Fusion, and other fusion-level bets: Part I

The post opens a series about what he calls fusion-level bets, and the definition does the work. What civilization needs is not fusion as such but something with fusion’s properties: primary energy that is globally scalable, ultra-low-carbon, independent of climate and geology, light on land, safe and affordable. He adds a second requirement that is easy to miss. Hydrocarbons do two jobs, and fusion answers only one of them. Coal, oil and gas are abundant energy sources, but they are also unusually good energy carriers, combining density, stability and ease of transport, and replacing that second function would be a comparable milestone in its own right. He says he knows of six such bets, two aimed at primary energy, two at energy carriers and two at both under the right circumstances, and that each will get the same three headings: the prize, the challenge, and the path to commercialization.

Part I then turns the series’ own test on fusion and very nearly fails it. The prize fusion offers is a better version of something already commercialized in the 1960s. Fission runs today, its safety record is better than almost every alternative, and spent fuel, though a genuine long-term risk, is in his judgment a very manageable one. What has held fission back, on his account, is public fear translated into regulatory cost, and he quotes his own post of three years earlier at length on that point. He then updates the quoted passage rather than endorsing it: opinion has begun to turn, nuclear is one of the few bipartisan subjects in Washington, and in countries that have made it a national priority plants are again finishing on time and on budget. Fission also does the carrier job. Uranium-235 is millions of times more energy-dense than diesel, which is why the navy’s long-deployment ships run on it, and there is no realistic fusion concept at that power density. Much of the space fusion is supposed to occupy turns out to be occupied already.

The challenge explains why the comparison is so unforgiving. A fusion reactor has to hold its fuel dense enough and hot enough for long enough to release more energy than starting the reaction consumed, which means confining a plasma at a hundred million degrees or more without a star’s gravity to do it. The ratio of energy out to energy in is the field’s scoreboard, and the break-even point has never been reached, at any cost. The two candidate reactions trade against each other: the easier one to ignite needs a fuel that is barely present in nature and throws off neutrons that wreck the machinery, while the one with abundant fuel and no neutron problem needs roughly ten times the temperature. He reports an elegant fix for the first case, a flowing liquid-metal blanket that both absorbs the neutrons and breeds the scarce fuel out of lithium, but the fix is another large engineering system bolted onto an already demanding one.

He still concludes the prize is real, for two reasons. Uranium is finite, geographically concentrated and dependent on a mine pipeline he reports as insufficient to support demand growth from the early 2030s, whereas fusion fuel is effectively everywhere. And fusion is inherently safe in a way fission is not, because the reaction stops the moment you stop feeding it, which should let fusion plants skip the containment structures, exclusion zones and armed guards whose cost he blames for fission’s spiral. On the path, he separates the scientific milestone from the commercial one, and that separation is the argument’s hinge. Even granting his own comparatively optimistic timeline for a first grid-connected plant, fission will in all likelihood still be significantly cheaper, at least where regulators allow it to be. The threshold fusion actually has to clear is a lower bill of materials than fission, which he judges much harder than reaching break-even. So the bet is worth making for the long-term future while fission, in his words, is worth tripling down on now.