Comparing 10 CFR Parts 50, 52, 53, and Proposed 57 for Advanced Reactor Licensing

Regulatory Pathways. Image is AI generated, of course.

Executive summary

For a small light water reactor (LWR) sold in multi-unit packs, the practical licensing choice today is still among Part 50, Part 52, and now Part 53. Proposed Part 57 is not yet effective, and it is not simply a “non-power reactor” rule. It is a new proposed commercial framework for microreactors and other reactors with comparable risk profiles, derived from the NRC’s non-power/utilization-facility approach but combined with elements of Parts 52 and 53 to support rapid licensing and high-volume deployment. As of May 2026, Part 53 is in the Code of Federal Regulations (CFR) and available as an optional framework for commercial plants; Part 57 remains a proposed rule published on May 1, 2026.

For a small, multi-unit LWR fact pattern, the strategic headline is this: Part 50 is the most mature but least standardized path (expect to ask for exemptions); Part 52 is now the most proven path for bankable finality; Part 53 is the most analytically demanding but potentially the best long-run fit for an advanced, small multi-unit plant; proposed Part 57 is only worth tracking if the design can actually satisfy its entry criteria and if fleet deployment is central to the business model. Those Part 57 entry criteria are unusually restrictive: the applicant must show that a postulated accident bounding a broad range of Design Basis Accidents (DBAs) does not exceed 1 rem TEDE in the unrestricted area, and that the total inventory of thorium, uranium, and plutonium in the reactor or any individual reactor in the plant does not exceed 10 metric tons. Note that it has been stated in public meetings with the ACRS that the NRC is thinking the dose rate is summed from all the units. In other words, the 1-rem unrestricted-area access criterion is cumulative from all the units, since an event like an earthquake would affect all the units in a pack. However, the 10 MTHM special nuclear material criterion is applicable to each of the units. A small, multi-pack of LWRs may or may not fit that screen depending on pack cumulative dose consequence analysis and individual core inventory.

The deepest substantive difference across the Parts is how much of the licensing basis must be analytically mature at application milestones. This is not a trivial feature for applicants, because we've all seen the announcements on LinkedIn for regulatory progress, intended to capture the interest of investors, and the follow-on announcements of oversubscribed Series X raises. Part 50 still centers on a PSAR-to-FSAR progression with separate Construction Permit (CP) and Operating License (OL) hearings. Part 52 and Part 53 give applicants more ways to lock down site and design issues early, and both provide stronger issue finality once approvals are issued. Part 53 goes further than Part 50 or basic Part 52 by making Probability Risk Assessment (PRA) / systematic risk evaluations, Structures, Systems, and Components (SSC) categorization, and graded special treatment part of the core framework rather than an overlay. Proposed Part 57 goes in a different direction: it would require a very specific up-front safety case keyed to a low-consequence eligibility screen and then allow unusually flexible constructs such as joint CP/OL applications, manufacturing licenses (MLs), standard design approvals (SDAs), generic finality for a referenced CP/OL, and even a general license to construct certain portions of facilities.

On Quality Assurance (QA), the most important practical conclusion is that the basic expectation remains remarkably consistent: for Parts 50, 52, and 53, Appendix B to 10 CFR 50 remains the regulatory anchor for safety-related QA during design and construction, and NRC RG 1.28 Rev. 6 endorses ASME NQA-1-2017/2019/2022 with clarifications. Part 53 explicitly carries Appendix B forward both through safety-related (SR) special treatment and through the requirement for an operating quality assurance manual. The proposed Part 57 application section likewise requires a Quality Assurance Program (QAP) for safety-related SSCs and continues to lean on other chapter I requirements as applicable. For a small, multi-unit LWR, that means a Quality Assurance Program Description (QAPD) aligned to Appendix B + RG 1.28 Rev. 6 + NQA-1-2022 is a reusable asset regardless of whether the company ultimately chooses Part 50, 52, or 53.

My overall recommendation is straightforward. Plan of record: keep a Part 50 / Part 52-compatible QA, Final Safety Analysis Report (FSAR), and site package because those routes remain the most licensable for a small LWR in the near term. Parallel hedge: structure the design and analysis program so it can be elevated into a Part 53-quality risk-informed case if the NRC’s first reviews under Part 53 mature favorably. Watch item: monitor Part 57, but do not base financing or schedule on it unless and until the company can demonstrate that its reactor pack satisfies the proposed eligibility criteria and the final rule survives substantially as proposed.

Comparative framework and licensing pathways

Part 50 remains the NRC’s foundational framework for licensing production and utilization facilities. It is broad, mature, with lots of guidance (Regulatory Guides, Interim Staff Guidance, NUREGs), and fully available, but for new commercial plants its default structure is still the traditional construction permit (CP) followed by operating license (OL) approach, with separate review milestones and less baked-in finality of NRC staff decisions than Part 52.

Part 52 was built to solve exactly those predictability problems. It adds early site permits (ESPs), standard design certifications (DCs), combined licenses, standard design approvals, and manufacturing licenses. It also gives applicants meaningful issue finality for referenced approvals and an ITAAC-based transition to operation under a combined operating license (COL).

Part 53, now codified, is an optional, technology-inclusive, performance-based framework for commercial plants. It still offers the familiar menu of early site permits (ESP), standard design approval, standard design certification, manufacturing license, CP, OL, and COL, but it reorganizes the licensing basis around technology-inclusive safety criteria, design and analysis requirements, PRA/SREs, SSC categorization, and graded special treatment. It also splits operational personnel requirements between interaction-dependent-mitigation and self-reliant-mitigation facilities.

Proposed Part 57 is narrower and more specialized. It would create a new framework for microreactors and other reactors with comparable risk profiles, including joint CP/OL applications, manufacturing licenses, standard design approvals, generic finality for referenced CP/OL approvals, and a general-license concept for certain construction activities. It does not mirror Part 52 exactly: notably, the proposal does not create a Part 57 design certification regime comparable to Part 52/53 design certification, and it is not an ESP-centered framework.

Path	Practical preparation burden	Likely best use	Path
Part 50 CP first	Lowest initial documentation burden, because PSAR maturity can be lower than FSAR maturity; but more sequential licensing risk later.	If a developer needs an earlier regulatory foothold before the full operating basis is mature.	Part 50 CP first
Part 52 COL	Higher than Part 50 CP initial package burden, but strongest commercial-grade finality and cleaner future replication story.	If financing and partnerships require regulatory certainty and the site/design are already mature.	Part 52 COL
Part 53 COL	Similar or higher analytical burden than Part 52 COL because PRA/SRE, SSC categorization, and graded treatment are central; potentially better long-term framework.	If a developer is willing to invest now in a modern risk-informed basis that can scale.	Part 53 COL
Proposed Part 57	Could be commercially powerful later, but today it is speculative because the rule is not final and eligibility is strict.	Watchlist only, unless the design clearly fits and the rule stabilizes.	Proposed Part 57

Why the answer is different for a small, multi-unit LWR than for a non-LWR

Part 53 was created precisely because older application frameworks were “primarily developed to address” light-water designs, and the final rule’s public materials emphasize advanced-reactor flexibility, PRA, and technology inclusiveness. But for a small LWR, Parts 50 and 52 are not inherently poor fits in the way they can be for some non-LWRs. A small LWR can often use the traditional LWR guidance stack more directly, including Appendix A logic, Appendix B QA, Appendix E/§ 50.47 or § 50.160 emergency planning, and LWR-based Standard Review Plan (SRP) review structures.

That is one reason Part 53 is best understood for a small LWR as an option with upside, not an automatic default. The potential benefits are real, especially for multi-unit configurations, explicit SSC categorization, and risk-informed operational structures. But the burden shifts earlier: more must be integrated and justified in the application itself.

What the application must cover

General information and financial information

The simplest comparison is that 10 CFR 50.33 is the base general-information section, and 10 CFR 52 COL applications explicitly incorporate all of § 50.33. 10 CFR 53 has its own general-information section, § 53.1109, and proposed 10 CFR 57 has its analog in § 57.55.

Proposed Part 57 is notably more specific than Part 50 on financing thresholds in the application text. It would require construction-cost and operating-cost estimates, a financial-capacity plan, and suggests proposed license conditions if available funding at the time of application is 50 percent or less. That is more explicit front-end financing discipline than the traditional Part 50 text.

Safety analysis maturity by pathway

The most consequential application-content difference is the maturity of the safety case required at filing.

Framework	Safety report at application	Design maturity implication
Part 50 CP	PSAR under § 50.34(a).	Allows meaningful design maturation between CP and OL.
Part 50 OL	FSAR under § 50.34(b).	Final operating basis closes later.
Part 52 DC	FSAR-level design information under § 52.47.	High up-front maturity to obtain reusable design finality.
Part 52 COL	FSAR under § 52.79, plus ITAAC-linked items.	Very high maturity at filing.
Part 53 CP	PSAR under § 53.1309.	Similar staged logic to Part 50 but within the Part 53 safety architecture.
Part 53 OL	FSAR under § 53.1369.	Final design and operating basis needed before OL.
Part 53 COL	FSAR under § 53.1416.	Comparable maturity expectation to Part 52 COL, but with Part 53-specific risk and SSC logic.
Proposed Part 57 joint CP/OL	One FSAR-style technical package under § 57.60 supporting both phases.	High front-end maturity despite later readiness finding for operation.

Required technical content by subject

The most important content comparison for a small, multi-unit LWR is below.

Topic	Part 50	Part 52	Part 53	Proposed Part 57
Site and population	Required in PSAR/FSAR under § 50.34.	ESP and COL have detailed site sections; COL FSAR includes site boundaries, hazards, and population.	ESP/COL/OL sections require site info, siting assessment, and site characteristics under subparts D and H.	§ 57.60 requires site description, site factors, and surrounding population; it also has a “large designated areas” concept for flexible siting.
Source term / consequence basis	Traditional postulated fission product release and Part 100-style siting logic.	Same basic postulated-release structure for ESP/DC/COL.	Safety criteria are set through DBAs, other LBEs, and explicit risk metrics.	Entry criterion is unusually strict: bounded postulated accident ≤ 1 rem TEDE in the unrestricted area, plus reactor inventory cap.
PRA / risk analysis	Not a universal front-door requirement in § 50.34. However, NEI 18-04 has been endorsed by the NRC for backfitting a risk-informed, performance-based licensing approach.	Explicit for DC under § 52.47; referenced COLs update PRA information.	Core requirement: PRA / SREs are required by § 53.450 and application sections call for their description and results.	Requires design-specific risk analysis methods and results, but not the full Part 53 architecture.
SSC classification	Traditional safety-related (SR) / non-safety-related (NSR) structure, with some risk-informed overlays possible elsewhere. NEI 18-04 is a pathway that has allowed incorporation of SR/ NSRSS / NSR graded approach. NSRSS = non-safety-related but safety significant	Same traditional quality structure as classic 10 CFR 50, though certified designs and PRA can support more graded treatment.	Explicit SR / NSRSS / NSR graded special treatment logic in § 53.460.	Focuses on safety-related SSCs and codes-and-standards justification; less elaborate than Part 53’s SSC framework.
Quality assurance / QAPD	Explicitly required in § 50.34(a)(7) and at OL in managerial/admin controls; Appendix B governs.	Explicitly required in § 52.47 and § 52.79; Appendix B still anchors the program.	Application sections require QAP descriptions; operating licensees must maintain QA under Appendix B, and SR special treatment includes Appendix B.	§ 57.60(a)(3) requires a QAP for design/fabrication/manufacture/construction/testing of safety-related SSCs.
Cybersecurity	Part 50 OL applicants subject to § 73.55 must include physical security, training/qualification, and cyber plans.	COL applications include security and cyber plan content.	Security program includes Fitness for Duty (FFD), access authorization, cybersecurity, and information security under § 53.860.	Explicit in § 57.60(a)(8)(v): physical security, cyber (§ 73.54 or proposed § 73.110), information security, access authorization.
Emergency planning (EP) / Emergency planning zone (EPZ)	Part 50 uses Appendix E and § 50.47, but SMRs/non-LWRs/non-power can use § 50.160.	ESP/COL can lock in major features or complete plans and rely on § 50.160 or Appendix E/§ 50.47 as applicable.	§ 53.855 allows either § 50.160 or Appendix E/§ 50.47; § 53.1109 requires EPZ analysis for eligible SMRs/non-LWRs.	Emergency plan is part of § 57.60; applicant must coordinate with local/offsite organizations and include their input.
Environmental / National Environmental Policy Act (NEPA)	Environmental Report (ER) required under Part 51 when applicable.	ESP/DC/COL/ML each have distinct Part 51 hooks.	Application-specific ER requirements appear in the relevant subpart sections.	Each application must justify a categorical exclusion or submit an ER/EA/EIS; proposed Part 57 also has its own subpart K.
Fitness for duty (FFD)	Part 26 applies to operating plants, but § 50.34 does not expressly call out an FFD program description the way later frameworks do. Part 26 scope explicitly covers Part 50 operating reactors.	COL applications expressly include an FFD program description.	COL applications expressly include FFD program description and implementation; § 53.860 makes FFD an operating requirement.	§ 57.60(a)(8)(xiv) requires an FFD program description; proposal also adds alternative FFD options.

Multi-unit implications

A multi-unit “pack” is not just a design detail; it increases the application burden. Part 50 requires CP and OL applicants for multi-unit sites to also evaluate hazards to operating units from construction activities and to describe administrative controls so limiting conditions for operation are not exceeded. Part 52 carries similar expectations into COL applications, and standard design certification applications for modular reactors must analyze module configurations, interface requirements, and interactions. Part 53 and proposed Part 57 go further by directly addressing multi-unit programmatic controls, shared systems, common systems, and different operating/construction/decommissioning configurations.

For a small, multi-unit LWR, that means the pack architecture should be treated as a licensing-basis object, not just a commercial product configuration. Shared SSCs, shared support buildings, common control logic, shared radwaste systems, common emergency resources, and construction interactions should be defined early enough to show up in the licensing basis, technical specifications where needed, and the QA scope.

Review timing, hearings, issue finality, and flexibility

Timing and hearings

Part 50’s traditional structure still means two major licensing gates: a CP decision and later an OL decision. That normally means more iterative design maturation, but it also usually means more process risk because fewer issues are finally resolved early.

Part 52’s central commercial path is the COL, where the NRC later authorizes operation once the ITAAC (inspections, testing, analysis, and acceptance criteria) findings are made. The hearing at the operations transition is narrowed to whether one or more ITAAC have not been or will not be met and whether the resulting operational consequences would undermine adequate protection; the NRC is directed to decide those issues as quickly as practicable and, to the maximum possible extent, within 180 days.

Part 53 largely preserves this structured finality logic. Its COL finality section expressly protects the COL and its ITAAC-like acceptance findings, and the duration section ties the COL term to the Commission’s finding that the acceptance criteria are met under § 53.1452(g) or interim operation under § 53.1452(c). Alternatively, Part 53 can still use a two-step “construct now, authorize operation after findings” logic under § 53.1309 and § 53.1369, even though the underlying technical basis is very different from Part 52.

Proposed Part 57 intentionally departs from these models. The preamble states that the NRC would authorize operation upon a finding that reactor construction conforms to the approved design and license requirements instead of using ITAAC under Part 52, and it presents this as a way to reduce delay for first-of-a-kind and fleet deployments.

Issue finality and backfitting

Part 50 has the well-known backfit rule in § 50.109, but it does not provide the same structured reusable-approval finality that Part 52 does. That is one reason Part 52 remains attractive for investors: its referenced approvals truly narrow later litigation and NRC revisiting of resolved issues.

Part 52’s finality protections are substantial. For example, a referenced design certification holds resolved matters closed in subsequent COL, CP, OL, ML, and § 52.103 proceedings, absent narrow exceptions. The same general logic applies to ESPs, MLs, SDAs, and COLs.

Part 53 imported this philosophy directly. It has finality provisions for ESPs, SDCs, MLs, CPs, OLs, and COLs, and a separate backfitting provision for operating licenses.

Proposed Part 57 is unusual because it would stitch together Part 50-style backfitting with Part 52-style issue finality, including a bespoke concept of generic finality for a referenced CP/OL and certain generic operational programs and requirements. It would also allow changes under a § 50.59(c)-style threshold for CP/OLs that have been afforded generic finality.

Flexibilities and alternative approaches under Part 53 and proposed Part 57

The flagship flexibilities in Part 53 are not merely procedural. They are substantive: PRA/SRE-informed LBEs, defense-in-depth grounded in analysis, SSC categorization with corresponding special treatments, explicit use of consensus codes and standards with justification, and differentiated operator/staffing structures for self-reliant facilities. Those are real opportunities, but they require a mature analytical organization.

The flagship flexibilities in proposed Part 57 are different. The rule would use a low-consequence eligibility screen, explicitly recognize remote operation, remote monitoring, and autonomous performance in the application content, permit generic finality for certain CP/OL matters, add a general-license construction concept for high-volume deployment, and provide alternative FFD pathways. These are powerful ideas, but because the rule is still proposed, none should be assumed bankable yet.

Practical implications for a small, multi-unit LWR

Which documents must be mature at application versus commitment-based

The table below is a practical planning tool for a 10-person developer.

Document / program	Part 50 CP/OL and Part 53 CP/OL	Part 52 COL	Part 53 COL	Proposed Part 57 joint CP/OL	Suggested maturity for a small, multi-unit LWR
Corporate / foreign ownership / financial qualifications	Mature general information at application.	Mature at application.	Mature at application.	Mature at application and more explicit cost/funding plans.	Application-ready in all paths.
Site package and site hazards	CP can be earlier-stage than OL.	Mature at COL unless using ESP.	Mature at COL unless using ESP.	Mature at application; “designated area” concept can shift some siting flexibility into procedures.	Application-ready for any first site.
Reactor design basis / FSAR	Stageable under CP→OL.	Must be highly mature at COL.	Must be highly mature at COL.	Highly mature at application.	Near-final before filing unless choosing Part 50 CP.
Multi-unit pack architecture / shared SSC basis	Required but can mature between CP and OL.	Mature enough for FSAR and site programs at COL.	Mature enough to define shared SSCs and configuration controls.	Explicitly mature in § 57.60(a)(4).	Application-ready if the pack is real business strategy.
QAPD	Needed early in all routes.	Needed early.	Needed early.	Needed early.	Application-ready.
Implementing QA procedures	Usually commitment-based at application, but enough must exist to credibly implement the QAPD. NRC SRP 17.5 and RG 1.28 inform this expectation.	Same.	Same, but Part 53 QA / special-treatment integration should start earlier.	Same.	Core set drafted before application; full set phased.
PRA / risk methods	For 10 CFR 50: helpful but not universally front-door mandatory. For 10 CFR 53: required, but some allowance for PRA to mature between CP and OL.	Important if using DC and referenced PRA content.	Foundational and must be mature.	Requires risk-method description and results.	Start now regardless; application-ready if pursuing Part 53.
EP / EPZ basis	Mature enough for chosen route; § 50.160 can narrow burden for eligible SMRs.	Mature at COL unless locked by ESP.	Mature at COL/OL with § 50.160 option where applicable.	Mature at application, with local/offsite coordination and input.	Application-ready for first site.
Security / cybersecurity / safeguards	More operationally oriented under Part 50, but plans are required for OL applicants subject to § 73.55.	Mature at COL.	Mature at COL.	Mature at application.	Application-ready for COL / Part 53 / Part 57.
FFD	More commitment-based under Part 50 application text; Part 26 applies at operation.	Express application content.	Express application content.	Express application content, with alternative pathways.	Concept and program basis ready at application.

Resource implications for a small company

These are planning estimates, not NRC deadlines. They are my inference from the maturity the regulations require at filing and from the extent of finality each part offers.

For a small developer of about ten people, a realistic estimation of burden is:

Framework	Current status	Core scope	Best fit for a small, multi-unit LWR
Part 50	In force	Broad domestic licensing framework for production and utilization facilities; traditional CP→OL path.	Most conservative near-term route if a developer wants maximum regulatory familiarity and can tolerate more bespoke hearings and more pre-approved building blocks.
Part 52	In force	Structured new-reactor framework with ESP, DC, COL, SDA, ML, and issue finality.	Strongest route if the company wants bankable finality, referenceable approvals, and a multi-site / multi-unit commercialization story.
Part 53	In force	Optional technology-inclusive framework for commercial plants; PRA/SRE-driven, graded, and performance-based.	Attractive if a developer wants a modern integrated framework and can invest early in risk methods, SSC categorization, and operational program maturity.
Proposed Part 57	Proposed, not effective	Commercial microreactors / comparable-risk plants; joint CP/OL, ML, SDA, generic finality, general-license construction concepts.	Only a watch item unless the multi-pack can clearly meet the proposal’s entry criteria and fleet deployment economics are central.

Recommended next steps for a small, multi-unit LWR

The first decision should be a formal pathway screen. A developer should compare three realistic options: Part 50 CP/OL, Part 52 COL, and Part 53 COL or CP/OL. Proposed Part 57 should only stay in the study set if a quick eligibility memo shows that the multi-pack can plausibly satisfy both the 1 rem unrestricted-area accident criterion and the individual units satisfy the 10-metric-ton inventory cap. 

The second step should be a licensing-basis package definition. Before picking the part, a developer should freeze the following at a “licensing-basis draft” level: pack configuration and shared SSC boundaries; first-site envelope; emergency-planning concept and assumed emergency planning zone (EPZ) basis; preliminary security architecture; source-term methodology; QA scope boundary; and the change-control philosophy for modular deployment. Those are the cross-cutting items that appear repeatedly in Parts 50, 52, 53, and proposed 57.

The third step should be to treat PRA and SSC classification as enabling infrastructure, not as an optional future enhancement. Even if a developer files first under Part 50 or 52, an early PRA / systematic risk program pays off immediately in multi-unit hazard analysis, tech-spec selection, design verification prioritization, emergency planning arguments, reduced quality levels (and cost) for certain SSCs, and later migration into Part 53 if desired.

The fourth step should be to finish the QAPD and core QA procedures now. That is the one workstream that is reusable regardless of the licensing path. The regulatory baseline is stable, the NRC guidance is mature, and the market signal is favorable because recent accepted QAPDs have continued to use the Appendix B / RG 1.28 / NQA-1 structure.

The fifth step should be a pre-application memo to the NRC that frames a developer as a small, multi-unit LWR developer and asks the staff to comment on the licensing path, the role of § 50.160 for the chosen configuration, treatment of shared SSCs and cross-unit construction hazards, and whether the company should keep a Part 53 migration strategy open. That kind of focused pre-application engagement is far more valuable than broad conceptual discussion. The existing rules already show where the demanding questions will land.

Open questions and limitations

Part 57 is still a proposed rule, so any Part 57 comparison in this report should be treated as directional, not bankable licensing advice. The final rule could change materially, especially around eligibility, generic finality, and alternative FFD.

Part 53 is newly codified, and although the rule is now in the CFR, the body of licensing precedent and detailed guidance is still immature compared with Parts 50 and 52. That does not negate the framework, but it does increase execution uncertainty for first movers.

This report focuses on application and submission requirements, not on downstream inspection manuals, fee strategy, export controls, or project-specific exemption packages. Those issues can materially affect path selection, but they sit one layer below the regulatory architecture compared here.

Originally published by Sarah Gibboney on LinkedIn, May 7, 2026.