains. The entire system? Not so much. Yet, that said, being able to successfully predict the gainful reaction in a computer model and make it happen using a decades old design is no small feat, and it shows that our understanding of the required physics is improving.

why is fusion so difficult?

How on, you might say, fusion is the most common reaction in the universe. We even know how to create it for a few milliseconds reliably enough to turn the mechanisms for igniting it into bombs. Why is it so hard to harness it without a massive explosion? How do stars keep it going for billions of years? Well, there are two ways to get fusion: insanely hot temperatures and immense pressure. Stars have monstrous gravity which generates the necessary pressure and stabilizes the reaction. We don’t have that nifty tool at our disposal.

Instead, we have to use heat in one of two general ways. The first is inertial confinement, like the NIF reactor. Precise arrays of 192 or more lasers are focused on a hollow tube filled with deuterium and tritium, and fire on it in unison so within a picosecond, the gas implodes into a searing hot plasma that creates helium. The other approach is to generate 150-million-degree plasma and keep it churning in dense enough currents to heat deuterium and tritium gas for at least 30 minutes in a device called a tokamak thanks to a somewhat clunky Russian acronym.

In inertial confinement, the idea is to have ongoing, rapid-fire shots, which means laser designs that can maintain amazing precision while putting out massive power. A slight defect in the fuel pellet or tiny misalignments in the lasers’ aim thanks to wear and tear, or interference from the plasma, and the reaction loses its potency and control. In tokamaks, magnetic fields that allow the plasma even a little bit of diffusion mean the reaction quickly breaks down as the ionized gas forms unstable, chaotic currents that cool and dissipate. In short, fusion is a finicky mistress unless you have roughly 70 Jupiters worth of gravity at your disposal.

what have we learned from this experiment?

Even though NIF’s inertial confinement reaction is only a net gain if you don’t count how much energy went into charging the lasers, it showed us that the breakeven point for these designs was a lot lower than that 5 megajoules we thought. Using far more efficient lasers, we may be able to optimize the system and inch closer to total net gain rather. On top of all this, the shot confirms that our fusion simulations are getting more accurate, predicting the results within tens of kilojoules.

Building on the NIF’s proof of concept, we can more accurately dial in what energies need to be achieved for break even and set realistic goals rather than estimate and be disappointed when yet another previously unknown problem gets in the way, as has been the patten for just about every fusion experiment for the past half century. Yet, at the same time, we finally appear to be running out of problems to solve before we start hitting break even points, which is why it was so exciting to see the latest results from the NIF.

This is also excellent news for tokamak designs, which have also been getting closer and closer to their own breakeven points. Just a little bit over a year ago, an experimental device called JET annihilated records by producing 59 megajoules of energy in just five seconds. That was only a third of the input, but it yielded promising proof of concept data for its successor ITER, which will be ten times larger and is expected to achieve net gain with its countless improvements in magnets, materials, software, and yes, physical size.

where do we go next?

Unfortunately, that gain will likely come at the end of 2030s even if everything goes perfectly, since the ITER project operates on a comparative shoestring. There simply isn’t enough cash to speed things up despite the experiment intended to demonstrate how to run a fusion plant, as well as achieve a tenfold return on energy investment. This is the same problem the NIF has, which is why it’s taken so long to achieve net gain on its reactions and refine the simulations to the point at which they’re becoming instructive rather than just raise more questions.

So, what’s next for fusion experiments? We seem to finally be hitting the point in which half a century of trial and error are starting to pay serious dividends, and even chronically neglected and underfunded experiments are slowly and defiantly creeping towards major milestones and anticipated benchmarks. We’ve come a very long way from trying to catch the shockwaves and heat of nuclear blasts to constructing building sized plasma generators ran by AI models, which actually have a shot at changing the world as we know it for the better.

But today, global investment in fusion experiments is tiny. The world will spend more than six times as much on pizza than we will on building the ITER experiment, and the NIF receives less and less of what is functionally a rounding error in the Department of Energy budget each year. Given the fact that we need to make fusion work to achieve many of our loftier goals for a shot at a decent future for humanity, it’s absurd we’re not investing far more than we are to speed up the rate of experiments. And until we do, fusion will remain a slow grind to the finish line.