Licensing a NEW US Commercial Spent-Fuel Reprocessing Plant — The Real Regulatory Pathway

TL;DR — the one-paragraph truth

A complete, dedicated, reprocessing-specific NRC framework does NOT exist. A partial path exists (an applicant could be shoehorned into 10 CFR Part 50 or Part 70), but the NRC's own 2009 gap analysis found 23 gaps those parts don't fully address — and the NRC says those gaps "remain." The NRC tried to build the dedicated rule (2008–2016), formally discontinued it in 2021 for lack of any applicant, and stated it did not expect to even review an application for 10–20 years. As of mid-2026, no company has filed an actual license application — only letters of intent, pre-application engagement, white papers, and MOUs. Realistic decision-to-operation horizon: ~15–25+ years. Build cost converges at ~$20–30 billion (lifetime $100B+). Startup "early-to-mid 2030s" dates are marketing, not regulatory schedules. Live caveat: on June 24, 2026 (two days ago) the NRC proposed a Part 70 rule that would, for the first time, write reprocessing into the regs — so the precise statement today is "no complete framework; one is now in proposed-rule stage."

Sections: 1) Framework exists vs. incomplete · 2) Timeline in years · 3) Precedent (Barnwell/West Valley/Morris) · 4) Startup claims vs. reality · 5) Construction time + cost · 6) Confidence & caveats

1 · Does a complete NRC framework EXIST?

PARTIAL framework exists   NO complete reprocessing-specific framework

The partial path (what exists today)

An applicant could be licensed under 10 CFR Part 50 (the two-step Construction Permit + Operating License process, because reprocessing meets the Atomic Energy Act §11 definition of a "production facility" — it separates plutonium) OR under 10 CFR Part 70 (Domestic Licensing of Special Nuclear Material — the SNM-possession vehicle). The NRC's own words (2026 proposed rule):

"Under existing regulations, the NRC could license a spent fuel reprocessing facility under 10 CFR part 50 or 10 CFR part 70. However, 10 CFR part 70 does not explicitly include spent fuel reprocessing facilities or address production facilities."

NRC, Modernizing Materials Licensing, proposed rule, 91 FR 38124, June 24, 2026 — public-inspection.federalregister.gov/2026-12702.pdf · landing: federalregister.gov/...modernizing-materials-licensing

Part 52 (combined licenses / certified designs) is not the reprocessing vehicle — the new Part 70 path is explicitly offered as the alternative to the Part 50 two-step process. No primary source frames Part 52 as the route.

The KEY GAP (why "partial" ≠ "complete")

"In a gap analysis performed in 2009 (SECY-09-0082), the NRC staff identified 23 gaps between existing regulations and requirements that would be necessary... Given that, under the current framework, neither 10 CFR part 50 nor 10 CFR part 70 fully address these gaps, an applicant seeking a license under either part may need to request exemptions, and may require additional license conditions... The gaps, however, remain..."

NRC, Modernizing Materials Licensing, June 24, 2026 (same doc as above).

"The regulations in 10 CFR Part 70, as currently written, do not provide a regulatory framework to license a reprocessing facility. ... Currently, licensing a reprocessing facility under 10 CFR Part 50 would pose a significant hindrance to effective and efficient licensing."

NRC, Federal Register, Aug 2, 2010 (75 FR) — govinfo.gov/...FR-2010-08-02...2010-18888.htm

The 2008–2021 effort to build the dedicated rule — and its death

NRC's stated reason for killing it in 2021:

"...based on the estimated costs to conduct the rulemaking and the limited interest expressed or expected from industry to submit an application for any type of facility involving reprocessing technologies in the near-term... no industry stakeholders indicated that they plan to submit an application to the NRC for a reprocessing facility in the foreseeable future. ... The staff estimates that $2.5 million would be needed to complete a regulatory basis, develop and issue guidance, and develop a proposed and final rule."

NRC, Federal Register, "Spent Fuel Reprocessing," Jul 29, 2021 — govinfo.gov/...FR-2021-07-29...2021-16173.htm · federalregister.gov mirror · trade corroboration: World Nuclear News, Aug 9, 2021 — world-nuclear-news.org

Calibration note: the popular "2014 deferral" date is approximate. Primary-source dates are 2013 (scope narrowed), 2016 (suspended), 2021 (discontinued). No discrete 2014 decision found — treat "2014" as a rounding of the 2013 SRM.

The GAO finding

GAO confirmed (quoting NRC's own conclusion) that the framework would have to be developed:

"Based on a preliminary assessment, NRC has concluded that changes in regulations and associated regulatory guidance would be necessary to support an efficient and effective licensing review of commercial GNEP facilities."

GAO-08-483, "Global Nuclear Energy Partnership: DOE Should Reassess Its Approach...," May 22, 2008 — gao.gov/assets/a274666.html. Skeptic's note: this is the strongest single GAO source on the framework gap, but it is framed around the GNEP-era recycling facilities. The cleanest modern synthesis is CRS R48364 (below); congress.gov returned 403 on direct fetch, so its exact contents are corroborated via cross-citation, not a primary read.

2 · How many YEARS would first-of-a-kind licensing realistically take?

No single government sentence gives a tidy total, but the components stack to ~15–25+ years decision-to-operation. The load-bearing data point is the NRC's own forward look:

"NRC... did not anticipate reviewing a license application for a reprocessing facility within 10-20 years."

CRS R48364, "Considerations for Reprocessing of Spent Nuclear Fuel," Jan 23, 2025, p.10 (citing NRC's 2021 FR notice, 86 FR 40764) — congress.gov/...R48364.2.pdf. That 10–20 years is time until a review even begins, not the review duration.

DOE goal "select by 2020, demonstrate by 2050" and "10 to 15 years to design and construct" supporting research facilities: GAO-12-70, "Nuclear Fuel Cycle Options," Oct 2011, pp.12–13 — gao.gov/assets/gao-12-70.pdf. "Many decades to more than a century of continuous recycling": NASEM, "Merits and Viability of Different Nuclear Fuel Cycles...," 2023, Ch.4 — nationalacademies.org/read/26500/chapter/7.

Skeptic flag — the aspirational counter-number: the House Appropriations Committee called for "commercialization of reprocessing by 2033" (H.Rept. 118-580, via CRS R48364 summary). That ~2033 target (~7 yrs out) is a congressional political goal, sharply more optimistic than every agency/expert horizon — not an engineering estimate.

Note: the National Academies "Going the Distance?" (2006) report is about safe transport of spent fuel, NOT reprocessing timelines — flagged so it isn't miscited.

3 · Precedent — built plants, almost none licensed to operate

The US has had exactly one commercial reprocessing plant that ever operated. Two others were built but never licensed to operate.

CRS RS22542 (via EveryCRSReport) — everycrsreport.com/reports/RS22542.html · GAO EMD-78-97 "Evaluation of Federal Support of the Barnwell Reprocessing Plant" — gao.gov/products/emd-78-97 · UCS, "A Brief History of Reprocessing and Cleanup in West Valley, NY" — ucs.org · NYSERDA Fuel Reprocessing History — nyserda.ny.gov · World Nuclear Association, US Nuclear Fuel Cycle — world-nuclear.org

Skeptic's correction: "Carter banned reprocessing" is shorthand. The precise mechanism was an indefinite executive deferral + withdrawal of federal funding + NRC termination of licensing — not a statute. Reagan lifted it in 1981 and Barnwell still never opened, proving economics killed it as much as policy.

Has anyone recently FILED an application?

NO As of mid-2026, no commercial reprocessing facility is operating or licensed, and no company has filed an actual NRC license application. Only pre-application engagement / letters of intent exist (see §4).

4 · Startup CLAIMS vs. regulatory REALITY — be skeptical

Every startup is at "pre-application," "letter of intent," "white paper," or "MOU" stage. None has filed. The pattern repeats history: AREVA/GE-Hitachi a decade ago promised NRC rules by 2015 and construction by 2020 — none happened (IPFM Blog, Jul 2021 — fissilematerials.org).

ANS Nuclear Newswire: "Curio to begin early talks with NRC..." May 29, 2026 — ans.org/...article-8078; "US nuclear fuel recycling takes two steps forward" Sep 8, 2025 — ans.org/...article-7348; "NSDA approved for Oklo's Aurora Fuel Fabrication Facility" Nov 13, 2025 — ans.org/...article-7538 · NRC Press 22-002, Jan 6, 2022 — nrc.gov/...22-002.pdf · World Nuclear News, "NRC denies Oklo licence application" Jan 7, 2022 — world-nuclear-news.org · IPFM Blog Jul 2021 — fissilematerials.org

Independent expert skepticism

• Paul Dickman (former DOE/NRC): "There's no economic benefit to reprocessing at this time." [MIT Technology Review, Mar 19, 2026 — technologyreview.com]
• Allison Macfarlane (former NRC Chair): the loop doesn't truly close — "fuel could be used twice, not infinitely"; spent MOX worsens repository heat constraints. [same MIT Tech Review]
• Edwin Lyman (UCS): "no matter how good your recycling process is, you're still going to need a geological repository in the end." [MIT Tech Review]
• Frank von Hippel (Princeton): reprocessing has run ~10× the cost of dry-cask storage. [IPFM]
• Chris Keefer (Decouple, Apr 9, 2026): when Reagan lifted Carter's ban in 1981, "the commercial response was revealing: nothing happened"; pyroprocessing demonstrated only "in batches measured in kilograms, not the hundreds of tonnes per year" needed. [decouple.media]

5 · Construction time + cost for a first-of-a-kind facility

Four independent estimate lines converge on ~$20–30 billion to build / 1–2 decades to construct; lifetime cost (build + ops + decommissioning) runs to $100B+.

IPFM Blog, Aug 29, 2024 (Rokkasho 27th delay) — fissilematerials.org · WNA Japan Fuel Cycle — world-nuclear.org · Wikipedia THORP / Sellafield (figures cross-checked) · GOV.UK "New era at Sellafield as Thorp reprocessing ends" — gov.uk · GRS "La Hague... until at least 2100" Mar 12, 2024 — grs.de · WNA USA Fuel Cycle (AREVA $25B / 12–15 yr) — world-nuclear.org · UCS "Nuclear Reprocessing: Dangerous, Dirty, and Expensive" — ucs.org

The economics verdict (reprocessing ≫ direct disposal)

Harvard/MIT (Bunn, Fetter, Holdren, van der Zwaan, 2005): the once-through fuel cycle stays "significantly cheaper than recycling... for at least the next 50 yr." Break-even requires uranium at ~$370/kg (≈7× the ~$50/kg reference) — a level not seen since. UCS quantifies the reprocessing premium at 0.4–0.6¢/kWh, ~$3–4.5B/yr extra for the US fleet.

Belfer Center (Harvard), "The Economics of Reprocessing vs. Direct Disposal," 2005 — belfercenter.org · full report (OSTI) — osti.gov/servlets/purl/822658. Caveat: the break-even is 2005-vintage; direction holds, exact numbers would shift on a modern re-run. UCS is an advocacy org (anti-reprocessing) — directionally reliable, methodology standard.

6 · Confidence & caveats (skeptic's ledger)

• Verbatim primary-source (high confidence): the 23 gaps + "gaps remain" (NRC 2026 rule); Part 50/70 are the vehicles; the 2008–2021 effort + 2021 discontinuation; NRC's "10–20 years before review"; West Valley/Morris/Barnwell histories; Carter Apr 7 1977 deferral; Oklo's Jan 2022 NRC denial; Rokkasho 27 delays; THORP £1.8B/1997–2018; La Hague €20–30B new-plant quote.
• Component-stacked (well-supported, not a single quote): the "~15–25+ year total" and "$20–30B build" — assembled from converging GAO/NASEM/NRC/WNA/UCS/Orano figures.
• Inferred / unverified: "~3–5 yr review duration" (no reprocessing application has ever been filed under modern rules, so this is extrapolated from fuel-fab/enrichment analogs); the "~5–7 yr focused framework" figure (directional, not verbatim).
• Corrections made: "2014 deferral" → actually 2013/2016/2021; "Carter banned reprocessing" → indefinite executive deferral + funding withdrawal, not a statute; £121B is the whole Sellafield site, not THORP; "Going the Distance?" (2006) is about transport, not timelines; Deep Isolation is disposal, not reprocessing.
• Could not open directly (corroborated via mirrors/cross-cite): CRS R48364 (congress.gov 403 — but its narrative is consistent across sources), several NRC ADAMS PDFs (timeouts), the 2022 CNBC Curio article (403 — the $5B/2035 figures are flagged single-source).
• The live wildcard: the June 24, 2026 NRC proposed rule could compress the framework timeline vs. the 2007–2021 record — but it's a proposal (comments due Aug 10, 2026), addresses only "some" gaps, and no finalization date is public. It does not change the bottom line that no application has been filed and no complete framework is final.