NNSA never checked the $1.2 billion-a-year nuclear workforce estimate

NNSA⧉ -- the semi-autonomous Energy Department agency that maintains and modernizes the U.S. nuclear weapons stockpile -- released an internal 20-year plan in July 2024 identifying 46 science, technology, and engineering (ST&E) facility investments its laboratories need between 2025 and 2045.▸ When the agency's own Integrated Project Team asked its management-and-operating (M&O) contractors -- the companies that run NNSA's national laboratories, like Los Alamos and Sandia -- what it would cost to staff those facilities, the contractors came back with a number NNSA didn't expect: $1.2 billion a year, every year, for 20 years, just for workforce and program costs beyond what NNSA had already budgeted. GAO⧉ -- Congress's independent auditor, whose recommendations agencies are expected to act on -- found that NNSA never had anyone review whether that number was right. It shelved the estimate instead, and left the 46-facility plan itself with public cost figures for only 7 of the projects on it.

NNSA's Integrated ST&E Plan covers the specialized labs, test beds, and accelerators the nuclear security enterprise needs to keep the stockpile "safe, secure, and effective" without live nuclear testing -- work like compressing plutonium with high explosives to study weapon physics just short of a nuclear detonation.▸ Of the 46 facility investments in the plan, 35 sit at NNSA's three national laboratories and the Nevada National Security Site: 11 at Los Alamos, 11 at Sandia, nine at Lawrence Livermore, and four at the Nevada site. NNSA rated 33 of the 46 as high mission importance and split them by urgency -- 10 needed within five years, 18 within six to 15 years, and 18 not for more than 15 years.

But a facility named in a 20-year plan isn't a facility with a budget.▸ GAO⧉ found NNSA has detailed public cost figures for only 7 of the 46 investments -- the ones already far enough along to appear as active line-item projects in the Department of Energy's fiscal year 2026 budget request⧉. Three of those seven -- upgrades to the U1a underground complex, a new Advanced Sources and Detectors facility, and the Zeus test bed -- form a cluster at the Nevada National Security Site that alone is projected to cost roughly $3.3 billion through fiscal year 2033. The other 39 investments in the plan have cost information NNSA itself calls "limited or non-existent," either because they're still in early planning or because they're sustainment work buried inside broader subprogram budgets.

The bigger gap isn't facilities -- it's people.▸ NNSA's tasking memo for the Integrated ST&E Plan directed the Integrated Project Team to also identify ST&E workforce and "programmatic" investments (things like equipment and experiment materials) needed beyond what NNSA had already planned to spend -- what the agency calls an "uplift." The M&O contractors at the three laboratories did the math and reported back: $1.2 billion a year in workforce uplift alone, for the full 20-year plan period. By NNSA's own preliminary analysis, that single figure exceeded what the agency had anticipated spending -- cumulatively, across workforce, programmatic, and facility investments, at every priority level combined.

That's the mechanism of the problem: a modernization plan can name every facility it wants and still fail if nobody can staff them, and right now nobody at NNSA can say what staffing them actually costs.▸ GAO also found the contractors weren't even estimating the same thing -- representatives at Sandia and Livermore said "programmatic" investments could include staff, so the three labs may have drawn the line between workforce and program costs differently when they answered NNSA's request. Rather than resolve the discrepancy, NNSA's Integrated Project Team decided the number needed more analysis, never sent it to program offices for review, and left it out of the published plan entirely.

As GAO put it in its Conclusions, NNSA officials "did not fully assess the workforce and programmatic investments" and, as of the report, "have no current plans to do so."▸ At $1.2 billion a year, the shelved estimate alone runs to roughly 28% of the entire $4.2 billion NNSA is requesting from Congress for its Stockpile Research, Technology & Engineering account for fiscal year 2026 -- and that $4.2 billion already funds the ST&E workforce NNSA currently employs; the $1.2 billion is what the labs said they'd need on top of it, every year, for the life of the plan.

NNSA's own Planning, Programming, Budgeting, and Evaluation process is supposed to treat long-range planning as continuous -- a living document that adjusts as circumstances change, not a one-time snapshot.▸ GAO found the Integrated ST&E Plan hasn't worked that way. Priorities have already shifted since the plan was finalized in July 2024: congressional funding decisions, delays to ongoing projects, and new administration initiatives in areas like artificial intelligence have all moved the ground under it, and NNSA has no process to regularly update the plan to reflect any of it. GAO's two recommendations target both failures directly: that NNSA's Deputy Administrator for Defense Programs ensure completion of a comprehensive workforce-and-programmatic-investment analysis, and separately establish a process to regularly update the facility-investment plan as priorities change.

In its written response, NNSA agreed to act on both.▸ The agency told GAO it plans to complete the comprehensive workforce and programmatic analysis -- the one that produced the shelved $1.2-billion-a-year estimate -- by December 2027, and to have a process for regularly updating the Integrated ST&E Plan built by December 2026. Both recommendations remain open as of this report; GAO tracks agencies against dates like these until it confirms the action was actually taken, and nothing besides NNSA's own written commitment obligates the agency to hit either one.