The HALEU Transport Problem Isn’t Radiation. It’s Criticality

The NRC’s proposed package rule solves a real problem. It isn’t the one blocking fresh fuel.

Guest Column by: Elliot Marsh

On July 27, 2026, the NRC published a proposed rule that would modernize how fuel transportation packages get certified. It carries Docket ID NRC-2025-1667, and the comment period closes August 26.

The nuclear trade press coverage framed it as a step toward easier microreactor deployment, and that framing is fair: the proposal would allow higher radiation level limits for certain Type B packages, including those carrying irradiated microreactors, and would let applicants use risk-informed or performance-based approaches to demonstrate compliance.

Both changes are genuinely useful. Neither one touches the constraint that actually governs moving high assay low enriched uranium (HALEU). Shielding is an engineering problem, and engineering problems yield to money and steel. You can make shielding thinner. You cannot negotiate with a fission cross-section.

The Key Issues

The NRC’s proposed package rule (Docket NRC-2025-1667) would modernize certification, but its substantive relief addresses radiation dose limits and testing methodology, not criticality.

• An advanced fuel cycle moves three things: fresh fuel, fresh fueled reactors, and irradiated reactors. Only the third is governed by dose, and it is the only one the rule helps. The fissile-package criticality rules in 10 CFR 71.55 and 71.59 are not amended, not discussed, and no comment was invited on them.

• The constraint on the first two legs sits at five percent. The moderation exception in 71.55(g) lets UF6 shippers assume water stays out of a package, and it is capped at 5 wt% U-235, thefloor of the HALEU band. Above that line, packages must be shown subcritical fully flooded, at optimal moderation and full water reflection. Packages built for higher enrichments carry internal criticality-control hardware that occupies volume the fuel would otherwise use.

• Behind the regulation is a data problem. Licensing a fissile package requires a validated criticality analysis, and validation requires benchmark experiments that for commercial-scale HALEU are still being funded. Congress directed the work in the Energy Act of 2020, and DOE’s first proposal call opened with UF6 transport under moderator exclusion and the 10-to-20 percent enrichment gap. In the interim the NRC has been requiring a larger administrative margin. The bottleneck is not casks; it is the experimental basis needed to license them.

• Approved packages do exist, some certified well above 20 percent. What does not exist is throughput. Centrus expects 12 metric tons of annual HALEU capacity from its initial build-out, with first new capacity by 2029. DOE is funding five companies on two-and-three-year schedules to develop and license transport packages. Enrichment capacity now has a funded schedule; transport capacity has a licensing schedule that depends on data that are still to be generated for use to license transport packages.

It helps to separate what actually moves in an advanced reactor fuel cycle. Readers here will recognise the three-leg framing, and it does real work:

Leg 3 is the one most people picture: a used core, hot with fission products, needing shielding and cooling time. That is a dose problem, and dose problems are what package rules have historically been written around.

Legs 1 and 2 are different, and leg 2 is the one that gets missed. Westinghouse describes the eVinci as factory-built, fueled, and assembled before it ships in a container. Radiant says its Kaleidos units are assembled, fueled, and tested in the factory, and the Part 70 licence it is pursuing would let it load fuel before shipping reactors to customers.

The tell is in Westinghouse’s own design description: shutdown rods are inserted during transport. That is not a package carrying inert cargo. It is a subcritical system being actively held subcritical while it moves — and nothing in the dose regime has anything to say about it.

A fresh core has no fission product inventory — no cesium, no strontium, essentially no gamma source term worth designing around. From a dose standpoint it is close to inert. From a criticality standpoint it is a fissile package with a deliberately reactive geometry, and it has to be licensed as one.

HALEU is enriched between 5 and 20 percent U235. That single number puts all three legs into fissile-package territory, but it does something more specific to legs 1 and 2 — which is where the rest of this article goes.

Read the amendment list rather than the press summary. The proposal would amend these sections: §§ 71.1, 71.4, 71.5, 71.17, 71.41, 71.45, 71.47, and 71.95 — a new paragraph (e) in 71.41, revised language in 71.47(b)(3), and a set of administrative changes.

Sections 71.55 and 71.59 — the fissile-package criticality standards — are not on that list. They are not amended, not requested for comment, not discussed.

The dose relief is also narrower than the headlines suggest.

• The existing 0.1 mSv/h (10 mrem/h) limit at 2 meters from the vehicle stays in place;
• The proposed 0.5 mSv/h (50 mrem/h) alternative applies only where package contents meet the highway route controlled quantity definition in 49 CFR 173.403 and
• The licensee coordinates in advance with Federal, State, or local inspection authorities.

The NRC is explicit about who that gate is built for. Irradiated microreactors will likely contain a quantity of radioactive material greater than 3,000 A₂, which is what brings them inside the HRCQ definition in the first place.

The stated motivation points the same direction: applications involving considerations not addressed in current regulations, such as limited cooling time for the fuel. Limited cooling time is, by definition, a post-irradiation condition.

Even the technical basis is scoped to the return leg. The supporting PNNL-38760 analysis evaluated microreactor transportation scenarios, and the NRC notes it may not bound radiation fields associated with other Type B package contents — anything other than a TRISO-based heat pipe microreactor would require the staff to confirm the design can be safely accommodated.

None of this is a criticism of the rule. The new § 71.41(e) would let applicants substitute an NRC-accepted methodology for the physical testing requirements in §§ 71.71, 71.73, and 71.61 — the pathway the agency already endorsed in SECY-24-0062 for the Project Pele TRISO-based package. That is meaningful flexibility, and package designers should use it.

The rule helps the return leg. It does not help the delivery leg.

Here is the mechanism, and it sits in a section the proposed rule does not open.

Most enriched UF₆ moving today travels under an exception. Section 71.55(g) excepts packages containing uranium hexafluoride only from the requirements of paragraph (b), provided four conditions hold — the last of which is that the uranium is enriched to not more than 5 weight percent uranium-235.

Lose that exception and paragraph (b) applies in full. The package must remain subcritical if water leaks into the containment system, assessed at the most reactive credible configuration, with moderation by water to the most reactive credible extent, and close full reflection of the containment system by water on all sides.

That is the cliff. Below 5 percent, you may design on the assumption that water stays out. Above it, you analyze the package flooded. Water is a moderator, and a flooded HALEU cylinder is not a marginally harder criticality problem than a flooded LEU cylinder — it is a different one.

The regulatory threshold at 5 wt% U-235, and the sections the proposed rule leaves untouched.

Industry says this plainly. Orano has described evaluations intended to define limits on the number of UF₆ packages that can be transported in a conveyance and consignment without using the moderation exception rule in 10 CFR 71.55(g), or SSR-6 para. 680, that limits enrichment to 5 percent.

One precision matters here: 71.55(g) is a UF₆-only exception, which is the 30B cylinder world. Fabricated oxide and powder fall under the general requirements of 71.55(b), with array behavior governed by 71.59. Different paragraphs, same cliff edge.

What the cliff costs shows up in payloads. Up to 48 VP-55 units can be loaded into an ISO 20-foot container for a total payload, for example, of up to 2,880 kg of UO₂ powder enriched to 20 percent U-235. That is roughly 60 kg of oxide per drum.

The VP-55 has a packaging tare weight of 390 lbs against a maximum gross weight of 750 lbs — and the cavity is not empty space waiting for product.

A high-capacity basket, constructed of an aluminum frame, insulation, and neutron moderating material, may be used in conjunction with a 5-inch pipe container for increased content limits. Criticality control is not a line item on the certificate. It is hardware, and it occupies the volume the fuel would otherwise use.

Now put that against where the enrichment work actually goes. About 90 percent of the separative work required to enrich natural uranium from 0.711 percent to 19.75 percent U-235 is expended in the 0.711 to 10 percent range.

The economics of enrichment are concentrated at the bottom of the ladder. The regulatory constraint on moving the product sits at the top. That mismatch is the story.

Licensing a fissile package requires a validated criticality analysis. Validating that analysis requires benchmark experiments resembling the system being analyzed. For commercial-scale HALEU, those experiments are still being funded.

This is not an inference. Congress wrote it into statute. Under Part (a)(2)(A) of the Energy Act of 2020, DOE is directed to develop, in consultation with the NRC, criticality benchmark data to assist the NRC in the licensing and regulation of fuel fabrication and enrichment facilities under 10 CFR Part 70, Domestic Licensing Of Special Nuclear Material, and in the certification of transportation packages under 10 CFR Part 71 Packaging And Transportation Of Radioactive Material.

The Joint DOE / NRC Program

The vehicle is a joint program: the DOE/NRC collaboration for criticality safety support for commercial-scale HALEU fuel cycles and transportation, known as DNCSH, established under Section 2001 of that Act and the Inflation Reduction Act.

The DOE/NRC Criticality Safety for Commercial-Scale High-Assay Low-Enriched Uranium (HALEU) for Fuel Cycle and Transportation (DNCSH) initiative is a collaborative effort between the U.S. Department of Energy (DOE) and the U.S. Nuclear Regulatory Commission (NRC) to establish new criticality benchmark data for commercial-scale nuclear energy.

The results will be crucial for the regulation and licensing of facilities handling special nuclear material and the certification of transportation packages for such materials, as outlined in the federal regulations. The project also  engages in activities that support the efficiency and robustness of licensing and criticality safety benchmark development processes such as nuclear data improvements.

Look at what it is buying. DOE is funding 16 projects (Project list – PDF file) through the first proposal call, across five topic areas: UF₆ transportation with moderator exclusion; the 10 to 20 percent enrichment gap; non-fissile material validation; fissile salts; and advanced moderator nuclear data.

These projects will support the development of data that will be useful to the NRC licensing evaluation process and industry’s licensing submittals pertaining to commercial-scale HALEU operations. The publicly available data developed from these projects will enable efficient future design and safety reviews and help the nuclear industry develop new and novel solutions to address data gaps.

The first topic area is the exception discussed above, named directly. The second is the enrichment band that exception excludes. A federal criticality-data program opened its funding list with precisely the two gaps that govern whether HALEU can move at commercial scale.

And missing data is not free while you wait for it. The NRC has been requiring a larger administrative margin because the experiments do not exist. That margin is not an abstraction — it is conservatism, and conservatism is paid for in payload. It is the same tax the package hardware collects, arriving through a different door.

The program is still extending into leg 2. Call #3 — the last of three — focuses on data needs supporting transportable microreactors.

So the bottleneck is not casks. It is the experimental basis required to license casks. Steel can be fabricated on a schedule. Critical experiments cannot be ordered from a catalog.

It would be wrong to say no approved packages exist. The Versa-Pac is a certified Type A fissile packaging under NRC Certificate USA/9342/AF-96, available in two drum sizes (55 gallons (VP-55) or 110 gallons (VP-110) , and multiple transport package options exist today for uranium oxides, metals, and alloys — some capable of carrying enrichments as high as 100 percent. The Versa-Pac is licensed for the following contents (also licensed for uranium hexafluoride in 1S and 2S cylinders):

• Uranium oxides
• Uranyl nitrate crystals
• Uranyl fluorides and carbonates
• Uranium metal and alloys
• Natural thorium
• TRISO fuel and compact

The problem is not existence. It is throughput, and the industry is acting like it. Orano and Urenco are developing a new cylinder designed for transport of uranium enriched up to 20 percent. DOE announced $11 million in awards to five U.S. companies to develop and license new or modified HALEU transportation packages, intended to establish long-term, economical transport capability. Projects developing new designs run up to three years; the modification effort runs up to two.

Now set that against demand and supply. NEI’s 2020 survey put demand for enrichment above 10 percent at 137 metric tons in 2030, rising to 501 tons by 2035. On the supply side, Centrus has said its initial build-out will include 12 metric tons of annual HALEU production capacity, with the first new capacity expected online by 2029 — under a task order awarded through DOE’s $2.7 billion domestic enrichment program.

Enrichment capacity now has a funded schedule. Transport capacity has a licensing schedule, and licensing schedules depend on data that is still being generated.

There is one more signal worth noting. When the NRC authorized the construction permit for Kemmerer Unit 1, the NRC staff granted four exemptions from existing regulations — including ones covering HALEU fuel handling and criticality requirements. A reactor that has not yet been built already needed relief from criticality rules written for a five percent world. The fuel has to reach the site before any of that matters.

Comments are due by August 26, 2026, under Docket ID NRC-2025-1667.

The proposed rule is good at what it does. It would let applicants use alternative demonstration methods in place of prescriptive physical testing, align tie-down standards with the IAEA’s SSG-26 values, and permit electronic submission of reports. The higher dose limit is well-supported for the case it was built around, and package designers should use the flexibility.

But note where the agency asked for help. The NRC posed specific questions on whether the increased dose-rate provision should apply broadly or under defined criteria, and what would justify higher radiation levels for packages outside the HRCQ definition. Every question concerns dose.

No parallel question was asked about criticality. The subpart governing fissile packages was not opened, and no comment was invited on it.

Thirty days is a short window to raise that. It is not too short to point out that the delivery leg is still waiting.

___________________________________________________
Elliot Marsh writes about reactor physics and nuclear fuel cycle engineering at NeutronRise (neutronrise.com).
Contact: contact@neutronrise.com