The NRC has just 18 months to execute an unprecedented, wide-ranging regulatory overhaul mandated by the May 23 executive order. This post breaks down the logistical mountain ahead, from tight deadlines to staffing constraints, and explores what shortcuts, strategies, and tech (including AI) might help the agency pull it off.

On Tuesday I shared a fast analysis of the May 23 NRC-reform executive order and promised several follow-ups. Here’s the first, and the one I’ve been thinking about the most—the logistics of the rulemaking. This is quite the multi-course meal to be serving in 18 months.

Gather the Ingredients

Future articles will give background and forecasts on the substance of the rule—logistics is enough for one day—but as a recap, the EO asks for 10 things:

1. Enforce Fixed Review Deadlines

2. Use Science-Based Radiation Limits

3. Update Environmental Review Rules

4. Expedite Proven Reactor Designs

5. Simplify Licensing for Small Reactors

6. Limit Mid-Construction Design Changes

7. Revise Oversight & Security Rules

8. Use Realistic Risk Thresholds

9. Extend License Renewal Terms

10. Streamline Public Hearings

Plan the Courses

The order directs the NRC to:

* undertake a review and wholesale revision of its regulations and guidance documents

* issue notice(s) of proposed rulemaking effecting this revision within 9 months of the date of this order (February 23, 2026)

* issue final rules and guidance to conclude this revision process within 18 months of the date of this order (November 23, 2026)

This schedule likely has NRC management petrified. It’s a timeline unheard of for any but the most routine of rules. The NRC cranks out a proposed and final fee rule every year, and it issues routine corrections as final rules quite easily. Anything harder? Not in my three years leading the agency’s centralized rulemaking organization, and not that I could easily recall in the rest of my two decades there. It’s going to take some major creativity and brainpower.

The closest analogy is the rulemaking led by the Subsequent License Environmental Directorate. They had a huge burning platform—the Commission’s adjudicatory decision that the license renewal generic environmental impact statement (GEIS) did not address subsequent license renewal (SLR). That meant previously extended licenses were instantly un-extended, and all SLR work was paused until this issue could be fixed generically or dealt with for each site.

In a little over 26 months, a hand-picked group revised the license renewal generic environmental impact statement and updated the rules to enable its use for subsequent renewals (60 to 80 year terms). Here’s what that looked like:

Setting aside any substance of the rulemaking at hand today and just looking at the schedule, what could be pared down to get from 26 months to 18 months?

* No rulemaking plan—there is simply no time for a traditional one. This piece of the process exists mainly so the staff doesn’t spend a bunch of time on rules without the Commission knowing or approving. Trust me, the Commission knows about this rule and has plenty of ways to give direction if needed.

* Shorter regulatory analyses—when you have to do something, the costs and benefits matter less. This rule might not be the time for the NRC’s standard “check every box” cost analysis. If it’s needed to choose among options, sure. But I’m not sure what we learn by trying to quantify the costs and benefits of an 18-month review schedule.

* Approval efficiencies—everyone has to play along. Several rules (like one on source security) have been sped up significantly by cutting bit players out of the approval chain, putting responsible managers in one room to agree on final language, and holding everyone to tough deadlines.

* Interagency partnering—OMB especially has to work fast. Notwithstanding the February EO that added review responsibilities, this part of the White House already owns a chunk of the timeline to review rules for compliance with the Paperwork Reduction Act and Congressional Review Act. Along with the other agencies that will be interested in the substance of the rule and the Federal Register staff who need to publish it, there will be a lot of favors to call in.

When these steps go wrong, even the best of rules can founder. Look at… well, there are plenty of examples.

Assemble the Kitchen Crew

The order gets pretty precise that the NRC must “create a dedicated team of at least 20 officials to draft the new regulations.” With 10 different parts of the rule called out specifically, along with the more nebulous “wholesale review,” there will surely be way more than 20 people interested. The key will be paring it down to a few experts each on reactor licensing, environmental reviews, radiation protection, and the other miscellaneous topics. Then, they’ll need managers with the clout to shove this rule through the process while shaping external expectations.

Except—the NRC also needs their best people in each of those areas to do the actual licensing, environmental, and managerial work at the exact same time (and oh by the way, doing a reorg/RIF). This is going to make the whole thing extra hard. I don’t envy those folks.

Keep Extra Cooks Out

The NRC is a very public and collegial place. They’re going to want to fully understand and address differing views among the staff. They’re going to feel bound to write lengthy dispositions of each public comment. They’re going to get asked for lots of public meetings.

I don’t think that’s happening in 18 months. It’s kind of like the old “pick two”—fast, good, or cheap. I’m not completely sure they can even get two of those. Fast and minimally disruptive (which is one version of good) is achievable with the right teamwork and management. Fast and thorough, with everyone happy with what you did? Probably not.

Outsiders can help the most by keeping out of it unless they have something substantive and helpful to contribute. “Me too” comments that have to get counted and dispositioned will distract. Non-actionable feedback will distract. Letters sent at the wrong time in the process will distract. Go to the meetings. Provide concrete suggestions when asked. Otherwise stay out of the kitchen.

Try Some Sous-Vide or Molecular Gastronomy

Writing from an AI startup, it’s essential that I mention how modern technology can help cook this feast. The NRC, its partner agencies, and public commenters all have access to powerful tools that can help them research, synthesize, and draft their pieces of this rulemaking puzzle.

* AI assistance can accelerate the most painful parts of the process, like:

* digesting documents put out for comment and comparing them to other sources or ideas

* developing comments, explanations, and transmittals

* creating boilerplate, historical, and summary sections of rulemaking documents

* checking for consistency across large documents or sets of documents

Oh, and Actually Make It Taste Good

Too much to ask after all that? Yeah, maybe. Stay tuned for 3 more articles on the substance of the rule that we set aside above:

1. Piece of (Yellow)Cake - The Easier Parts of the NRC-Reform Rule

2. ALARA Alarm - Radiation Protection Standard Changes

3. Atomic Highway - Streamlined Reactor Licensing