The Navy is paying $18 billion to fix a submarine problem money alone can't solve

The submarine fleet is a two-sided machine: one set of shipyards builds new boats, another keeps the existing ones running, and both depend on the same scarce thing — trained welders, pipefitters, and machinists.▸ For two decades that machine has run short of money. Now it is running short of people, and the Government Accountability Office's February 2026 testimony and the Congressional Budget Office's April 2026 testimony to Congress both say the labor shortage, not the budget, is now the binding constraint.

The Navy's plan, laid out in CBO's April 2026 testimony on the shipbuilding industrial base, is to spend about $18 billion rebuilding the submarine supply chain — split roughly $10 billion for new-construction capacity (Virginia- and Columbia-class production) and $8 billion for maintenance and retention.▸ Almost all of the new-construction money is aimed at people, not steel: "recruitment, training, and retention of shipyard workers," per the testimony, alongside infrastructure, supplier development, and paying other yards to take on overflow work.

It is not the first check the Pentagon has written for this problem.▸ The Department of Defense had already put nearly $6 billion into the shipbuilding industrial base by the end of fiscal year 2023, per GAO's February 2026 testimony — and GAO found that, despite that spending, none of the seven shipbuilders that build Navy battle-force ships were positioned to meet the Navy's own delivery goals. Money bought equipment and buildings faster than it bought people. Separately, across the Navy's most recent lead ships — the first-of-class vessels that absorb the steepest design risk — the last 11 have together cost at least $8 billion more than planned, GAO found, a different $8 billion than the maintenance figure above, and a reminder that overruns keep compounding even as new investment goes in.

The clearest read on the labor problem isn't a headcount — it's the boats sitting idle.▸ Virginia-class attack submarines were being built at one per year as of June 2025, per GAO, half the Navy's stated goal of two per year; the two Virginia boats the Navy did accept in 2025 were each more than three years late. The lead Columbia-class ballistic-missile submarine, SSBN 826, is now projected at least 18 months behind its contract delivery date, and GAO's own shipyard survey found that "workforce challenges at Navy shipbuilders persist" for welders, pipefitters, and machinists in particular.

Maintenance tells the same story from the other end.▸ The nation's four public shipyards handle nearly all upkeep on the Navy's nuclear submarines and carriers, and the Navy's goal is to keep no more than 20% of its attack- submarine (SSN) fleet tied up in or awaiting maintenance at any time. Per CBO's April 2026 testimony, the actual figure today is 33% — the worst share since 2008 outside of 2021 — which means that out of 48 SSNs in the fleet, the Navy can deploy at best 32 of them. If the Navy hits its own 20% maintenance target by 2035, when it projects a 54-boat fleet, deployable submarines would rise to 43.

There are early signs the labor side is turning.▸ Voluntary attrition at Bath Iron Works and Electric Boat has fallen to under 10% a year, and Ingalls Shipbuilding's rate has dropped to around 15%, per CBO — before a new labor agreement raises wages there 18% immediately and roughly 40–45% over five years. Whether that shows up as more submarines on time is a multi-year question; it hasn't yet.

Figures cover Navy submarine programs specifically; parallel Coast Guard cutter delays (Polar Security Cutter, National Security Cutter) and Navy surface-ship programs (Constellation-class frigate, DDG-51 destroyers) cited in the same GAO and CBO testimony are not included in the totals above.