The Economics of Candu Reactor Refurbishment Versus New Construction

The global nuclear energy sector faces a decisive moment as dozens of reactors built during the 1970s and 1980s approach the end of their original operating licences. For owners of CANDU (Canada Deuterium Uranium) reactors, the choice between extending the life of an existing plant through a major refurbishment and constructing a brand-new reactor is a multi-billion-dollar strategic calculation. The answer has profound implications for electricity system reliability, carbon abatement trajectories, and the long-term competitiveness of heavy water reactor technology. This analysis unpacks the economics behind that decision, drawing on real projects, capital cost data, and the market and regulatory frameworks that shape investment outcomes.

What Makes CANDU Refurbishment a Distinct Economic Proposition

CANDU reactors, originally developed by Atomic Energy of Canada Limited, are pressurized heavy water reactors that use natural uranium as fuel and heavy water as both moderator and coolant. Their horizontal fuel channel design – a set of approximately 380 to 480 pressure tubes that contain the fuel bundles – is fundamentally different from the large pressure vessel designs of light water reactors. This architecture is the key to the refurbishment economics: the pressure tubes and their enclosing calandria tubes are subject to irradiation, creep, and dimensional change over decades of operation, and they were never intended to last for the full 60‑plus years of feasible plant life. Everything else – the concrete containment, the steam generators, the turbine hall, and the balance of plant – can, with proper maintenance, function for much longer.

Refurbishment therefore focuses on a mid‑life “heart transplant” rather than a full rebuild. The core activity is removing and replacing all pressure tubes and calandria tubes, a process that requires shutting down the reactor for a period of three to four years per unit. During this outage, utilities also replace feeder pipes, inspect and often replace steam generators, install modern digital control and safety instrumentation, and upgrade the plant’s emergency heat removal systems to meet current safety standards. Because the heavy water inventory is largely retained and the existing containment structure and site infrastructure are reused, the material and construction volume is far smaller than that of a greenfield project.

That reuse of infrastructure is the first major economic differentiator. A refurbishment project leverages the original investment in site preparation, transmission connection, cooling water intake and outfall, and the immense mass of concrete and steel that makes up the reactor building. It also preserves the plant’s existing labour agreements, security clearances, and the tacit knowledge of the operating workforce. From a regulatory standpoint, the licensing pathway for refurbishment is considerably shorter than that for a new build, even though it still requires robust environmental assessment and safety case approval from the Canadian Nuclear Safety Commission (CNSC).

Current Refurbishment Cost Benchmarks

The most comprehensive data on CANDU refurbishment costs come from Ontario, where the provincial government and its agencies have undertaken the simultaneous life extension of ten reactors at the Darlington and Bruce nuclear generating stations. Ontario Power Generation (OPG) completed the refurbishment of Darlington’s Unit 2 in 2020 at a cost of approximately CAD 5.1 billion for that single 878‑megawatt (electric) unit. The overall Darlington refurbishment program, which covers all four units in sequence, is budgeted at CAD 12.8 billion (2015 dollars). When adjusted for inflation and scope, this equates to roughly CAD 3,700 to CAD 4,200 per installed kilowatt of capacity.

At the Bruce Power site – operated as a public‑private partnership – the Major Component Replacement (MCR) project for six units carries an estimated total cost of CAD 13 billion (in 2020 dollars). While these figures are substantial, they are markedly lower than the overnight capital cost of a new large nuclear reactor, which the most recent international scaling studies place in a range of USD 5,500 to USD 9,000 per kilowatt for advanced light water and heavy water designs in Western markets. A new build CANDU 6, if pursued today, would likely fall toward the upper end of that range due to re‑establishment of the supply chain and regulatory learning costs. A 2023 study by the OECD Nuclear Energy Agency placed the median overnight cost for a new nuclear plant in a first-of-a-kind environment at over USD 7,700 per kilowatt, reinforcing the cost advantage of refurbishment.

Importantly, the cost of refurbishment has proven to be far more predictable than that of new construction. The Darlington Unit 2 project was delivered on schedule and within its budget envelope, earning confidence from bond rating agencies and the provincial treasury. The predictability stems partly from the fact that the majority of the project cost is for manufacturing and installing components at a site where the geological, hydrological, and logistical characteristics are already fully characterized. Overruns in large new builds have historically been driven by unanticipated ground conditions, supply chain failures, or design immaturity – three risks that are greatly reduced in a refurbishment.

Economic Advantages of Refurbishment: Beyond Capital Savings

Focusing only on the upfront cheque misses the full economic picture. Refurbishment delivers electricity into the grid sooner, carries lower financing costs, and defers the decommissioning liability by thirty years or more. Here are the key advantages when compared with a new CANDU construction project.

  • Lower overnight capital outlay. Even a high‑end refurbishment at around USD 4,500 per kilowatt represents roughly half the unit cost of a new reactor built under current market conditions. This reduces the equity and debt that must be raised and the financial risk passed through to ratepayers or taxpayers.
  • Faster return to service. A unit undergoing its single mid‑life refurbishment outage is offline for 36 to 48 months. A new nuclear plant typically takes 7 to 10 years from first concrete to grid connection, not including the 3‑5 years of planning, licensing, and site preparation that precede it. Earlier cash flows improve the net present value dramatically.
  • Re‑use of regulatory approvals and site envelope. The existing environmental assessment, site licence, and operating licence can be amended rather than created from scratch. Community acceptance is generally higher because local populations have lived alongside the station for decades and understand its socio‑economic contribution.
  • Supply chain maturity. The CANDU pressure tube and feeder replacement supply chain – involving specialized Canadian manufacturers such as BWXT Canada, SNC‑Lavalin (now AtkinsRéalis), and various precision machining firms – has been revitalized and optimized through the Darlington and Bruce programs. The learning curve has already been paid for.
  • Deferral of decommissioning costs. Every additional year of operation pushes the final shutdown date farther into the future, reducing the present value of the decommissioning trust liability. For a four‑unit station, a 30‑year life extension can shift the payment window by a generation, freeing up capital for other energy transition investments.

Levelised Cost Comparisons from Real Projects

When all these factors are combined into a levelised cost of electricity (LCOE) calculation, refurbishment comes out highly competitive. Studies by the Canadian Nuclear Association and independent academic groups such as the Ontario Centre for Engineering and Public Policy have estimated that the LCOE for refurbished CANDU power can range from CAD 40 to CAD 65 per megawatt‑hour (2023 dollars), assuming typical discount rates for regulated utilities. This puts it in the same band as new combined‑cycle natural gas without carbon pricing and significantly below new nuclear greenfield projects, which often exceed CAD 100 per megawatt‑hour in comparable economic analyses. A 2024 report by the International Energy Agency noted that extending the life of existing nuclear plants is one of the most cost-effective ways to maintain low-carbon flexibility in electricity systems.

The Darlington and Bruce programs are also instructive on the balance of trade and employment effects. Refurbishment of ten reactors is expected to sustain approximately 50,000 person‑years of direct and indirect employment, much of it in skilled manufacturing, engineering, and construction trades. The non‑exportable nature of these jobs means that the economic multiplier stays largely within the provincial and national economy, strengthening political support for the expenditure. Furthermore, the work is largely shielded from global supply chain disruptions because the majority of components are fabricated domestically.

How Refurbishment Affects Ratepayers and Utility Balance Sheets

From a utility perspective, refurbishment offers a predictable cost recovery model. In Ontario, OPG uses a “deferred project cost” regulatory mechanism that spreads the refurbishment investment over 30 years of post‑refurbishment operation. This avoids sudden spikes in electricity rates and aligns costs with the benefits received by consumers. The Bruce Power MCR program is financed through power purchase agreements with the Ontario IESO, providing revenue certainty that allows private lenders to offer competitive interest rates. For ratepayers, the impact is modest: the Ontario Financial Accountability Office estimated that the Darlington refurbishment adds less than CAD 3 per month to an average residential bill, far below the cost of alternative new generation.

In contrast, financing a new CANDU build would require either a government backstop or a contract for difference that prices in construction risk. The higher upfront capital and longer interest‑during‑construction period would inevitably raise the cost of capital, which for a new build can add 30% to 50% to the overnight cost. Ratepayers in jurisdictions without strong state ownership would face significant price risk during construction delays. Refurbishment’s proven track record of schedule and budget discipline makes it a safer bet for balance sheets.

The Case for New CANDU Construction

While refurbishment is economically compelling for existing sites, it is not a universal solution. New construction may be the only viable path in circumstances where there is no reactor to refurbish, where rapid capacity expansion is needed, or where a country wishes to launch a nuclear power program for the first time. A greenfield CANDU reactor can also take advantage of evolutionary design improvements that simply cannot be retro‑fitted into an old containment building.

The Enhanced CANDU 6 (EC6) design, for example, incorporates passive features for post‑accident cooling, a more robust containment spray system, and provisions for a longer 60‑year design life without a major mid‑life refurbishment. These advancements could eventually narrow the operational cost gap by reducing the heavy water losses and outage frequency. However, no EC6 has yet been built, and the first‑of‑a‑kind premium would be severe. Canada’s export push for the EC6 in markets such as Argentina, Romania, and China has yielded only limited uptake, with the most advanced discussions – at Romania’s Cernavodă units 3 and 4 – still facing financing and construction partnership hurdles.

New CANDU construction would also carry a different set of economic risks:

  • Higher overnight capital cost. Estimates for a new twin‑unit CANDU 6 station in a Western regulatory environment start at CAD 10‑15 billion for the first pair of units, translating to CAD 8,000‑12,000 per kilowatt until serial‑build benefits are realised.
  • Longer revenue‑deferred period. Capital is tied up for a decade or more with no cash return, exposing investors to interest rate volatility, electricity price risk, and regulatory changes.
  • First‑of‑a‑kind engineering. Even using a proven reference design, re‑establishing the detailed engineering and licensing for a new site involves tens of thousands of hours of professional work that must be paid for upfront.
  • Fuel cycle infrastructure. While CANDU reactors are remarkably flexible on fuel (natural uranium, slightly enriched uranium, and even thorium blends), the market remains small compared with light water reactor fuel, and the heavy water inventory alone costs hundreds of millions of dollars per reactor.

Despite these hurdles, new builds may be justified in specific scenarios. Expanding an existing CANDU fleet on an already‑licensed site – such as the proposed Darlington New Nuclear Project (now pivoting to small modular reactors, but originally conceiving a full‑scale CANDU unit) – would mitigate some site‑specific risks. Export projects backed by government‑to‑government financing and sovereign guarantees could overcome the financing barrier, as the earlier CANDU 6 exports to Romania, South Korea, and China showed in the 1980s and 1990s. In those cases, the cost of heavy water and the need for long-term fuel supply agreements were manageable within the bilateral framework.

Comparative Analysis: Refurbishment vs. New Build in Specific Scenarios

To illustrate the trade-offs, consider two hypothetical projects. Scenario A: A single-unit CANDU station built in 1990, entering its mid-life refurbishment window at age 35. The plant has performed well, has a strong licensing basis, and operates in a deregulated market with carbon pricing. Refurbishment at USD 4,000/kW with a 4-year outage will produce power for an additional 30 years at LCOE around CAD 55/MWh. Scenario B: A greenfield twin-unit CANDU 6 design in the same market, requiring 10 years to build, at USD 8,000/kW. The LCOE, including financing and carbon costs savings, would be above CAD 110/MWh for the first decade, declining only after capital is fully amortized. The refurbishment unit operates at a significant cost advantage for the entire period until 2055, while the new build would need a power price guarantee to attract investment.

In jurisdictions without existing CANDU expertise, the calculus shifts. A country like Saudi Arabia or Poland considering a heavy water program would face high first-of-a-kind costs, supply chain development, heavy water leasing, and workforce training. For them, new CANDU construction would be compared with other technologies like AP1000, HPR1000, or SMRs, and may be viable only if heavy water reactor characteristics (e.g., use of natural uranium, online refuelling) align with energy security goals.

Factors That Tip the Balance in Refurbishment Decisions

For asset owners and government planners, the choice between refurbishing or replacing is rarely a pure engineering‑economic optimisation. Several external factors heavily influence the final decision.

Carbon pricing and climate policy. In Canada, the federal carbon price is scheduled to rise to CAD 170 per tonne of CO₂ by 2030. This dramatically improves the economic competitiveness of existing nuclear capacity, which emits essentially no greenhouse gases. Refurbishment, by keeping zero‑emission baseload on the grid, directly avoids the construction of new gas‑fired peaking plants and their associated carbon costs. Modelling by the Canadian Climate Institute indicates that without the life extension of Ontario’s CANDU fleet, the province would face an emissions gap of tens of millions of tonnes per year, requiring expensive offsetting measures.

Electricity market design. In a deregulated or hybrid market, the revenue certainty for a new nuclear plant is difficult to achieve. Refurbished units, however, sit within existing power purchase agreements, regulated rate‑base recovery, or contract‑for‑difference mechanisms that have already been established. Ontario’s Long‑Term Energy Plan explicitly recognised that refurbishment is the lowest‑cost option for maintaining clean baseload, and the government structured a unique “deferred project cost” model to allow OPG to recover its refurbishment investment through regulated rates over 30 years.

Supply chain readiness and labour force. The resurgence of CANDU refurbishment work in Ontario has revitalised a domestic nuclear supply chain that had been dormant since the completion of Ontario’s last new reactors in the early 1990s. Companies that had lost skilled toolmakers and engineers began rehiring and training a new generation of workers. This ecosystem, once rebuilt, makes subsequent refurbishments cheaper and more efficient. A decision to build new large CANDU reactors would, in contrast, require a scale‑up of a different magnitude, potentially drawing resources away from the highly successful refurbishment programs and creating a costly competency gap.

Public and indigenous ownership. Refurbishment of existing stations on long‑occupied indigenous lands often carries stronger community support because the project is framed as stewardship and continuity rather than new encroachment. The Bruce Power MCR initiative, for example, has included significant investment in local infrastructure and partnerships with the Saugeen Ojibway Nation. Such social licence is easier to maintain for a life‑extension project than for a new build, which involves fresh site‑specific land rights and environmental baseline studies.

Policy and Regulatory Implications for Refurbishment Economics

Government policy plays a decisive role. In Canada, the CNSC has streamlined the licence renewal process for refurbishment projects, requiring a project-specific safety case but allowing reuse of existing site documentation. The World Nuclear Association notes that such regulatory efficiency reduces the licensing cost of refurbishment by 40% compared to a new build. Internationally, Romania’s Nuclearelectrica has benefited from EU state aid guidelines that classify nuclear life extension as a clean energy investment, enabling access to preferential loans from the European Investment Bank.

The fiscal treatment of decommissioning trusts also matters. In Ontario, decommissioning funds are held in segregated trusts with conservative return assumptions. Refurbishment defers the timing of withdrawals, allowing the trust to grow through investment returns and reduce the net present value of the ultimate liability. New builds, by contrast, require immediate contributions to a new trust from day one, imposing a cash flow burden that adds to the front‑ended cost structure.

Real‑World Case Studies: Ontario’s Refurbishment Renaissance

The experience of the Darlington and Bruce refurbishment programs offers a live economic laboratory. OPG’s Darlington Refurbishment was the first full‑scale CANDU mid‑life refurbishment in Canada since the Point Lepreau and Wolsong 1 projects of the 2000s. The Unit 2 overhaul achieved commercial operation ahead of its revised schedule and at a final cost that matched early‑stage estimates – a rarity in heavy industrial projects. This performance strengthened market confidence and directly influenced the government’s 2022 decision to proceed with the refurbishment of the remaining three Darlington units and to explore further life extensions at the Pickering station.

Bruce Power’s Major Component Replacement project is the largest private sector infrastructure investment in Canadian history. Executed by Bruce Power, TC Energy, and the Power Workers’ Union, the MCR programme will extend the life of six units by an estimated 30 to 35 years. The financing was arranged through a consortium of Canadian and international banks, underpinned by power purchase agreements with the Ontario Independent Electricity System Operator. The economic structure demonstrates that refurbishment can attract private capital even for multi‑billion‑dollar outlays, provided the regulatory and revenue framework is durable.

Internationally, the refurbishment approach is also being applied to CANDU reactors elsewhere. Argentina’s Embalse reactor completed a life extension project in 2019, and Romania’s Nuclearelectrica is evaluating similar work for Cernavodă Unit 1. These examples reinforce that the global CANDU fleet – approximately 30 reactors in seven countries – can remain in service for the better part of this century, deferring the need for massive new build investments while maintaining zero‑carbon capacity. The Korean CANDU units at Wolsong have also undergone life‑extension programs, demonstrating that the refurbishment model is replicable across different regulatory and economic environments.

Lessons from the Chinese Qinshan CANDU Experience

China’s two CANDU 6 reactors at Qinshan Phase III, built in the early 2000s, offer a contrasting perspective. These units were constructed as new builds in a rapidly growing economy with strong government support. Their construction time of about 6 years per unit was competitive for the era, and they have operated at high capacity factors. However, their lifecycle economics were shaped by China’s low cost of capital and state‑owned banking system. For most other countries, the Chinese approach is not replicable. The Qinshan experience underscores that new CANDU construction only makes sense when capital is patient and the builder has deep experience with the technology – conditions that currently exist only in a few markets.

When New Build Becomes the Pragmatic Choice

Refurbishment cannot go on indefinitely. Eventually, the containment building itself, the massive reinforced concrete structure that encloses the reactor, will reach a point where continued licence renewal becomes technically and economically infeasible. At that stage, a decision on replacement capacity must be taken. For jurisdictions that expect sharply rising electricity demand due to electrification of transport, heating, and industrial processes, adding new nuclear capacity may be necessary to meet clean energy targets.

Some of the economic impetus for new CANDU builds could come from the growing interest in small modular reactors (SMRs). While SMRs are not the same as new large CANDU stations, the development of a modern, advanced CANDU‑derived SMR design by AtkinsRéalis could bridge the gap between refurbishment and next‑generation heavy water technology. The intellectual property and the heavy water supply chain preserved through ongoing refurbishment work create a platform that lowers the barrier to future new builds. In this sense, every refurbishment is also an investment in maintaining the capability to eventually construct new CANDU plants.

The International Energy Agency has warned that without life extensions for the existing nuclear fleet, the world will find it much harder to meet net‑zero emissions targets. The 2023 update to its Net Zero Roadmap highlights that closing nuclear plants before end of life would increase cumulative CO₂ emissions by over 4 billion tonnes globally, primarily through increased reliance on natural gas. Refurbishment of CANDU reactors, therefore, holds an economic value that transcends national boundaries, contributing to a lower‑cost global decarbonisation pathway. Additionally, a 2024 study by the OECD NEA indicated that extending the life of an existing nuclear plant can avoid up to 100 times more CO₂ per dollar spent than building a new renewable farm or gas plant, making it one of the most efficient climate investments available.

Technological Pathways: From Refurbishment to Advanced CANDU Designs

The heavy water technology ecosystem developed through refurbishment could pivot to advanced CANDU designs like the CANDU SMR or the EC6. AtkinsRéalis is advancing a 300 MWe heavy water SMR that leverages the same fuel channel architecture and pressure tube manufacturing supply chain that refurbishment programs have rejuvenated. The economic viability of such an SMR depends on factory fabrication and modular assembly, which could reduce onsite construction costs by 30% compared to a large CANDU. While no orders exist yet, the technology readiness is high. If such designs reach commercial deployment, they could offer a path for replacing retired CANDU capacity on existing sites with a more compact, lower-cost unit, blurring the line between refurbishment and new build.

Synthesis and Strategic Considerations

The economics of CANDU reactor refurbishment versus new construction can be summarised as a clear preference for the former whenever such a choice exists. Refurbishment offers a lower capital requirement, higher cost predictability, faster revenue realisation, and a significantly reduced risk profile compared with greenfield heavy water reactor projects. It leverages fully‑depreciated site infrastructure, stabilised regulatory frameworks, and a workforce intimately familiar with the plant.

New CANDU construction, while possessing engineering merit and the potential for improved safety and operational efficiency, faces headwinds in the form of enormous upfront capital needs, long construction timelines, and stiff competition from alternative technologies including advanced light water reactors and SMRs. In the near to medium term, the industry’s finite technical and financial resources are rationally allocated to keeping the existing fleet running, and the evidence from Canada’s refurbishment program indicates that this strategy is delivering value.

Decisions will vary by jurisdiction. A country with no existing nuclear capacity would obviously find refurbishment irrelevant; for them, the relevant comparison is between a new CANDU unit and other baseload options. But for the nations that already host CANDU stations – Canada, Romania, Argentina, South Korea, China, India, and Pakistan – the economic logic points strongly toward maximising the life extension of those assets before committing to a new build cycle. The clean energy transition demands pragmatic solutions that balance cost, reliability, and decarbonisation speed, and CANDU refurbishment checks all those boxes.