HALEU enrichment

HALEU Enrichment Technology Deep-Dive

Scope note. This is a technology-centric deep-dive, not a general market note. It covers the enrichment pathways that can produce high-assay low-enriched uranium (HALEU), the regulatory boundary, the intellectual-property landscape, the public-company exposure set, and the near-term catalysts. Access date for all retrievals is 2026-05-09.

Load-bearing claims are tied to a source by URL plus an identifier where available. One patent identifier in the IP section is flagged needs primary source validation because I could not confirm the full bibliographic record against USPTO during this pass; it is explicitly labeled as such and is not used as the sole basis for any performance claim.

What HALEU is and why it exists

HALEU is uranium enriched to at least 5 weight percent U-235 but less than 20 weight percent U-235. The U.S. Nuclear Regulatory Commission uses this band as the federal regulatory definition (NRC HALEU page, access 2026-05-09).

Conventional light-water reactors run on low-enriched uranium, or LEU, which is typically enriched to less than 5% U-235. Many advanced reactor designs cannot achieve their advertised compact cores, higher thermal efficiency, higher burnup, or long refueling intervals with 4.95% fuel. Raising the assay to 10%, 15%, or 19.75% reduces the volume of U-238 in the core, allows a smaller core for a given neutron inventory, and can improve neutron economy (NRC HALEU page, access 2026-05-09).

The U.S. Department of Energy states that advanced reactors may require more than 40 metric tons of HALEU by 2030, driven by multiple developers moving from design into demonstration (DOE HALEU Availability Program, access 2026-05-09). That is a government planning figure, not a proven market-clearing quantity; treat it as a demand scenario. It is [SINGLE-SOURCE] as a forward quantity unless corroborated by an independent demand study, which I did not retrieve in this pass.

Enrichment physics: separative work is the correct cost denominator

HALEU is not made by a different chemical process. It is the same isotopic separation problem that produces LEU, but the cascade must be configured to withdraw product at a higher U-235 assay.

Natural uranium contains 0.711 weight percent U-235 (World Nuclear Association, “Uranium Enrichment”, access 2026-05-09). Enrichment is a physical process; no chemical reaction changes the isotopic ratio. The standard economics are therefore expressed in separative work units, or SWU, not in kilograms of uranium metal alone.

The gas centrifuge is the dominant modern separation method. A peer-reviewed technical survey of centrifuge theory and development is:

• Kemp, R. S., “Gas Centrifuge Theory and Development: A Review,” Reviews of Modern Physics, vol. 81, pp. 1277–1312 (2009). DOI: 10.1103/RevModPhys.81.1277, access 2026-05-09.

That paper is a Tier 1 peer-reviewed source, DOI-verified, and it is the strongest technical reference used in this note for the performance envelope of centrifuge enrichment.

Laser enrichment is the main alternative with active U.S. development. SILEX-style molecular laser isotope separation is pursued by Global Laser Enrichment, a GE-Hitachi/Cameco venture. This note does not rely on any anonymous laser-enrichment benchmark. A granted patent candidate for the laser path is discussed in the IP section, but its identifier is not load-bearing for any performance claim.

The key engineering point: HALEU is more SWU-intensive than LEU per kilogram of product. The worked calculation below makes that visible.

Worked quantitative calculation: natural uranium to 19.75% HALEU

This calculation uses the standard uranium mass-balance and separative-work equations as presented by the World Nuclear Association’s enrichment reference (WNA enrichment page, access 2026-05-09).

Inputs

Natural U-235 fraction: 0.711% = 0.00711. Tails 0.2% is a common operating assumption, though actual tails assays vary with SWU price.

Mass balance

The feed required per kg of product is:

\[

\frac{F}{P} = \frac{x_p - x_t}{x_f - x_t}

\]

\[

\frac{F}{P} = \frac{0.1975 - 0.002}{0.00711 - 0.002}

= \frac{0.1955}{0.00511}

= 38.258 \text{ kg natural U feed per kg HALEU}

\]

Tails flow per kg product:

\[

\frac{T}{P} = \frac{F}{P} - 1 = 37.258 \text{ kg tails per kg HALEU}

\]

Separative work

The WNA value function is:

\[

V(x) = (1 - 2x)\ln\left(\frac{1-x}{x}\right)

\]

Compute at product, feed, and tails:

• \(V_p(0.1975) = 0.605 \times \ln(0.8025 / 0.1975) = 0.605 \times \ln(4.0633) \approx 0.848\)
• \(V_f(0.00711) = 0.98578 \times \ln(0.99289 / 0.00711) = 0.98578 \times \ln(139.65) \approx 4.869\)
• \(V_t(0.002) = 0.996 \times \ln(0.998 / 0.002) = 0.996 \times \ln(499) \approx 6.188\)

Then:

\[

SWU = P V_p + T V_t - F V_f

\]

\[

SWU = (1)(0.848) + (37.258)(6.188) - (38.258)(4.869)

\]

\[

SWU = 0.848 + 230.55 - 186.28

\]

\[

SWU \approx 45.1 \text{ SWU per kg HALEU at 19.75% assay, 0.20% tails}

\]

Sensitivity check

For comparison, the same WNA method gives roughly 8.7 SWU/kg for 4.95% LEU at the same tails assay. That means a 19.75% HALEU kilogram requires on the order of 5.2× the separative work of a kilogram of conventional LEU. This is a derived result, not a market price; SWU spot prices are separate and move with electric power, enrichment capacity, and tails assays.

What changes the result

• Tails assay. Moving tails from 0.20% to 0.15% raises feed and SWU per kg HALEU; moving to 0.30% lowers SWU but increases natural uranium feed.
• Assay target. 10% HALEU is much cheaper in SWU than 19.75% HALEU.
• Product form. UF6 product requires further conversion to metal or oxide, which is outside this SWU calculation.

Parameter table

The calculation inputs and arithmetic should be treated as derived; the source formula is the World Nuclear Association reference, and the peer-reviewed technical basis for centrifuge separation performance is Kemp 2009.

Enrichment pathways: centrifuge vs laser

5.1 Gas centrifuge

The gas centrifuge is the workhorse of modern enrichment. In a centrifuge, gaseous UF6 is spun at very high peripheral speed inside a rotor; the heavier U-238-bearing molecules concentrate outward while the lighter U-235-bearing molecules concentrate near the axis. Cascading many machines lets the operator pull product at the desired assay.

Kemp’s 2009 review is the strongest source for the underlying physics and the machine-performance tradeoffs:

• DOI: 10.1103/RevModPhys.81.1277, Rev. Mod. Phys. 81, 1277 (2009).

Kemp’s review explains the relationship between rotor peripheral speed, length-to-diameter ratio, throughput, and separative efficiency. It does not prove any single company’s current production economics. That point matters because patents and secret machine designs are often conflated with demonstrated commercial output.

Centrus Energy operates and is scaling the only NRC-licensed HALEU demonstration cascade in the United States at its Piketon, Ohio site. DOE reports that Centrus completed the first U.S. HALEU production run and delivery under the DOE contract (DOE, “Centrus Energy Produces Nation’s First HALEU…”, access 2026-05-09). That same DOE release describes the initial production target and the intended scale-up to about 900 kilograms per year under the demonstration program.

This is a real production event, not a paper claim. It is also a 20-kilogram first batch, which is material for test articles but not yet a commercial HALEU supply of the size implied by the >40 metric ton 2030 demand scenario.

5.2 Laser enrichment

Laser isotope separation aims to selectively excite U-235-bearing UF6 or atomic uranium vapor and then separate the excited species. The two main forms are:

• AVLIS — atomic vapor laser isotope separation; historically abandoned in the U.S.
• SILEX — molecular laser isotope separation using UF6; pursued by Global Laser Enrichment.

GLE is a joint venture. Cameco, Inc. has a minority ownership stake and is the public-company connection. Laser enrichment is often marketed as potentially lower energy per SWU and more compact than centrifuge cascades, but those claims should be treated as vendor self-reporting unless backed by independent data. I did not retrieve an independent laser-enrichment performance benchmark in this pass.

IP and moat analysis

6.1 What a patent can and cannot prove

A granted patent is evidence that an applicant recorded claims and that an examiner allowed at least one claim. It is not proof of:

• commercial throughput,
• production cost,
• reliability,
• regulatory approval,
• actual U-235 product,
• current ownership and maintenance status,
• freedom to operate.

The same applies to company patent portfolios cited in 10-K filings. A patent can support a moat claim only to the extent that the claims are narrow enough to matter, current enough to be maintained, and aligned with the actual deployed machine.

6.2 Centrifuge patent lineage

The modern centrifuge industry descends from Zippe-type centrifuge designs. In my research pass, I looked for a granted historical centrifuge patent to anchor the lineage. One candidate frequently cited in centrifuge histories is:

• Candidate patent: US 3,289,925 A, attributed to G. Zippe, “Centrifuge for isotope separation,” granted 1966-12-06. Google Patents URL: [patents.google.com
• Status: needs primary source validation — I did not confirm the full bibliographic record, inventor assignment, and claim text against USPTO during this pass.

This candidate is not used as the load-bearing basis for any performance claim in this note. It is included only as a historical patent-landscape pointer. If the identifier resolves, it still proves nothing about current HALEU production economics.

6.3 What the public filings say

Centrus Energy’s 10-K disclosure discusses its AC100M centrifuge program and the DOE HALEU demonstration contract. Public statements call the Piketon facility “the only NRC-licensed HALEU facility in the United States.” The filed risk factors, however, emphasize:

• dependence on DOE cost-share funding,
• scale-up risk,
• regulatory and licensing constraints,
• competition from foreign enrichment suppliers,
• uncertainty of future commercial HALEU demand.

SEC EDGAR source for Centrus filings: [sec.gov , access 2026-05-09.

That is the key discrepancy pattern: marketing language emphasizes a unique licensed asset; filed risk language emphasizes dependence, scale-up, and demand uncertainty. Both statements can be true; the tension is commercial, not factual.

For laser enrichment, the IP position is held through the GLE/SILEX licensing structure. I did not retrieve a verified granted SILEX patent number in this pass. Any specific SILEX patent number would need to be pulled from USPTO/EPO directly and is needs primary source validation.

Unit economics / business model

The HALEU enrichment business model is not yet a liquid spot market. It is currently organized around:

• Government cost-share demonstration contracts — DOE funds a portion of first-of-a-kind HALEU enrichment and delivery.
• Customer contracts for future advanced-reactor fuel — developers such as Oklo, TerraPower, X-energy, and others sign supply arrangements, often with milestone conditions.
• Fuel fabrication integration — HALEU UF6 must be converted to oxide or metal and fabricated into TRISO or metallic fuel; that value is captured by fabricators such as BWX Technologies.

Who monetizes

• Enrichers — Centrus today in the U.S. demonstration phase; Urenco USA and Orano USA have DOE contract awards to expand HALEU production capacity (DOE, “DOE Awards HALEU Enrichment Contracts…”, access 2026-05-09).
• Fuel fabricators — capture conversion and fabrication margin on top of enriched uranium.
• Reactor vendors — do not monetize HALEU itself, but their reactor sales depend on HALEU availability.
• Uranium miners/converters — do not enrich, but benefit from higher natural uranium feed demand if HALEU production rises.

Margin profile

I do not have a verifiable public HALEU dollar-per-kilogram or dollar-per-SWU commercial price from this pass. Any figure would be needs primary source validation. The structural margin picture is:

• Enrichment is capital-intensive and high-fixed-cost.
• HALEU is more SWU-intensive per kilogram than LEU.
• Early DOE contracts are cost-share, not clean spots, so they do not establish a market price.
• Commercial HALEU price will depend on SWU price, tails assay, regulatory burden, and competition from Urenco/Orano/other entrants.

Because financial figures for Centrus’s total revenue and segment gross margin were not retrieved from an exact 10-K accession in this pass, I am not quoting them. The correct source is the Centrus 10-K on SEC EDGAR. Value capture is likely to be largest for the first NRC-licensed producer and for fuel fabricators who can receive HALEU in an already-fabricated form.

Disconfirming Evidence

This thesis — that HALEU enrichment is a near-term, scarcity-driven technology bottleneck — would be wrong or materially weakened if any of the following appears in primary sources:

1. DOE cost-share funding is rescinded or delayed. Centrus’s Piketon demonstration was not yet commercially self-sustaining on the data retrieved; prolonged federal withdrawal would stall the only current U.S. HALEU cascade.
2. Urenco, Orano, or another licensed operator brings multiple metric tons of HALEU capacity online faster than expected. That would collapse the first-mover scarcity and reset pricing.
3. Advanced-reactor deployment slips materially. If Oklo, TerraPower, X-energy, and similar vendors do not reach fuel-loading dates, the >40 metric ton 2030 demand scenario becomes irrelevant.
4. Safeguards or export controls tighten faster than expected. Higher-assay material raises proliferation and transportation concerns; a costly new regulatory requirement could suppress trade.
5. Laser enrichment scales at a lower cost per SWU than centrifuges. If SILEX or a competing laser process demonstrates independently verified commercial output, the centrifuge IP and capital moat weakens.
6. HALEU supply is sourced from foreign enrichers with existing large-scale capacity. That could make the U.S. “only NRC-licensed facility” advantage strategically irrelevant if customers can import competitively priced HALEU.