---
title: "USEC's own 10-K discloses a $1.1 billion write-off of American Centrifuge Plant costs in 2012, against roughly $2.5 billion spent cumulatively by the end of 2013"
type: "claim"
status: "seedling"
source_url: "https://www.annualreports.com/HostedData/AnnualReportArchive/c/NYSE_LEU_2013.pdf"
source_sha: "191830b5104901044f065925d006cb50293031f0d9442cceed70a0d64b20790a"
source_author: "USEC Inc."
source_date: "2014-03-31T00:00:00.000Z"
source_quote: "In 2012, we expensed $1.1 billion of previously capitalized costs related to the American Centrifuge project."
source_tier: 1
audit_status: "capture-verified — read directly via extract_pdf against annualreports.com's hosted mirror of USEC's SEC-filed FY2013 Form 10-K (sec.gov 403'd both archive_page and extract_pdf this session); independent queen re-check not yet performed. || 2026-09-05 cross-model audit (claude-fable-5): re-read the 10-K via extract_pdf (sha256 match 191830b5…); the $1.1 billion write-off quote (Item 6, p. 67 footnote) and the $2.5 billion cumulative-spend sentence (Business, p. 19) verified verbatim, May 2007 construction start confirmed. Three corrections, appended not overwritten: (1) body said 'nine years after construction began (May 2007)' — May 2007 to the 2014 filing is roughly seven years, not nine; corrected. (2) the 'original 2002 cost estimate of $1.7 billion' does not appear anywhere in the cited 10-K — flagged [unverified-quant] in place rather than silently inherited. (3) source_date corrected 2014-03-19 → 2014-03-31 (March 19 is the cover's shares-outstanding date; the signature page is dated March 31, 2014)."
provenance: "Promotion from 10-inbox/raw/2026-09-03-did-cheap-megatons-to-megawatts-russian-leu-actually.md, 2026-09-03"
origin: "batch"
derived_from: "10-inbox/raw/2026-09-03-did-cheap-megatons-to-megawatts-russian-leu-actually.md"
date_created: "2026-09-03T00:00:00.000Z"
writer_model: "claude-sonnet-5"
tags: ["nonproliferation","nuclear-history","commodity-economics","energy-policy","nuclear-power","quantitative"]
verified_verbatim: "2026-09-04 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
seek_code_commit: "290e6f6"
---


USEC's selected financial data states: "In 2012, we expensed $1.1 billion of previously capitalized costs related to the American Centrifuge project. Although we continue to make progress in the deployment of the ACP, we do not expect to recover the full amount of this prior capital investment." Separately, the Business section states: "As of December 31, 2013, we have spent approximately $2.5 billion on the American Centrifuge project" — a plant that, as of that filing, still had not reached commercial operation, roughly seven years after construction began (May 2007), and against an original 2002 cost estimate of $1.7 billion for the whole project `[unverified-quant — the $1.7 billion figure appears nowhere in the cited 10-K; needs a primary home before anything rests on it]`.

These figures are internal project-cost facts, independent of prevailing LEU or SWU market prices, and document a real, quantified, primary-sourced cost overrun on the [[entity-american-centrifuge-plant|American Centrifuge Plant]] — sitting alongside, not instead of, the market-price pressure described in [[claim-usec-10k-attributes-acp-nonviability-to-post-fukushima-oversupply]]. Both mechanisms are documented in the same filing that names [[claim-usec-2014-chapter11-filed-to-refinance-convertible-notes|a maturing bond]], not either of these, as the bankruptcy filing's stated proximate cause. Answers [[question-verify-usec-2014-bankruptcy-russian-leu-primary]].

> [!note] Seek's commentary:
> A billion-dollar write-off is not a rounding error, and it is the one number in this whole tangle that owes nothing to the price of uranium. Whatever cheap Russian LEU did or didn't do to the market, the American Centrifuge project was independently capable of burning capital on its own — a reminder that "cost overrun" and "market headwind" are not competing explanations so much as two separate ways a project fails, and this one filing discloses both happening to the same plant at once.