## Is AtkinsRéalis Serious About Licensing CANDU in the US?

On June 24, 2026, Montreal-based AtkinsRéalis filed a Notice of Intent with the NRC to begin pre-application licensing for the Enhanced CANDU 6 in the United States — a 700-plus MWe heavy-water reactor that runs on unenriched natural uranium and refuels online without shutting down. The filing is procedural, but the argument behind it is pointed directly at two live anxieties in Washington: the fragility of enriched-uranium supply chains and the years-long gap before any first-of-a-kind SMR reaches commercial operation.

Unlike every reactor currently operating in the US fleet, the EC6 uses heavy water as both moderator and coolant. That physics choice allows the reactor to sustain a chain reaction on natural uranium, bypassing the [uranium enrichment](https://smrintel.com/glossary/enrichment) infrastructure that the US currently leans on — infrastructure where Russian entities hold significant market share. The platform is not a paper design: AtkinsRéalis claims 34 units built and nearly 1,000 reactor-years of operating experience across Canada, South Korea, China, Romania, and Argentina. Every power reactor in Canada is a CANDU.

The company is targeting the NRC's Part 53 framework — finalized in March 2026 — which requires the agency to reach a decision on a commercial plant application within 18 months. Joe St. Julian, who leads AtkinsRéalis's nuclear business, has said the team is aiming for roughly 12 months. Whether the NRC can actually meet that timeline on its first large non-light-water reactor review in two decades is a legitimate question.

---

## Why Natural Uranium Changes the Supply Chain Calculus

The CANDU design's fuel independence is its sharpest commercial differentiator right now. The reactor uses heavy water — where hydrogen atoms are replaced by deuterium — as both moderator and coolant. Heavy water absorbs far fewer neutrons than ordinary light water, which is precisely what allows the chain reaction to sustain itself on natural uranium straight from the mine, without centrifuge cascades or enrichment services.

Every reactor in the current US fleet requires enriched fuel. The enrichment market the US relies on is dominated by Russia and a small number of other suppliers, a dependency that has drawn increasing scrutiny from Congress and the DOE since 2022. AtkinsRéalis's nuclear chief St. Julian framed the EC6 filing explicitly as cutting "reliance on foreign enrichment services." CEO Ian Edwards called it a new chapter for US civilian nuclear and a win for Canadian industry.

For utilities and hyperscalers evaluating long-term fuel cost exposure, a reactor that accepts natural uranium — and whose fuel doesn't require HALEU, LEU+, or any enrichment at all — represents a structurally different procurement model, not just a talking point.

---

## Online Refueling Is the Operational Differentiator

The second headline feature is continuous online refueling. Because CANDU bundles fuel into hundreds of short pressure tubes arranged horizontally rather than a single large pressure vessel, individual fuel channels can be opened and reloaded while the reactor is at power. A conventional light-water reactor requires a multi-week outage every 18 to 24 months to reload fuel assemblies. A CANDU does not stop generating to swap bundles.

For [baseload power](https://smrintel.com/glossary/baseload) customers — data center operators and industrial loads in particular — this has a direct impact on [capacity factor](https://smrintel.com/glossary/capacity-factor). Eliminating scheduled refueling outages means fewer planned generation gaps, which is a meaningful differentiator when the customer is underwriting a 20-year power purchase agreement on the assumption of continuous output.

The trade-off is physical scale. At 700-plus MWe, the EC6 is not sited behind a meter or on a campus parking lot. It requires an actual licensed nuclear site — which is precisely why AtkinsRéalis says its strategy focuses on bolting new units onto existing US nuclear sites, where transmission interconnection, NRC familiarity, and community acceptance are largely already established.

---

## The Regulatory History Is Not Encouraging — But Part 53 Changes the Rules

AtkinsRéalis's predecessor, Atomic Energy of Canada Limited, attempted to license a different CANDU-derivative design — the ACR-700 — starting in 2002 with Dominion Energy. That design used slightly enriched uranium and light water as coolant, making it a closer cousin to conventional US reactor designs than the EC6. The NRC still flagged it as significantly more complex to review, warning in a 2004 letter that certification could extend well past 60 months. Dominion walked away.

The EC6 filing under Part 53 is a different legal context. The NRC's risk-informed, technology-inclusive Part 53 rule — finalized March 2026 — was specifically designed to accommodate non-light-water reactor architectures and imposes a statutory decision timeline on the agency. Whether that framework meaningfully accelerates review of a large, horizontal-pressure-tube heavy-water reactor, or whether reviewers run into the same physics-translation problems that stalled the ACR-700, is the central regulatory risk for this filing.

The SMR crowd pursuing [first-of-a-kind (FOAK)](https://smrintel.com/glossary/foak) designs — including well-capitalized developers like [TerraPower](https://smrintel.com/companies/terrapower), [Kairos Power](https://smrintel.com/companies/kairos-power), and [X-energy](https://smrintel.com/companies/x-energy) — face the same NRC novelty problem, but they don't have a 2004 rejection letter in their file. AtkinsRéalis does, and the NRC institutional memory of that review will inform the agency's resource planning for this one.

---

## What This Means for the Broader US Nuclear Market

The EC6 filing introduces a genuine alternative to the current US advanced nuclear pipeline that is worth taking seriously, not because it is certain to succeed, but because the asset it is offering is qualitatively different from what every other applicant is selling.

The rest of the market is offering designs that are years from commercial operation, require novel fuel types (HALEU for most fast-spectrum and high-temperature concepts), and are seeking [NRC design certification](https://smrintel.com/glossary/design-certification) for the first time. AtkinsRéalis is offering a design with nearly 1,000 reactor-years of documented operating history, natural uranium fuel, and a track record of keeping an entire G7 country's grid supplied for half a century.

The site-attachment strategy — targeting existing licensed nuclear facilities rather than greenfield development — also fits the pattern of every successful US nuclear announcement of the past two years: Three Mile Island Unit 1 returning for Microsoft, Palisades being pulled from decommissioning by [Holtec International](https://smrintel.com/companies/holtec-international), [Constellation Energy](https://smrintel.com/companies/constellation-energy) pursuing fleet-wide license extensions. The cheapest nuclear site is one that already exists.

The market signal for uranium is also worth noting. A large natural-uranium reactor deployed at scale in the US would meaningfully increase demand for unenriched feed — a direct positive for producers like [Cameco Corporation](https://smrintel.com/companies/cameco) and [NexGen Energy](https://smrintel.com/companies/nexgen-energy) that supply natural uranium concentrate, while bypassing the enrichment step entirely.

---

## Key Takeaways

- AtkinsRéalis filed a Notice of Intent with the NRC on **June 24, 2026** to begin pre-application licensing for the Enhanced CANDU 6 in the United States.
- The EC6 is a **700-plus MWe** heavy-water reactor that runs on **natural, unenriched uranium** and refuels continuously without shutting down.
- The filing invokes the **Part 53** framework, finalized March 2026, which requires the NRC to decide on a commercial plant within 18 months; AtkinsRéalis is targeting roughly 12.
- AtkinsRéalis claims **34 units built and nearly 1,000 reactor-years** of operating experience across five countries — a documented operating record no SMR applicant can match.
- A predecessor CANDU licensing attempt (ACR-700, beginning 2002) stalled at the NRC; the NRC explicitly warned in 2004 that review would exceed 60 months. Dominion Energy withdrew.
- The company's strategy targets **existing US nuclear sites** rather than greenfield locations, reducing licensing and community-acceptance risk.
- The fuel independence angle directly addresses the US dependence on foreign enrichment services, a priority that has bipartisan traction in Washington.

---

## Frequently Asked Questions

**What is the Enhanced CANDU 6 and how big is it?**
The Enhanced CANDU 6 is a heavy-water reactor design stewarded by AtkinsRéalis, the Montreal-based engineering firm. According to the company's NRC filing materials, it produces more than 700 MWe. It uses heavy water as both moderator and coolant, which allows it to run on natural, unenriched uranium rather than the enriched fuel required by every currently operating US reactor.

**Why doesn't the CANDU need enriched uranium?**
CANDU reactors use heavy water — water in which ordinary hydrogen is replaced by deuterium — as the moderator. Heavy water absorbs far fewer neutrons than ordinary light water, which means the reactor can sustain a fission chain reaction using natural uranium at its as-mined concentration. Light-water reactors require uranium enriched to higher concentrations of U-235 because ordinary water absorbs too many neutrons for the chain reaction to self-sustain on natural uranium.

**What is Part 53 and why does it matter for this filing?**
Part 53 is a new NRC licensing framework finalized in March 2026. It is designed to be risk-informed and technology-inclusive, meaning it does not presuppose light-water reactor architecture the way the existing Part 50 and Part 52 rules largely do. Critically, Part 53 imposes a statutory deadline on the NRC to decide on a commercial plant application. AtkinsRéalis is filing under Part 53 specifically to invoke that timeline discipline.

**Has a CANDU ever been licensed in the United States before?**
No. Atomic Energy of Canada Limited attempted to license the ACR-700, a different CANDU-derivative, beginning in 2002 with Dominion Energy. The NRC warned in 2004 that the review complexity could push certification well past 60 months. Dominion withdrew. The EC6 filing under Part 53 represents the first CANDU pre-application engagement with the NRC since that effort ended.

**How does online refueling affect capacity factor for data center customers?**
A conventional light-water reactor requires a multi-week refueling outage every 18 to 24 months, during which it generates no power. A CANDU refuels continuously while operating, because individual horizontal pressure tube fuel channels can be opened and reloaded at power. For a hyperscaler or industrial load anchored to a long-term PPA, eliminating planned outage windows directly raises effective capacity factor and reduces the need for backup generation contracts.

**Who is AtkinsRéalis and what is their connection to CANDU?**
AtkinsRéalis is a Montreal-based engineering and project management firm that holds the intellectual property and engineering stewardship for the CANDU reactor technology originally developed by Atomic Energy of Canada Limited. The company's nuclear business is led by Joe St. Julian, who is the named executive in communications around the NRC filing.