The joke is older than my parents: fusion is the energy source of the future, and always will be. Thirty years away, forever. Every energy writer inherits the joke with the beat, and every energy writer eventually has to decide what to do with it.
I am writing it down this summer because of a machine. In a hall outside Boston, a company spun out of MIT is assembling a device called SPARC around a set of magnets that did not exist, anywhere, when the joke was coined: high-temperature superconducting coils that produce fields of twenty tesla, several hundred thousand times Earth's magnetic field, the strongest ever built for a fusion device. The first of eighteen went in this winter. The full ring is due in place by summer's end, with the switch-on campaign targeted for next year.
And in July something happened that I found more remarkable than any plasma physics. Pension funds, the institutions that manage teachers' and firefighters' retirement money, the most careful money there is, put a billion dollars into the company. Fusion has been funded by governments, billionaires, and true believers for seventy years. It has never before been funded by actuaries, which is a boring sentence that took seventy years to become true.
One paragraph of physics, as plain as I can make it. Fusion means squeezing light atoms together until they merge and give off energy, the same thing the sun does. To do that you have to hold a plasma at a hundred million degrees, and nothing material can touch something that hot, so you cage it in magnetic fields. The key fact of the moment is this: the fusion power you can wring from a given machine scales roughly with the fourth power of the magnetic field strength. Double the field, and the same-sized machine gets sixteen times better. That is why the old road to fusion ran through sheer size, like the international ITER project in France, big as a cathedral and budgeted in decades. With the old magnets, the only fix was to build bigger.
The new superconducting tape rewrote that math. Fields twice as strong mean a machine a fortieth the volume can, on paper, reach the same milestone. That is exactly SPARC's bet: a tokamak that fits in a warehouse, built by a few hundred people in a few years, designed to produce meaningfully more energy from its plasma than goes in. Not a power plant. SPARC is the demonstration; the power plant, already under development with a utility site in Virginia, comes after. But it would be the proof, at commercial scale and commercial speed, that the joke's premise is dead.
I keep coming back to the pension money, because money like that does not move for a good story. Venture capital buys stories. Billionaires buy legacies. Governments buy options. Pension funds buy probabilities: long, boring streams of value that have to actually show up. Their arrival does not mean fusion is solved. Smart money is wrong all the time, and this summer's energy panic, a strait closed, oil spiking, every government on Earth rediscovering the words "energy security," has made any domestic, fuel-free power source look temporarily irresistible. So I discount for that.
But the bet is on something more specific than "fusion works." The wager is that the sequence now has acceptable odds: magnets, which are built, the hardest engineering already bolted to the floor; then plasma, which is next year's exam; then net energy; then a first plant in the 2030s. Every step has a way to fail. Plasmas have humiliated seventy years of confident schedules, and this column will report the humiliations if they come. The honest status is that fusion has moved from unknown physics to brutal engineering with a checklist, and brutal engineering with a checklist is the kind of problem people turn out to be good at. That move between categories is what a billion careful dollars was buying.
The fine print
Even on the rosiest schedule, fusion arrives too late to matter for the 2020s and most of the 2030s, the decades the climate math actually cares about. The transition being won right now is being won by the unglamorous things this column covers all year: solar, which just helped renewables pass coal; batteries; grids; efficiency. Anyone selling fusion as a reason to relax about this decade is selling delay with better physics.
Fusion's real promise sits one horizon out. The 2040s-and-after question is firm, dense, always-on clean power for a world that will by then have electrified nearly everything and doubled its demand, the steady floor under the solar century. That is not a consolation prize. It is the difference between a transition that stalls partway and one that finishes.
The plasma will give its verdict soon enough. I don't know which way it breaks, and I am suspicious of anyone who says they do. What I know is that the joke finally has a date attached, and the world's most cautious money is in the room, watching. This summer, in a hall outside Boston, the rest of the eighteen magnets go into the ring. Next year, somebody throws the switch.
