NRC Part 57 and Autonomous Operation, Explained

Where the proposed Part 57 rule stands today

Part 57 is at the proposed-rule stage. The NRC published the draft in the Federal Register (91 FR 23628), invited public comment, and closed that comment period on June 15, 2026 ^[1]. A proposed rule is a draft the regulator puts forward for review. It is not in effect, developers cannot be licensed under it yet, and the NRC may revise or withdraw provisions before issuing any final rule.

This matters for how you read every other claim about the rule. Vendor material and even some trade coverage describe Part 57 in the present tense, as though its provisions are settled. They are not. The proposed text signals the direction the NRC is exploring for the microreactor class, and it reflects the modernized, risk-informed approach the agency has been building across its advanced-reactor work ^[3]. What it does not do is grant anyone permission to operate under its terms. Until a final rule is published, the accurate way to speak about Part 57 is in the conditional: what it would do, what it contemplates, what it proposes. This post uses that framing throughout, on purpose.

How the proposed rule frames autonomous and remote operation

The proposed rule defines autonomous operation and contemplates remote operation with reduced on-site staffing for the microreactor class ^[1]. In plain terms, it sets out language for reactors that are not assumed to have a full crew standing at the controls. It supplies definitions and a licensing pathway the NRC is proposing for that operating model, rather than requiring the legacy staffing assumptions written for large light-water plants.

Read carefully, the draft is doing something narrower than the headlines suggest. Defining a term is not the same as declaring that unattended reactors are approved reality. The proposed rule provides a regulatory vocabulary and a framework the NRC would use to evaluate such designs; it does not certify that any particular reactor can run that way safely. That determination would still depend on a specific design, its analysis, its testing, and NRC review. Reduced on-site staffing also is not the same as no oversight: safety-significant decisions in these operating concepts keep a human in the loop, and no facility today is licensed to operate unattended. The proposed rule frames the model. It does not settle whether any given machine has earned it.

How Part 57 compares with the final Part 53 rule

Part 57 and Part 53 are two different NRC rules at two different stages. Part 57 is proposed and aimed at microreactors; Part 53 is final and is a broad, risk-informed, technology-inclusive framework for advanced reactors, published in March 2026 ^[2]. They are related pieces of the NRC's licensing modernization, but only Part 53 is in effect, and the two differ in scope and in how they treat the staffing and autonomy question.

The table below compares them on the points buyers ask about most. Read the status row first: it is the one that changes how you interpret everything else.

The short version: Part 53 is the finalized, general advanced-reactor framework, while Part 57 is the proposed, microreactor-specific rule that introduces the autonomous-operation language. If you see the two names used interchangeably, that is a signal to slow down and check which rule, and which stage, is actually being discussed.

The trust surface the proposed rule opens

By contemplating reduced on-site staffing, the proposed rule surfaces a question it does not itself answer: when fewer people are present, who independently confirms that the reactor is doing what its operator reports. The rule sets the regulatory frame for the operating model. It does not supply the technical layer that lets a regulator, insurer, lender, or grid operator check the reactor's reported state without taking the operator's word.

That gap is the reason this rule matters beyond the licensing bar. When a reactor is designed to run with a reduced crew, more of the day-to-day safety story rests on what the reactor reports about itself, which raises the obvious follow-up: can an outside party trust a system's account of its own condition. Our view is that this is the most consequential thing the proposed rule exposes, more than the autonomy definition itself. The answer is independent verification: continuous confirmation of reactor state and commands by a party structurally separate from the operator, recorded so it can be checked later. We explore that model in [How to Verify an Autonomous Microreactor Is Operating Safely] and the difference between a self-report and an outside check in [Self-Attestation vs Independent Verification for Autonomous Reactors]. Part 57 frames the operating model. The trust layer is a separate problem.