NNSA's weapons-production rebuild runs 30 years behind schedule

The National Nuclear Security Administration is building 28 major construction projects, each costing $100 million or more, to modernize the infrastructure that produces the country's nuclear weapons.▸ Together they're estimated to cost more than $30 billion, according to a February 2026 GAO review⧉. Of the 16 projects already under construction, the cumulative cost overrun above their original approved baselines grew from $2.1 billion in 2023 to $4.8 billion by June 2025; the cumulative schedule delay grew from 9 years to 30 years over the same span. Two projects -- both part of the Uranium Processing Facility at the Y-12 National Security Complex in Tennessee -- account for almost 80 percent of that overrun. A September 2025 GAO report⧉ examined that facility in detail: re-baselined in December 2024 at $10.35 billion, it will now take 16 years, not 8, to go from construction start to full operations.

GAO reviews NNSA's major projects on a biennial basis under a standing congressional mandate.▸ Its 2023 report found a $2.1 billion cumulative cost overrun and a 9-year cumulative schedule delay across projects under construction. Two years later, as of June 2025, those figures stood at $4.8 billion and 30 years. The number of projects in the execution phase fell from 18 to 16 over that period -- five reached completion, and NNSA approved baselines for three new ones -- but the projects that remained, or newly entered, construction did not perform better.

Of the 10 execution-phase projects reporting overruns as of June 2025, nine had incurred or expect to incur cost growth of more than 20 percent above their original baselines.▸ NNSA attributes the overruns to inadequate project management by its management-and-operating contractors, poor performance by vendors and subcontractors those contractors oversee, and rising costs of equipment and materials. Two projects -- the UPF Main Process Building and the UPF Salvage and Accountability Building, both part of the Uranium Processing Facility -- account for $3.8 billion of the portfolio's $4.8 billion overrun and a combined 12 years of schedule delay. The other 14 execution-phase projects account for the remaining $1.0 billion and 18 years.

NNSA began planning the Uranium Processing Facility in 2004, to replace Building 9212 -- a uranium-processing building at Y-12 built in 1945, before modern nuclear safety codes existed.▸ The facility's preliminary cost estimate has grown at every subsequent review: about $1.1 billion in 2004, a range of $1.4 billion to $3.5 billion by 2007, a range of $4.2 billion to $6.5 billion approved in 2012 alongside a decision to defer part of the project's scope, and $6.5 billion with a 2026 full-operations target as of March 2018. In December 2024, DOE approved a re-baseline at $10.35 billion, with construction complete in January 2032 and full operations beginning in 2034 -- $3.85 billion more than the 2018 estimate, and a project that will now take 16 years, instead of 8, from the start of construction to full operations.

GAO has designated NNSA's program and project management a "high risk" area for waste, fraud, abuse, and mismanagement for years.▸ The pattern is not confined to the Uranium Processing Facility. NNSA's fiscal year 2026 budget justification shows the Savannah River Plutonium Processing Facility's Main Process Building -- still in the earlier, unbaselined design phase -- may cost more than $22 billion, up from a preliminary range of $6.9 billion to $11.1 billion. The Lithium Processing Facility Main Processing Facility, also still in design, may cost more than $6 billion against a preliminary range of $871 million to $1.5 billion, with about a 6-year delay.

The Uranium Processing Facility's delay has a direct cost of its own.▸ Building 9212 was originally supposed to cease operations once the new facility opened in 2026. It will now run for at least a year past the new facility's expected full operations in 2034 -- and NNSA and its Y-12 contractor have identified about $463 million needed from 2026 through 2035 to keep it running safely: over $308 million for maintenance and operations, $100 million for deferred maintenance, $38 million for personal protective gear, and $16 million for recapitalization work like roofing and HVAC replacement. The protective gear is required because the 1945 building predates modern nuclear safety codes; workers in the new facility, which uses modernized equipment such as gloveboxes, won't need it.

Y-12's deferred-maintenance backlog, separate from routine upkeep, has grown from about $133 million to $219 million since 2023, with $100 million of that attributable to Building 9212 alone.▸ Since 2020, high-priority repairs -- roof work, air-supply fans, emergency lighting, potable water lines, electrical transformers -- have been deferred in part because officials expected the building to close by 2026. Roofing degradation has caused leaks; piping, floors, and fire-barrier walls and doors are flagged as high risk to disrupting operations. NNSA has comprehensive plans for keeping three other aging uranium-processing buildings safely running -- buildings now 70 to 83 years old -- through the 2040s. It does not yet have one for Building 9212, which must bridge a gap of at least 10 to 15 years until a replacement exists.

GAO recommended NNSA's Administrator direct the Office of Defense Programs to establish a comprehensive plan, with relevant stakeholders, for keeping Building 9212 safely operating through 2035.▸ NNSA concurred and has directed its contractor to evaluate the building's site, infrastructure, and equipment, with findings due by March 2026. More broadly, GAO's 2026 portfolio report found NNSA has not fully addressed 8 of 21 open recommendations from prior reviews, including a recommendation originally made in 2020 to develop an integrated master schedule for plutonium pit production, which NNSA agreed with but had not yet acted on as of September 2025. On the broader portfolio review, NNSA provided only technical comments to GAO's draft, not substantive comments on the findings.