On September 29, pv magazine reported a number the energy commentariat has been predicting and dreading in roughly equal measure: according to the World Nuclear Industry Status Report 2026, wind generation alone overtook nuclear on a monthly basis in March 2025, and solar did it in April — the first time global monthly solar output has exceeded nuclear. From March 2025 through June 2026, one or the other beat nuclear every month. The report is overseen by Mycle Schneider, who has been counting reactors since before most people tweeting about it were born.

The milestone is real. The argument over what it means is where everyone gets lazy — including, especially, my own side.

The Capacity-Factor Objection Is Correct and Irrelevant

Yes: solar produces nothing at night, and its capacity factor is a fraction of a reactor’s. Every pro-nuclear writer will say this in the next 48 hours, and every one of them will be right, and none of them will be making an argument.

Because capacity factor is not what the milestone measures. It measures where capital went. And capital does not care about your physics.

Here is the ledger. Britain’s Hinkley Point C will now cost at least £49 billion ($65 billion) if it enters service in 2030, with another £1 billion ($1.3 billion) if the first unit slips to 2031, per The Quad Report. Industrial Info put the slip itself at €2.5 billion ($3.3 billion) — and the project has gone from £18 billion ($24.2 billion) at launch thirteen years ago to £35 billion ($47 billion) in 2015 prices. Sizewell C currently estimates at £38 billion ($50.6 billion). Georgia’s Plant Vogtle cleared $30 billion for two reactors, a figure that doesn’t include the $3.68 billion the original contractor paid the owners after going bankrupt — PIRG’s writeup of that episode is worth reading twice.

Solar did not out-compete that on efficiency. It out-competed it on scheduling: you can build 50 megawatts of it, at a known price, and have it producing before the financing on the alternative has closed.

Every Reactor Is a Bet-the-Company Wager

The megaproject model isn’t a technology. It’s an industrial organization — one design, one balance sheet, one date, and no option to stop. When it slips, the slip is measured in billions, and nobody can walk away, because the state has already guaranteed the offtake.

That is the structure the Anglophone right has spent a decade championing as the serious climate answer. Note what it requires: nationalized construction risk, regulated returns, a sixty-year horizon no private investor would underwrite on merchant terms, and a political commitment longer than any government survives. If a left-wing party proposed that for anything else, the same columnists would call it what it is.

A project-finance associate at a London infrastructure fund, answering a Slack message at eleven the night EDF’s last delay hit the wires, put it more economically than I can: “We’d love the asset. We will never own the construction risk, and neither will anyone else at this table.” That’s not ideology. That’s a spreadsheet with a memory.

The uncomfortable part is that this isn’t an argument for the other team either. Solar’s economics are propped up by mandates, tax equity, and grid rules that push its costs onto everyone else, and it earns its returns partly by making the evening peak expensive. I’ll take the win column on that another week.

The Right Is Pro-Modular and Hasn’t Noticed

The serious pro-nuclear argument in 2026 is not about Hinkley. It’s about small modular reactors — factory-built, incrementally deployable, standardized, riding a learning curve. WNISR 2026 notes SMR deployment remains limited, which is fair, and also not a rebuttal; it’s a description of an early supply chain.

But look at what that pitch actually is. Small. Modular. Manufactured in a plant and shipped. Additive rather than monolithic. Financed in tranches. Any conservative who has praised that description for a data center, a factory, or a franchise has just described the argument for distributed generation — the argument they mock when a homeowner puts panels on a roof.

You don’t have to like solar to notice the inconsistency. The lesson of April 2025 is not that photons are virtuous and fission is sinful. It’s that in energy, the incrementally deployable thing beats the bet-the-company thing, essentially always, and the political coalition that claims to believe in markets, learning curves, and dispersed decision-making has spent a decade on the wrong side of its own principle.

The honest right-of-center energy position, if anyone wants it, is unglamorous: gas for firmness, storage for the evening, a grid that actually pays for reliability, and a regulatory regime that lets nuclear compete at the size it can actually get built. None of that produces a stirring conference speech. All of it beats writing another column explaining that the milestone is a statistical artifact.

It isn’t an artifact. It’s a verdict on who can still finish a project. The industry that split the atom lost to the one that ships panels by the pallet, and it lost on delivery, not on physics.

Sources