The National Nuclear Security Administration, a semi-autonomous agency inside the Department of Energy, manages the nuclear weapons stockpile through life-extension programs that refurbish or replace aging warheads. Those programs appear each year in the NNSA congressional budget justification, the Stockpile Stewardship and Management Plan, and related justifications submitted to the House and Senate Armed Services and Appropriations committees.

The same programs are described in open hearings. Administrators and laboratory directors have, according to the hearing record, told lawmakers that designated life-extension efforts remain executable within approved cost and schedule baselines. The written products tell a more qualified story.

Budget justifications for the weapons activities account typically present total estimated cost, remaining “to-go” cost, and a contingency line. Footnotes and table notes, according to successive fiscal year volumes, flag that estimates exclude some future scope, that contingency is not a reserve against all identified risks, and that independent cost reviews have at times produced higher figures than the program office baseline. Those notes are public. They are not classified annexes.

The Government Accountability Office has, in a series of reports on NNSA acquisition, described a pattern in which early cost estimates for warhead work prove optimistic once production engineering and first-production-unit dates approach. GAO has stated that NNSA has improved some estimating practices but that several life-extension and modernization efforts still carry substantial schedule and cost risk. Inspector General material at Energy has similarly noted weaknesses in how contingency and management reserve are presented.

Congressional testimony, by contrast, often compresses that texture. Officials have said, in paraphrase of public hearing transcripts, that a given life-extension remains “on track,” that first production is “achievable,” or that the program is executing within the current approved baseline. Those formulations are not, on their face, false. They omit the footnotes.

Life-extension work is not a single line item. The record includes the B61-12 gravity bomb life extension, the W88 Alt 370, the W80-4 for the Long-Range Standoff missile, the W87-1 for the Sentinel intercontinental ballistic missile, and planning for the W93. Each has its own cost table, its own production complex dependencies at Pantex, Y-12, Kansas City, Savannah River and the national laboratories, and its own footnotes on what is included.

Public budget documents have shown, over multiple cycles, that “to-go” costs can rise even when officials describe a program as stable, because remaining work is re-priced, because the production complex is shared, or because other programs compete for the same high-explosives, pit, and non-nuclear component capacity. Footnotes sometimes state that figures are in then-year dollars, sometimes that they exclude certain capital projects, and sometimes that an independent cost estimate differs from the program estimate. Testimony rarely walks through those distinctions.

The production complex itself is a source of the discrepancy. NNSA has told Congress that pit production, depleted uranium, and lithium processing must be recapitalized if life-extension schedules are to hold. Those recapitalization lines sit in other budget chapters. A warhead life-extension table can therefore look contained while the enabling infrastructure, according to the same justification, is not yet at required capacity. Officials have acknowledged infrastructure risk in written answers for the record more often than in opening oral statements.

Committees have noticed. Armed Services and Energy and Water appropriators have, in report language, directed NNSA to reconcile program-office estimates with independent cost estimates and to explain contingency. Those directions appear in public committee reports. They do not assert that any particular classified yield or design detail is at issue. They concern money and schedule as presented to Congress.

The agency’s position, as restated in budget transmittal letters, is that life-extension programs are essential to an aging stockpile and that cost estimates will be updated as design definition matures. That is a standard acquisition claim. The footnotes, however, already record that maturation has repeatedly moved numbers. Testimony that treats the current baseline as a settled fact therefore sits beside written material that treats the same baseline as provisional.

None of this requires a leaked document. The comparison is between two public genres: the footnote apparatus of the budget justification and the compressed oral record of a hearing. Named institutions—NNSA, GAO, the Energy inspector general, the nuclear security laboratories, and the congressional defense committees—already publish enough for the contrast to be drawn.

Whether the next round of life-extension testimony will cite the footnotes as prominently as the baselines is a matter for the next hearing. The written record already shows the gap.