When will we get serious about nuclear? (or something ‘nuclear-esque’?)

The forecast rests on a premise he imports from elsewhere in his series and does not argue here: transmission bottlenecks and siting constraints on wind and solar will become major limits on renewables growth through the 2020s and be undeniable by the early 2030s. The world will then conclude, he expects, that full decarbonization needs another globally scalable zero-carbon primary source, ideally firm and, more crucially to him, power dense enough to drop into populated areas: a coal plant without the emissions. That resource already exists, and has for decades: a fleet designed before the computerized spreadsheet that still runs reliably and that he argues is the safest energy source per unit produced. So the constraint is not technological.

It is public and then regulatory, in that order, and the order is the point. The partial meltdown at Three Mile Island in 1979 interrupted an American construction boom; the response stretched timelines and ballooned costs on plants already under way, and the country then stopped building for twenty years. Polls still register strong disapproval, and where the public turns against nuclear, governments follow; no program survives without a supportive or at least neutral policy framework. His account of the Nuclear Regulatory Commission is that its official job is safety while its effective one is making new plants uneconomic to build and new designs unaffordable to license, and his evidence is a licensing bill, a Georgia overrun, and a reform the commission’s own staff proposed in 1999 and never delivered. Countries with friendlier opinion and what he calls rational regulation have meanwhile kept building at reasonably competitive cost on much the same light-water technology the United States pioneered.

He then repeats that deployment needs no new technology, and concedes that new technology might matter anyway, because it could address the root problem of acceptance. He is cool on small modular reactors: scale economies in single large plants are demonstrated while factory scale economies are theoretical, and he does not believe modularity fixes public perception on its own. Passive safety is the design change aimed at perception, and the wider Gen IV field at everything else. But he suspects the number of competing designs is itself holding the industry back, since fifty deployments of one design would serve a renaissance better than one deployment each of fifty, which bites hardest on the small modular reactors that need volume to get cheap. The irony he draws is that the United States hosts many of these developers and is not on track to be their proving ground; East Asia and Eastern Europe, with less abundant renewable resources, reached the same conclusion first, and he expects their policy and domestic industrial preferences rather than any technical review to pick the winners.

The final turn is that none of this has to be fission. The gap needs a profile rather than a fuel: inexhaustible, zero-carbon, power dense, geography independent, around the clock. Fusion fits in principle, because the reaction cannot run away and its waste is far less radioactive, so it ought to win public opinion and a lighter regulatory touch. Private investment has brought it close, he thinks, but he is careful about what close means: near to demonstrating scientific milestones, a long way from commercial viability, and most approaches need near-perfectly shaped superconducting magnets, or ultra-powerful lasers that do not quite exist yet, or both. Deep geothermal is the other candidate, available almost anywhere if you drill far enough, and the drilling is the problem: exploration risk like oil and gas but a smaller prize, since heat is worth far less than liquid fuel, and a cost at useful depths nowhere near competitive. A breakthrough in deep drilling, he concludes, would be practically equivalent to fusion.