Regulatory pathway

Licensing: Part 57 primary, Part 53 backup.

HELIX is being developed against a defined regulatory path, not a hope. Our primary target is the NRC's proposed 10 CFR Part 57 microreactor framework, with the final Part 53 framework held as the backup pathway, and we scoped the topical-report sequence to a verified compliance register and are structuring a DOE-authorized test unit so its data credits into the commercial case. Below is the honest state of that path: what is targeted, what is planned, and what is owed before any hardware is built.

Licensing is where advanced-reactor programs most often overstate their position, so we are going to be precise about ours. A pre-application developer has not been granted anything by the NRC; it has chosen a pathway and is preparing for the structured engagement that precedes a formal application, an engagement we have not yet begun. That is exactly where HELIX is, and our primary pathway is itself still a proposed rule, which we say plainly rather than bury. What makes the position credible is not a claim of approval, it is that every step is named, sequenced, and tied to a verifiable compliance register rather than to a marketing timeline.

What is Part 57, and why is it our primary target?

Part 57 is the NRC's proposed licensing framework written specifically for microreactors, proposed on May 1, 2026, with a final rule expected on November 23, 2026. Two things about it matter for HELIX. First, it provides fleet approvals of identical reactors, which is exactly the shape of a number-up site built from identical factory-sealed modules: approve the machine once, then deploy it as a fleet rather than relicensing each unit as a bespoke plant. Second, it is aimed at simple machines with simple safety systems, and a pumpless, walk-away design whose shutdown cooling is carried by natural-draft air and radiation is built to fit that description. We say clearly what this is not: Part 57 is proposed, not final, we hold no approval under it, and no licensing application is underway.

Where does Part 53 fit now?

Part 53 is the NRC's risk-informed, technology-inclusive framework for commercial nuclear plants, made final in 2026, and it remains our backup pathway. It matters for a design like HELIX because rather than forcing a sodium-cooled microreactor to fit rules written around large light-water plants, it lets a developer make a safety case on the actual physics and risk profile of the design. The work we scoped against it, the compliance register, the topical-report sequence, and the analysis structure, transfers rather than being discarded, so holding it as the backup costs the program nothing and keeps a final, in-force framework available if the Part 57 rule shifts or slips.

How does the application actually get built?

An application is not a single document; it is a sequence. The Licensing Project Plan lays out the topical-report sequence and ties each report to a specific regulatory requirement in a compliance register we maintain and verify. That sequence includes selecting the licensing-basis events the design must withstand, classifying every structure, system, and component by its safety significance, developing a mechanistic source term that describes what could actually be released and under what conditions, and qualifying the pressure-boundary materials under ASME Section III Division 5 for the high-temperature regime, which is guided by Regulatory Guide 1.87. We have not yet fixed the vessel material or the vessel temperature, so that code case is named here as owed rather than claimed as done. Each of these is a discrete, checkable deliverable, and each is scoped against the register rather than asserted.

What is the DOE test-unit track, and why does it matter?

There is a faster, more rigorous way to generate the data an application needs than analysis alone: build a test article under Department of Energy authorization and collect quality data from it from day one. Structuring a DOE-authorized test unit, with data gathered under an NQA-1 quality program from the first hour of operation, feeds real measured behavior into the eventual NRC application. This is consistent with the NRC's own proposed pathway to credit DOE-authorized designs, and it converts what would otherwise be a purely paper submission into one anchored by test data. For a first-of-a-kind reactor, that difference is the difference between a credible application and an optimistic one.

What about manufacturing? It is a second license, and it comes later.

Deploying a fleet needs two separate authorizations, and they are often conflated. The first is the reactor licensing path above. The second is the right to build units in a factory, and it runs on its own track: an Appendix B quality-assurance program, then a Part 21 defect-reporting program, then a contract with an Authorized Inspection Agency, then an ASME Section III Division 5 Certificate of Authorization, and only then a manufacturing license under 10 CFR Part 52 Subpart F.

Two facts about that track are worth stating plainly, because they set the sequence and most vendor sites leave them out. No manufacturing license has ever been applied for under Subpart F. And a reactor built under one may only be transported to and installed at a site that already holds a construction permit or a combined license, which means the manufacturing license follows customers rather than creating them. ASME also surveys a physical shop with an implemented quality program, and its certificates last three years, so obtaining one before there is a factory to certify would simply burn it.

The useful consequence is that the two tracks share their first step. The same Appendix B quality program that gates every topical report on the reactor side also gates the entire manufacturing side. One workstream unlocks both, which is why it is the item we treat as the bottleneck rather than the paperwork.

What is honestly still owed?

Two gates sit ahead of any hardware commitment, and we label them as owed rather than dressing them up. The first is a stood-up NQA-1 quality-assurance program and a preliminary safety analysis report. Until that program exists, no analysis can be credited at all, which is exactly why we describe every physics result on this site as unqualified screening: it is produced outside a QA program and is not carried forward as credited work. The second is independent physics validation, confirmation with independent codes and ultimately test data, together with validated offtake demand. We do not commit hardware on the strength of our own screening and our own optimism; both of those external confirmations come first.

Being pre-application is where the frontier is

Every credible advanced-reactor developer is either in pre-application or early licensing. That is the current phase of the entire cohort: 10 CFR Part 53 itself only became final in 2026, and the microreactor-specific Part 57 is still a proposed rule. We treat the honest labels, design targets, unqualified screening, and pre-QAPD status, not as caveats to apologize for but as the milestones of a path we are actively executing. The discipline of naming exactly where we are is what makes everything else on this site credible, and it is the same discipline that produced a reactor designed to be checked rather than merely trusted.

See the validation program that feeds the safety case →

Read why verification is the whole strategy →