Why Canada Never Enriched

Canada is the world's second-largest uranium producer and has never enriched a gram of it. Ontario Power Generation is now building a reactor at Darlington that will run on fuel enriched in New Mexico. The gap between what Canada has been legally permitted to build since 1955 and what it has actually built is about to stop being a historical curiosity and start being the fuel line for the next generation of Canadian reactors.

Three details further update the work I published last week. Recapping last week's piece, I found that Canada's nuclear fleet uses exclusively natural-uranium CANDU reactors, and is one of the only nations in the world to do so. Canada's decision to use these reactors comes out of a combination of restrictions on enriched uranium set by the United States in the 1950s, the existing heavy-water know-how that applied to natural reactors already gathered by Canadian scientists during the Manhattan Project, and a workaround for Canada's limited heavy-forging capacity at the time. However, when considering that Canada is the world's second-largest uranium producer, mining 24% of the world's uranium in 2024, and that every comparable Western fleet opted for enriched uranium reactors, it is puzzling that Canada has never enriched domestically. In this article I want to add further details on when restrictions on enriched uranium ended, when other US allies gained access to enriched uranium reactors, and why Canada did not switch then.

1When the door opened

The United States lifted restrictions on enriched uranium for its allies in 1954. Up to that point the 1946 Atomic Energy Act barred sharing nuclear material or information with anyone, allies included. Eisenhower signed the Atomic Energy Act of 1954 on August 30, 1954, following the December 1953 Atoms for Peace speech. The revision allowed nuclear technology and material exports conditional on the recipient pledging not to use them for weapons, and became the foundation for bilateral "Agreements for Cooperation."

Agreements between the US and other nations generally come in two tiers: research agreements and power agreements. Research agreements were put in place in roughly 35 countries, starting with Turkey in mid-1955, and running through Israel, the Philippines, Japan, Brazil and South Korea. These nations were supplied with enriched uranium — in many cases high-enriched uranium (HEU) — for small research reactors. Not power-fleet fuel. A few select allies received power agreements, including Belgium, Canada, France, Australia, the Netherlands, Norway and Switzerland. Agreements were additionally used as negotiation leverage. Power bilaterals were withheld from Italy and West Germany to push them toward Euratom, while authorizing France and the Netherlands and improving Belgium's terms. Britain is the conspicuous absence: it never needed American LEU, because Capenhurst was already running. Canada being on the power list means that Canada is not forbidden from building enriched reactors.

Subsequently, nations running natural uranium reactors began, in a fairly short time, to pivot to enriched uranium.

2The natural-uranium fleets and when they died

An event that plays a crucial part is that the Atomic Energy Commission (AEC) in the United States effectively stopped signing new enrichment contracts in 1974. The AEC was both the monopoly enrichment supplier and its own regulator, which worked somewhat well up to the 1970s when nuclear plants were government owned and operated, but started to fail when the US pivoted to privately run plants. The AEC had been selling enrichment on flexible requirements contracts, where utilities could purchase to their required demand rather than in pre-determined quantities. In 1973 the AEC rewrote the criteria to long-term fixed-commitment terms, with minimum requirements (take-or-pay) and years of advance notice required. Utility companies quickly locked in capacity before the new terms bit, and the order book against the three existing gaseous diffusion plants (Oak Ridge, Paducah, Portsmouth) filled up. By July 1974 the AEC was fully committed and stopped taking orders. Utilities that did not sign before that date — including foreign ones building reactors on the assumption that American fuel was simply purchasable — were forced out of building enriched reactors. What is crucially important for Canada is that in 1974 the argument for just purchasing enriched uranium from the US broke. Comparable nations to Canada like France, Britain, the Netherlands and West Germany all committed to their own supplies independent of the US. Canada is the notable exception, and in hindsight the absence of any committed supply meant the fallback — buy American fuel, then pivot to enriched reactors on the back of it — was never actually available.

1974 compounds. 6 March — Messmer announces the PWR programme in France, an initial tranche of 13 units that scales over the decade to 58, with Eurodif already committed. May 18 — India's test, using plutonium from a Canadian-supplied research reactor and American heavy water; this eventually produces the NSG and the 1978 NNPA, and is also Canada's own credibility disaster. August — the White House publicly affirms that the US intends to remain a reliable source of enriched uranium, a reassurance issued precisely because the order book was shut. October — the Energy Reorganization Act abolishes the AEC, splitting it into ERDA and the NRC effective January 1975, so the institution responsible for the problem ceases to exist mid-crisis. Also 1974 — the Shah lends $1 billion to the French CEA for Eurodif, entitling Iran to 10% of Eurodif output.

3France and Britain

France and Britain are the cleanest case studies and the obvious parallels to Canada, because of broadly similar economies and relationships to the United States.

After gaining access to enriched uranium from the United States, a January 1967 CEA/EDF technical report found UNGG electricity cost roughly 20% more per kWh than a PWR of the same 500 MWe output. The formal switch to light water came with the 1969 decision to abandon the gas-graphite line. In conjunction, enrichment facilities began to be built in Europe. The first was Urenco. The Treaty of Almelo was signed 4 March 1970 between the UK, the Netherlands and West Germany, explicitly to break the American monopoly on enrichment.

France sat outside Urenco and assembled Eurodif in response. In late 1969 Pompidou had initiated the enrichment project, essentially simultaneous with the reactor decision. The second European consortium formed with Belgium, Italy, Spain and Sweden; Sweden withdrew in 1974, and its place is what created the opening for Iran. France already had a military enrichment plant at Pierrelatte, decided in the late 1950s and operating through the 1960s, but it was not deployed for power generation. Eurodif was co-financed by partners who agreed to buy from it as their sole supplier of enrichment services. The financial partners took no part in operating the plant, and operation was governed solely by French law. Foreign shareholders held roughly 40%. Iran's 10% entitlement ran through a 40% stake in Sofidif, which held 25% of Eurodif. France lost the resulting arbitration at the ICC in 1991 and paid Iran over a billion dollars.

Enrichment was the final step in France running the full cycle, including mining, initial processing, enrichment, conversion to fuel, and power generation. France was able to source uranium from a colonial-era African supply.

By the end of 1973, partners committed to building at Tricastin. The target was first units by 1979 followed by full capacity in 1982, which was considered very ambitious at the time. Georges Besse I started up in 1979 at 10.8m SWU/yr, which at the time was roughly a quarter of world demand — enough product for ~90–100 reactors at 3.7% enrichment. It ran for 33 years, shutting down in June 2012.

Britain started with natural reactors for the same reasons Canada did. The McMahon Act cut off US nuclear cooperation, and Britain's only enrichment capacity was committed to the weapons programme, so the Magnox reactor ran on natural uranium metal in magnesium-alloy cladding. The AGR was the middle ground, keeping the graphite moderator and CO₂ coolant of the Magnox, but increasing gas outlet temperature to ~645°C so the plants could drive conventional steam turbines, lifting thermal efficiency to about 41% against roughly 20–33% across the Magnox fleet. Enrichment is what makes those temperatures possible, even though the AGR is built on the Magnox lineage. Uranium metal can't survive 645°C, so the fuel became UO₂; Magnox cladding can't either, and was changed to stainless steel. As stainless absorbs significantly more neutrons than magnesium alloy, natural uranium could no longer sustain criticality. Britain could pay that price because the Capenhurst enrichment facility already existed from the weapons programme. The programme, however, failed on execution. Dungeness B was contracted in 1965 for 1970 completion and only reached the grid in 1983, thirteen years late and roughly 3–4× over budget in real terms, though later stations like Torness came in far better. The programme's real defect was five designs from three consortia. Britain eventually abandoned the indigenous line for a Westinghouse PWR at Sizewell B.

4Britain's decade of indecision

Also in Britain, in February 1975, the UK government picked the SGHWR, a heavy-water reactor sharing some resemblance to a CANDU, ordering four units at Sizewell and two at Torness, 660 MWe each. However, spending cuts plus SGHWR cost escalation froze the contracts. In mid-1977 the National Nuclear Corporation recommended AGRs and PWRs instead. In early 1978 Benn abandoned the SGHWR, and the CEGB and SSEB built twin AGR stations (Heysham 2, Torness) instead. In 1980 Sizewell B was announced as a PWR; the CEGB approved the Westinghouse SNUPPS design in October 1981, and the plant became operational in 1995. Thatcher's government wanted a standardised design and had no interest in technological independence from the US.

Again, what is salient to Canada is the possibility of Britain adopting a heavy-water reactor in the mid-1970s, while most nations were opting for enriched reactors. This is possibly a point scored for a heavy-water regime, but as it did not play out, I argue it is more noise than signal.

5Overall

6What the two cases tell you about everyone else

The underlying rationale behind nearly all nations switching from natural to enriched reactors is primarily a unit-size and capital-cost problem, rather than underlying physics. Natural reactors have a ceiling of scale — UNGG topped out at around 700 MWe. Enriched fuel lets you shrink and densify the core and put it in a pressure vessel. The clearest example from the period is the January 1967 joint CEA/EDF study, which compared a UNGG and a PWR at the same 500 MWe output, in the same country, under the same cost accounting, and found natural-uranium electricity cost roughly 20% more per kWh. That isolates the fuel line as the variable rather than confounding it with different countries, eras, or plant sizes.

Of the US allies that switched, all had a fuel supply sourced first, or within two to four years of the reactor decision. Britain had Capenhurst already. France launched Eurodif alongside the 1969 decision. The Netherlands and West Germany went to Urenco in 1970. Every reactor decision was made with a fuel solution already in place. Smaller nations like Italy and Spain retired their natural-uranium units and took equity in Eurodif instead of attempting greenfield enrichment. Consortium membership was the cheap route to security of supply. Canada has never seriously considered a joint enrichment project, which may be the single most fundamental reason it did not make the switch.

After 1974 the countries still on natural-uranium lines were mostly the ones locked out on US national-security grounds, including India's PHWR programme, Pakistan and Argentina. Canada is the notable anomaly. It had allied access from 1955–56 and stayed natural anyway.

7Why Canada did not follow

So why didn't Canada opt for enriched uranium?

1. No military enrichment legacy to convert. Canada did not have a weapons programme to kick off civil enrichment. Britain's civil enrichment is a hand-me-down from a weapons programme, and France's Eurodif was built onto an existing military site at Pierrelatte/Tricastin. As Canada never had a weapons programme, there was no existing plant, or know-how to civilianise. So the option was always greenfield, which never justified the capital given an all-natural fleet.

2. The economic penalty France measured never showed up in Canada's numbers. The scaling ceiling of natural reactors was a graphite-moderator problem. Pressure-tube heavy-water didn't have it. Canada successfully scaled to 500–900 MWe at Pickering, Bruce and Darlington, all with natural uranium, and CANDU has the highest neutron economy of any commercial thermal reactor: World Nuclear Association fuel-cycle figures put natural-uranium PHWR consumption at roughly 150 tonnes of mined uranium per GWe-year, about 80% of what a comparable enriched-fuel reactor needs.

3. Greenfield enrichment was expensive in a specific way. France's financial data reflects the cost of market participation rather than a simple energy narrative. At its peak Georges Besse I drew roughly 3–4% of France's total electricity, effectively consuming the output of the dedicated reactors built alongside it at Tricastin. Centrifuges used in Georges Besse II dropped costs, but expanding Georges Besse II represents a highly optimized path, integrating four additional modules into a licensed, fifty-year-old operational hub. Even with these advantages, Orano's board approved the expansion in October 2023: a €1.7bn investment for 2.5 million SWU, a roughly seven-year build to 2030. Conversely, a Canadian greenfield project would incur heavy upfront costs for site acquisition, licensing, and novel engineering. Furthermore, Canada lacks access to restricted centrifuge designs held by Urenco and Orano. Since CANDU reactors utilize natural uranium, a domestic enrichment facility would essentially target an export market already saturated by incumbents, lacking the guaranteed offtake that secures the French model.

4. Canada had already bought the alternative. Canada faced a sunk cost. CANDU substitutes heavy-water capital cost for enrichment operating cost. Once the D₂O plants and the fleet existed, switching meant writing off a paid-for advantage to acquire a new foreign dependency.

5. The customer was a provincial utility; the vendor was selling the design abroad. Ontario Hydro, not Ottawa, drove fleet decisions. AECL's business case was CANDU export. Switching in the 1970s would have destroyed the export product at the moment it was being sold to India, Pakistan, Argentina, Korea and Romania.

6. After May 1974 the politics closed — but there was never a legal bar. India's test used plutonium from a Canadian-supplied research reactor. Canada's response was to become one of the strictest nuclear exporters in the world. Nothing has ever legally barred Canada from enriching. Professor Esam Hussein, a nuclear engineer and professor emeritus at the University of Regina and the University of New Brunswick, points out that no legal impediment exists, and that what has held Canada back is the absence of policy on enrichment and reprocessing. Frank von Hippel — a Princeton physicist, professor of public and international affairs emeritus, co-founder of the university's Program on Science and Global Security, and an established figure in American nuclear arms control and nonproliferation — says the same from the nonproliferation side: Canadian enrichment would contravene no law or international commitment, but would be hard to justify economically. The constraint was policy, not permission. The question isn't whether Canada may, but whether anyone will pay.

7. When Canada did ask, the answer was no. Steve Aplin of the Canadian Nuclear Workers' Council says Canada sought US support for a domestic enrichment facility in 2006 and was refused outright. This rests on a single source, but if accurate, the US stance is a real barrier.

8Where this leaves Canada

Canada is the anomaly on its own terms now, not just historically. It mines a quarter of the world's uranium, built a reactor line specifically engineered to need none of it enriched, led the world in that technology, and continued using it for decades after every comparable ally had switched. That combination — abundant ore, a working alternative, and permission it never exercised — was a defensible position for as long as the fleet stayed CANDU. It stops being defensible the moment the fleet doesn't.

That moment is Darlington. Ontario Power Generation is building a BWRX-300 there, the first commercial unit of its design anywhere in the world, and it runs on enriched fuel rather than natural uranium. That fuel is enriched at Urenco's plant in Eunice, New Mexico, with France's Orano as backup capacity. For the first time, a Canadian reactor's day-to-day operation depends on an enrichment plant Canada does not own — and did not build a domestic alternative to, in the seventy years it had to do so.

Ontario's answer so far is diversification, not production. Supply commitments run through the US, France and Britain rather than one country alone, and the federal government has convened a Nuclear Fuels Table with provinces and industry — though the longer-term question of whether Canada builds enrichment capacity itself remains under assessment. The only identified fuel-assembly fabricator in the current plan is still American.

That is the distinction Mark Winfield draws: multiple suppliers reduce the risk of disruption, but they do not eliminate the risk of leverage. A dispute doesn't need to cut off every supplier to work — it only needs to touch the one Canada is actually short on. Whether that residual exposure is something Canada is willing to live with, or the reason it finally builds the piece of the fuel cycle it always had the ore and the permission for, isn't a question this piece can answer. It's the one the next chapter will.