On Monday, Energy Minister Tim Hodgson stood at a podium in Newmarket, Ontario, and unveiled Canada’s new Nuclear Energy Strategy — a document that calls for up to 10 new reactors by 2040, a doubling of uranium exports, and an aggressive global sales push for Candu technology. Hodgson called it a “new civilian nuclear renaissance” and declared there is “no credible plan” to double the grid and build a low-carbon economy without nuclear baseload power. The announcement landed with predictable enthusiasm on Hacker News, where the comments quickly sorted themselves into the familiar camps: nuke-pilled optimists celebrating baseload triumphalism, and the usual skeptics muttering about waste and safety.

But something was missing from the strategy document and from Hodgson’s remarks. It’s the same thing that’s been missing from every Western nuclear announcement for two decades, and it’s the thing nobody in the cheering section seems to want to ask about.

The strategy contains no cost estimate. Not a single dollar figure for what 10 reactors will cost Canadian taxpayers, ratepayers, or both.

Vogtle, Hinkley, Olkiluoto: the numbers that already exist

Canada doesn’t need to guess what new nuclear costs. It has three recent Western case studies, and they all point in the same direction.

Georgia Power’s Vogtle Units 3 and 4 — the only new reactors built in the United States this century — came in at roughly $35 billion for two AP1000 units, more than double the original $14 billion estimate and seven years late. Hinkley Point C in the United Kingdom is now projected at £46 billion for two EPR units, up from an initial £18 billion. Finland’s Olkiluoto 3, a single EPR, took 17 years from construction start to commercial operation and cost roughly €11 billion — about triple the original €3.2 billion estimate.

These are not cherry-picked horror stories. These are the only data points available for new-build Western nuclear in the 21st century. If Canada builds 10 reactors, it will not be immune to the forces that produced those numbers — and those forces are not primarily technical.

One project-finance lawyer who works on infrastructure deals and asked to speak candidly about a client’s portfolio put it this way: “The reactor design is the easy part. The regulatory uncertainty, the supply-chain atrophy, the skilled-labor shortage — those are the things that compound. And they compound at 7 percent a year on a construction loan.”

What the strategy actually funds

The strategy is not an appropriations bill. But the only concrete dollar figures attached to it so far are modest: a $40-million investment announced in April to explore a microreactor in the North, and various regulatory-streamlining commitments. The rest is ambition — and ambition is not a financing mechanism.

This is the pattern. Governments announce reactor targets with fanfare. The private capital that would actually need to build them stays conspicuously quiet. The reason is straightforward: no utility or pension fund will finance a first-of-a-kind nuclear project in a jurisdiction that hasn’t built one in decades unless the government backstops the construction risk. Ratepayers, in turn, get handed the bill through pre-construction cost recovery or mandatory capacity payments — mechanisms that make the $35 billion at Vogtle look like a consumer-protection success story by comparison.

Hodgson’s strategy nods at this obliquely, mentioning “innovative financing approaches” without specifying them. When a government document uses the phrase “innovative financing,” it usually means one of two things: off-balance-sheet borrowing, or a bill that arrives after the next election.

Candu’s export problem

The other pillar of the strategy — selling more Candu reactors abroad — has its own quiet difficulty. Canada currently services nine Candu reactors in countries including Romania, India, and South Korea. The global market for heavy-water reactors is narrow. Most countries pursuing new nuclear are opting for light-water designs: the AP1000, the EPR, the APR-1400, or small modular reactors from vendors like NuScale and GE Hitachi. Candu’s technology has real merits, particularly its ability to run on natural uranium. But it is not the design the market is choosing.

If Canada wants to be a nuclear exporter in 2040, it will need to compete on price and delivery timelines against South Korea’s KEPCO, which built the Barakah plant in the UAE on budget and ahead of schedule, and against Russian and Chinese state-backed vendors that don’t face the same export-credit constraints. Saying “we’ll sell more Candus” is not a strategy; it’s a hope dressed up in policy language.

The actual argument for doing this anyway

None of this means nuclear is the wrong call. The case for building reactors — particularly if Canada’s grid genuinely needs to double by 2050 — is strong on its merits. Nuclear provides clean, dense, reliable power that wind and solar cannot replicate without storage at a scale that doesn’t yet exist. If the alternative is burning more natural gas forever, the cost overruns on a reactor start to look like a bargain when measured against 60 years of emissions.

But that argument requires honesty about what things cost and who pays. The strategy released Monday is instead a document that talks about “energy sovereignty” and “global leadership” while being silent on the one number that will determine whether any of these reactors actually get built. Canada has just watched Georgia, Finland, and the UK learn the world’s most expensive lesson about nuclear construction timelines. Pretending the same dynamics won’t apply north of the border is not ambition. It’s avoidance.

A construction manager who spent four years at Vogtle and now consults on megaprojects in Alberta put it more bluntly. “Until the press release includes a cost range and a financing structure, the reactors are imaginary. They’re political reactors. They generate press, not power.”

Ten reactors by 2040 is a vision. Visions are free. The concrete has not been poured, and no one in Ottawa has yet named a price.

Sources