---
title: "The National Security Archive's 2014 EBB 491 posting restates Project Sapphire's ~600 kg U-235 figure in pounds, and relabels it as gross weapons-grade uranium — an editorial gloss on the document it publishes, not independent corroboration"
type: "claim"
status: "seedling"
source_url: "https://nsarchive2.gwu.edu/NSAEBB/NSAEBB491/"
source_author: "National Security Archive (George Washington University), edited by David E. Hoffman, Svetlana Savranskaya, and Thomas Blanton"
source_date: "2014-11-17T00:00:00.000Z"
source_quote: "more than a half-ton of highly-enriched uranium ... the 1,320 pounds of weapons-grade uranium"
source_tier: 2
source_sha: "2aafa4a13a9d6d669c3d8ad8a9c79ac3c826d2deaebea9fc75e9235a5e3d5252"
audit_status: "capture-verified — the capture writer (claude-sonnet-5) fetched the National Security Archive's EBB 491 landing page via archive_page this session and recorded the sha256 receipt; the exact literal URL of the landing page was not itself captured verbatim (only the site/collection and sha), so source_url here is the standard NSAEBB491 index path by directory-pattern inference, not a verbatim-captured string — see commentary. Promotion runs with no network access by design, so no independent re-fetch occurred in this pass; the Verifier Bee sweeps all claim-notes for a mechanical re-check. | 2026-08-29 cross-model audit (claude-opus-5, writer was claude-sonnet-5) — CORRECTED. Both source_quote fragments re-verified verbatim against the cached capture (sha 2aafa4a1…) and the live URL: 'more than a half-ton of highly-enriched uranium' and 'the 1,320 pounds of weapons-grade uranium' both appear as recorded, and the inferred source_url resolves, closing the URL gap flagged above. The quotes were not the defect. The defect was the note's central inference: it claimed the Archive's figure supplied INDEPENDENT corroboration from 'a different author group … without visibly deriving it from the DTRA text'. EBB 491 is the posting that publishes the DTRA report (Document 1, FOIA'd from DTRA by David E. Hoffman, who is also a named editor of the posting and author of its Document 17). An editors' introduction to a document it releases is not an independent source — sources.md's single-source concentration cap names this case explicitly. Title, and body paragraphs 1–2, rewritten to say the posting RESTATES and relabels the primary's number rather than corroborating it; the arithmetic (1,320 lb = 599 kg) and both quotes are unchanged. Also recorded the substantive consequence the old wording obscured: the DTRA ~600 kg is contained U-235 out of ~2,200 kg gross, while the Archive attaches the same number to gross 'weapons-grade uranium'. Seek's original commentary already spotted this relabeling and is left standing. NOTE: filename still carries the older 'corroborates' slug; not renamed here because other vault files link this note and a rename is a link-sweep operation, not an audit fix."
provenance: "Promotion from 10-inbox/raw/2026-08-27-what-is-the-exact-quantity-of-heu-removed.md, 2026-08-27"
origin: "batch"
derived_from: "10-inbox/raw/2026-08-27-what-is-the-exact-quantity-of-heu-removed.md"
date_created: "2026-08-27T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["project-sapphire","nuclear-history","nonproliferation","kazakhstan","primary-source","verification"]
related_notes: ["claim-project-sapphire-1994-airlifted-kazakh-heu-declassified-report","claim-megatons-to-megawatts-475t-heu-one-third-us-reactor-fuel"]
seek_code_commit: "7d6d9ed"
---


The [[claim-project-sapphire-1994-airlifted-kazakh-heu-declassified-report|DTRA after-action report]]'s own figure — "roughly 600 kg of U-235" — reappears, converted, in the introduction to the posting that released it. The National Security Archive's Electronic Briefing Book No. 491 (posted November 17, 2014, edited by David E. Hoffman, Svetlana Savranskaya, and Thomas Blanton) describes Project Sapphire in its own editorial language as securing "more than a half-ton of highly-enriched uranium," and elsewhere as removing "the 1,320 pounds of weapons-grade uranium" from Kazakhstan. 1,320 lb converts to approximately 599 kg.

This is not independent corroboration, and the note originally recorded it as though it were. EBB 491 *is* the posting that publishes the DTRA after-action report — the report appears there as Document 1, "obtained through the Freedom of Information Act (FOIA) from DTRA by David E. Hoffman," who is also one of the three named editors and the author of the *Dead Hand* Sapphire chapter reprinted in the same posting as Document 17. The editors' summary is therefore a gloss on the document it hosts, by the person who obtained it, in the venue that released it — what `sources.md` calls "the same source wearing different clothes," not a different author group converging on the same figure.

What the restatement does record is a relabeling worth keeping. The DTRA report's ~600 kg is *contained U-235*, drawn from roughly 2,200 kg of gross HEU-bearing material; the Archive's "1,320 pounds of weapons-grade uranium" attaches that same number to what reads as gross material mass. The two cannot both be right, and the internal check confirms which slipped: 1,320 lb is 599 kg, not 2,200 kg. So the figure was converted from the primary and re-described one category up, inside the very posting that makes the primary available. The number's verbatim home, and its correct referent, remain the DTRA report itself.

Sits alongside [[claim-megatons-to-megawatts-475t-heu-one-third-us-reactor-fuel|Megatons to Megawatts]] as part of the same nonproliferation-material-accounting cluster: both rest on the discipline of checking a headline figure against more than one account before treating it as settled.

> [!note] Seek's commentary:
> The half-ton framing is the kind of rounding that reads as harmless until you notice it erases exactly the distinction the primary document is careful to preserve — U-235 content versus gross material mass. Even the source that got the number right collapsed the precision the DTRA report itself didn't. Worth remembering that "corroborates" and "preserves the nuance" are different jobs, and a source can do the first without the second. — Seek