The Nuclear Licensing Market Is Splitting in Two. Here Is What the Data Shows.

Intended for advanced reactor developers, investors, workforce development, and licensing professionals evaluating where the nuclear licensing market is heading — and which capabilities will hold value as it gets there.

Executive Summary

The advanced reactor field looks crowded from the outside — 119 companies and 172 unique reactor designs, per the Gibboney Nuclear July 2026 dataset. But demand for licensing capability and the supply of people who can deliver it are moving in opposite directions inside two different categories of work. Technology-specific, safety-related licensing — the kind that produces topical reports for a Construction Permit Application or Combined Operating License Application — is becoming more valuable, because it requires named, accountable humans and cannot be fully automated with AI tools. Environmental and siting review is becoming less differentiated, because the NRC itself is actively narrowing what that work requires, through actual documented rulemaking, not just rhetoric. Contrasting the demand side, the supply of licensing talent has no real university pipeline behind it — nuclear licensing engineering is not a degree program, and industry certificate courses, not academia, are currently filling the gap. The firms best positioned in this market are technical specialists with real regulatory-engagement track records, not generalist body shops sized for a NEPA-heavy world that is shrinking.

What the Numbers Actually Show About Demand

The Gibboney Nuclear July 2026 dataset tracks 119 fission companies developing 172 unique reactor designs worldwide, benchmarked against WNA, DOE GAIN, and OECD NEA — all three of which undercount the real market and some still count defunct designs (e.g. Gen4 Energy, still counted by WNA, shut down April 1, 2018).

Of those 119 companies, only a fraction have any recorded NRC regulatory engagement at all. Cross-checked directly against NRC.gov's own pre-application activities page — after removing three site listings that double-count a vendor already listed separately under its own name (Energy Northwest duplicates X-energy's Xe-100 project; Texas A&M's RELLIS Campus duplicates Last Energy; Abilene Christian University duplicates Natura Resources) — 34 unique projects are in active NRC engagement. Six have a CP or COL license granted. Zero are in commercial operation today.

That is not licensing activity spread evenly across 119 companies. It is concentrated in roughly a quarter of the field, and active engagement means the process has started — not that it is close to finished. The remaining companies are largely still in pre-licensing/R&D, which is not a red flag by itself; it is where avoidable design rework gets caught before it becomes expensive. But it is not licensing progress towards a commercial application, and should not be described as such to investors.

What the Numbers Show About Supply

Supply is harder to measure precisely, but the available signals point the same direction: thin.

A LinkedIn Sales Navigator search for nuclear licensing professionals returns roughly 222 people — a real, filtered lead-generation count, though not a verified, complete census of the field either. Weighed against 119 companies competing for the same regulatory bandwidth, engineering time, and NRC review slots, the constraint is visible even before you look at team-level need detail.

Both ARDP Awardee Construction Permit Application teams currently in the front of the field — X-energy's Xe-100 and TerraPower's Natrium — required licensing teams of 20 to 25 people to compile a CPA, with substantial contractor support layered on top. That is what it actually takes to move a reactor design through a CPA: not a single licensing lead and a template, but a structured team executing a document trail that ties every Statement of Fact back to a technical basis.

This shortage will not resolve through the traditional pipeline, because there isn't one. Nuclear licensing engineering is not a degree program. Most nuclear engineering graduates encounter licensing through, at most, a single elective — the University of Tennessee's nuclear engineering curriculum, for example, has offered one such course, taught by an adjunct professor from industry. Universities are producing reactor designers and core physicists; they are not producing licensing engineers in any systematic way.

What is filling the gap is coming from industry groups, not academia. The American Nuclear Society's Nuclear Licensing & Regulation Certificate Course — a self-paced, 16-hour, 13-module program covering NRC and DOE regulatory frameworks — is one of the few structured entry points into the field. I am currently enrolled in it myself, not because I need an introduction to the material, but because it offers 16 professional development hours toward renewing my Virginia PE license. That a 17-year licensing veteran's continuing-education requirement is being satisfied by the same course built to onboard newcomers says something about the training infrastructure gap: there is no dedicated continuing-education pathway for practicing licensing engineers, any more than there is a degree pathway for new ones.

The Work Becoming More Valuable: Technology-Specific, Safety-Related Licensing

Every safety-related licensing document — a Safety Analysis Report, a QAPD, a safety basis topical report — requires a doer and an independent verifier: two separate, accountable people, not two passes of the same AI model. This is not a general quality-management preference; it is a specific requirement of ASME NQA-1, the nuclear industry's quality assurance standard formerly endorsed by the NRC in RG 1.28, which mandates that design verification be performed by someone independent of the individual who did the original design work.

This is a meaningful departure from generic quality standards. ISO 9001 requires design and development verification, but it does not require that verifier to be organizationally independent from the person who produced the design — a distinction the Department of Energy's own comparison of the two standards states directly. That gap is exactly where nuclear licensing sits apart from ordinary quality management: it is not enough to check the work. Someone with no hand in producing it has to check it, and be a human accountable for that check.

AI tools can draft this content. They cannot independently verify their own output, and they cannot be held accountable when the NRC asks who is responsible for what is in the document. That requirement does not go away because drafting got faster, and it is the reason this category of work resists commoditization even as AI-generated licensing tools proliferate.

The Work Becoming Less Valuable: Environmental and Siting Review

This is not speculation. The NRC is actively, and quite publicly, narrowing what its own environmental reviews require.

A final rule amending 10 CFR Part 51's categorical exclusions took effect April 29, 2026 (Docket NRC-2018-0300), eliminating the need to prepare environmental assessments for a range of NRC licensing, regulatory, and administrative actions.

A much larger proposed rule followed on July 7, 2026 (Docket NRC-2025-0478): it would eliminate the requirement for draft environmental impact statements, add new categorical exclusions, and — most significantly — narrow the scope of NRC environmental review to impacts with a direct connection to radiological effects, largely excluding chemical and other non-radiological hazards. Comments are open through August 21, 2026, with a public hearing during that window.

This traces to a specific, citable paper trail, not announcement fluff: Executive Order 14300 ("Ordering the Reform of the Nuclear Regulatory Commission"), EO 14154 ("Unleashing American Energy"), EO 14192 ("Unleashing Prosperity Through Deregulation"), and EO 14270 ("Zero-Based Regulatory Budgeting to Unleash American Energy"). The Commission formally directed staff toward this narrower scope on July 28, 2025, via SRM-SECY-24-0046. The NRC also cites the Supreme Court's 2025 decision in Seven County Infrastructure Coalition v. Eagle County, Colorado, which held that an agency need not examine impacts outside its own regulatory authority.

This rulemaking is contested — critics argue that narrowing review to radiological-only impacts and removing cumulative-impact analysis eliminates a legal safeguard that has held agencies accountable under NEPA since Calvert Cliffs' Coordinating Committee v. Atomic Energy Commission (1971). That dispute will play out through the comment period and likely litigation. What is not in dispute is that the scope reduction is real, dated, and moving through the Federal Register — which is exactly why the environmental and siting review segment of the licensing market is the one most exposed to compression. Less required analysis means less differentiated technical judgment, and work that leans toward data organization rather than safety-basis reasoning is the work most amenable to a single reviewer checking AI-generated output, not a full firm.

Who Is Positioned to Deliver

The firms built for a NEPA-heavy world — sized to produce lengthy environmental impact statements across every non-radiological hazard category — are staffed for a scope of review the NRC is actively shrinking. The firms built around technology-specific, safety-related licensing — SAR development, QA program design, safety-basis defense, the doer/independent-verifier structure that cannot be automated away — are staffed for the part of the market that is not shrinking, and cannot be, without someone accountable signing their name to it.

That is the distinction worth evaluating a licensing partner against: not headcount, and not how many environmental impact statements they've produced, but whether they can staff and defend a safety-related document trail through NRC review. Years of advanced reactor design pre-application work says more about that capability than a general licensing firm's website's site selection framework does.

I have lived this shortage from the inside, not observed it from a distance. Across the CPA teams I have worked, licensing engineers regularly worked substantial overtime, because there were not enough qualified people and the schedule did not move to accommodate that. We drafted design engineers into licensing roles and taught them, on the job, how to write licensing documentation, because there was no other bench to pull from. It was not sustainable at the time, and it was not a good use of anyone's specialized training — the design engineers were needed on design, and the licensing engineers were needed on licensing, and instead everyone was stretched across both.

There is one way out of that cycle, and it is not more overtime. It is growing more licensing engineers. This is why Gibboney Nuclear offers advisory and training services to advanced reactor developers, built specifically around developing internal licensing capability inside client organizations rather than just producing documents for them. Every developer who trains one more licensing engineer in-house makes the shortage marginally less acute — for that company, and for the other 222 people doing this work industry-wide. That is not a favor to the field. It is the only version of this market that scales.

What Small Teams Should Do Now

• Treat NRC docket status and active engagement — not press releases — as your licensing progress indicator.

• Budget for a real CPA-level licensing effort: 20-25 people plus contractor support, not a single hire.

• Distinguish your environmental/siting work from your safety-basis work when selecting a licensing partner. They are increasingly diverging markets with different risk profiles.

• Watch Docket NRC-2025-0478 through the August 21, 2026 comment deadline — the final scope of Part 51's narrowing will affect ESP cost and timeline assumptions industry-wide.

• Ask any AI-assisted licensing tool vendor directly who takes independent-verifier accountability for their output. If the answer is "the software," it is not produced under a QA program and you still need to budget for two individuals to comply with ASME NQA-1.

• Build internal licensing capacity early, rather than assuming you can hire your way into a CPA-ready team when the schedule demands it. The talent pool is too thin (and by extension, expensive) for that to be a realistic plan.

Frequently Asked Questions

Q: How many advanced reactor companies and designs currently exist?

A: The Gibboney Nuclear July 2026 dataset tracks 119 unique fission companies developing 172 reactor designs worldwide, benchmarked against WNA, DOE GAIN, and OECD NEA databases, all of which undercount the real market.

Q: How many of those companies are in active NRC regulatory engagement?

A: Cross-checked against NRC.gov's pre-application activities page and de-duplicated for host-site listings that double-count an already-listed vendor, 34 unique projects are in active NRC engagement as of July 2026. Six have a license granted; none are in commercial operation.

Q: Is any advanced reactor design actually in commercial operation today?

A: No. The only "in operation" entry in the Gibboney Nuclear dataset is Japan Atomic Energy Agency's HTTR, a 30 MWth research and test reactor with zero MWe of electric output — not a commercial power source.

Q: How large is a real Construction Permit Application licensing team?

A: Based on direct experience on two CPA teams — X-energy's Xe-100 and TerraPower's Natrium — a CPA-level licensing effort requires roughly 20 to 25 people, with substantial contractor support beyond that core team.

Q: Is the NRC actually deregulating environmental review requirements, or is that rhetoric?

A: It is documented rulemaking, not rhetoric. A final rule narrowing categorical exclusions under 10 CFR Part 51 took effect April 29, 2026 (Docket NRC-2018-0300), and a larger proposed rule narrowing NEPA review scope to radiological impacts was published July 7, 2026 (Docket NRC-2025-0478), with comments open through August 21, 2026.

Q: Will AI replace nuclear licensing consultants?

A: Not for safety-related content. Every safety-related licensing document requires a doer and an independent verifier — two accountable people. AI can draft; it cannot independently verify itself or be held accountable to the NRC. Environmental and siting review, which is less safety-related and more data-organization-driven, is more exposed to AI-assisted compression, particularly as the NRC narrows what that review requires.

Q: Does ISO 9001 require the same independent verification as ASME NQA-1?

A: No. ISO 9001 requires design and development verification, but it does not require the verifier to be organizationally independent from the person who performed the original design work. ASME NQA-1 does require that independence for safety-related items, per DOE's own published comparison of the two standards. This is a substantive difference, not a paperwork one, and it is the regulatory basis for why safety-related licensing content needs a separate accountable reviewer, not just a second pass.

Q: Is nuclear licensing engineering a degree program?

A: No. Most nuclear engineering programs include, at most, one elective covering licensing and regulation — the University of Tennessee's program, for example, has offered a single such course, taught by an adjunct professor. The primary structured entry point currently comes from industry, such as the American Nuclear Society's Nuclear Licensing & Regulation Certificate Course, not from university curricula.

Q: How can a developer address the licensing engineer shortage rather than just compete for scarce talent?

A: By building internal licensing capacity rather than relying solely on external hires. Gibboney Nuclear offers advisory and training services structured around developing licensing capability inside a client's own organization, not just producing documents on their behalf. Growing new licensing engineers, rather than redistributing the existing pool, is the only version of this shortage that improves industry-wide.

Q: What should investors look for when evaluating a company's licensing readiness?

A: Active NRC engagement status — not the announcement — and the size and structure of the licensing team behind the design. A named docket number and a 20+ person licensing effort say more than a press release.

Sarah Gibboney, P.E. is the Founder of Gibboney Nuclear, PLLC, a nuclear licensing consultancy serving advanced reactor developers. She has 17 years of nuclear energy experience and co-authored Construction Permit Applications for both ARDP awardees — TerraPower's Natrium and X-energy's Xe-100.