GAO-26-108457 was released on 2 July 2026. Figures below are from the report’s analytical section unless otherwise noted. Programme cost and schedule data in the report come from Department of Defense reporting, not from independent measurement.
Every account of this year’s assessment led with the same figure: it now takes the Pentagon more than 12 years to deliver a weapon system. The explanation for that figure is in the same document, about forty pages further in, and almost nobody went to get it.
What is there is a workforce that has been cut while the schedule expectations stayed where they were. Thirty-seven of the 48 programmes GAO questioned in depth reported losing staff. One lost 38 percent of its core personnel in twelve months and managed to replace a third of them. And the data GAO used to produce the headline number is, in places, two and a half years old, because the Pentagon stopped publishing the reports it was drawn from.
- Published
- 2 July 2026
- Statutory deadline
- 30 March, per 10 U.S.C. § 3072
- Length
- 249 pages, 24th annual edition
- Programmes assessed
- 104 — 75 MDAP, 23 MTA, 6 future
- Questionnaire subset
- 72 programmes; 48 assessed against leading practices
- Portfolio value
- over $2.4 trillion, excluding sustainment and classified
- Audit period
- April 2025 – July 2026
- Recommendations
- 1 — DoD concurred
The number, and why it understates
The expected cycle time across the major defence acquisition programme portfolio rose past 12 years from programme start to initial operational capability. GAO says plainly that this is a floor. Six programmes are excluded from the average entirely because they have not set a new delivery date, most of them rebaselining after extensive delays. Fifteen more have no published date or a classified one.
The comparison inside the figure is the part worth holding onto. Of the 75 programmes in the portfolio, 43 have already reached initial capability, and they took an average of 132 months, about eleven years. The 17 that have not reached it estimate 149 months. There is almost no gap between what finished programmes actually took and what unfinished programmes are still promising, which means the promises have already absorbed the delay and are unlikely to hold.
Cost moved the same way. The portfolio lost four programmes over the year and gained more than $50 billion. Among the 72 programmes with usable cost data, 46 reported increases totalling $122.08 billion, 16 reported decreases totalling $47.3 billion, and 10 reported no change.
What is on page 53 and after
The report’s section on the acquisition workforce runs from page 53 to 56, and it is the most consequential material in the document that did not make the coverage.
Forty of the 48 programmes assessed against leading practices reported workforce difficulties. Thirty-seven reported reductions in military, civilian or contractor staff. Thirty-three reported trouble hiring or retaining people. Twenty-five said the department’s civilian hiring freeze made it harder to fill posts. Most could not backfill at all.
The individual accounts are worse than the aggregate. One programme office lost 38 percent of its core personnel between December 2024 and December 2025 and replaced one in three. Another lost seven civilian positions and could not backfill any of them under the hiring freeze; officials told GAO the same work is now distributed across fewer people who work weekends to keep the schedule from slipping visibly. A third reported losing both its most experienced staff and its junior staff at once, cutting off, as GAO put it, both ends of the experience spectrum.
Hiring someone from outside government into one of these offices takes between eight months and a year once clearances and familiarisation are counted. Several programmes reported re-advertising the same post repeatedly to find one qualified candidate.
None of this appears in the summary page, and none of it appeared in the trade coverage. It is, however, the most direct available explanation for why schedules that were optimistic in 2024 are less achievable in 2026.
The instruments are degrading too
A footnote in the methodology explains where the numbers came from, and it is not reassuring. The Pentagon did not issue comprehensive Modernized Selected Acquisition Reports for fiscal 2024, because it did not include a Future Years Defense Program in the FY2025 budget request. The comptroller then decided not to publish a Green Book for FY2026. Where current reporting was unavailable or incomplete, GAO fell back on Modernized Selected Acquisition Report data from December 2023.
What this means for anyone citing the figures. Parts of a July 2026 assessment rest on programme data dated December 2023, the most recent available at the time of review. The report is explicit about this, and it does not invalidate the findings. But it does mean the 12-year figure is a measurement taken with instruments the Pentagon has stopped calibrating, and that the direction of any error is unknown rather than conservative.
The report itself was due to Congress by 30 March under statute. It arrived on 2 July.
The fast pathway is where the slow programmes went
The middle tier of acquisition exists to prototype or field a capability within five years. The Pentagon now plans to spend at least $49 billion across the 23 largest programmes using it, and GAO is direct about why that total grew: more of the largest and most time-consuming programmes moved onto the pathway this year. Space Force programmes account for close to half the MTA portfolio’s cost.
Between 2018 and 2025, 18 of 40 programmes entered the pathway with immature technologies, some at technology readiness level 3 or 4, against the level 7 that leading commercial developers require before development starts. Seven of the eight programmes currently on the pathway remain immature. Eleven programmes since 2018 entered without identifying their critical technologies at all; the Army’s XM30 combat vehicle did not identify its own until this year, seven years after designation, and then found several of them immature.
The exit numbers are worse than the entry numbers. Fourteen of the 23 current MTA programmes do not intend to develop a minimum viable product at all. Since 2018, only one of the four programmes that entered as rapid fielding efforts has fully transitioned to production. The Navy’s Hypersonic Air-Launched Offensive Anti-Surface Warfare effort finished its MTA run ahead of schedule without delivering a capability. The B-52 Commercial Engine Replacement Program left the pathway for the conventional one and now expects seven more years before initial capability, or fourteen years from the start of the fast route.
One programme did what the pathway was built for. The F-15EX entered rapid fielding with its critical technologies at level 7 or above and reached full-rate production inside five years. It is the counter-example, and it is a single one.
Three programmes stopped this year
| Programme | Time and money spent | Outcome |
|---|---|---|
| M-10 Booker | Over 7 years in development | Terminated to end procurement of obsolete systems; contract termination costs put at $65 million on top of a 62 percent cost decrease |
| FFG 62 Constellation-class frigate | Over 5 years, following years of design instability | Cut from 20 ships to two in November 2025; $2.0 billion redirected to alternative workload and workforce retention at Fincantieri Marinette Marine |
| Next Generation Operational Control System (OCX) | Over 13 years and more than $7 billion | Cancellation recommended by the Air Force acquisition executive in late 2025; departmental decision still pending as of March 2026 |
These are the visible end state of the pattern the rest of the report describes. GAO’s companion assessment for the House Oversight subcommittee, published three weeks earlier, supplies the most physical illustration of the same failure: nearly 10,000 Integrated Visual Augmentation System headsets built across two versions and heading into storage rather than to soldiers, after eight years and three separate acquisition attempts.
Reforms are being adopted at the edges
GAO assessed 48 programmes against the product-development practices it has been recommending since 2022. The pattern is consistent: the practices that require asking are widely adopted, and the practices that require committing are not.
Forty-five of 48 programmes gather end-user feedback. Thirteen have formal user agreements that would oblige them to keep gathering it. Four have a digital twin meeting GAO’s criteria, the distinguishing requirement being a real-time automated data connection. Four further programmes explicitly plan to omit that connection from what they still call a digital twin.
Programmes were candid about why. Digital engineering demands front-loaded spending on tools, infrastructure and specialised staff inside an already constrained budget. Classified and air-gapped environments make a digital thread genuinely hard to build. And programmes started before December 2023 are not required to do any of it.
One consequence surfaced in the supply chain. Three interconnected programmes drew on the same production line. The Army’s Long Range Hypersonic Weapon, the Navy’s Conventional Prompt Strike and the DDG 1000 destroyer integrating the CPS missile: none had conducted an industrial base assessment before production. The line bottlenecked, LRHW’s second battery slipped at least six months, CPS hit quality problems, and DDG 1000’s integration was delayed behind both.
What to watch
GAO made one recommendation in 249 pages: require programmes on rapid pathways to start with mature technologies, or develop the immature ones separately. The Pentagon concurred. Concurrence is where most of these recommendations have stopped before. GAO has been asking for the underlying policy revamp since 2022, and holds recommendations from January 2025 on modular open systems of which three of 14 have been implemented.
Sources
- US Government Accountability Office — Weapon Systems Annual Assessment: Requiring Mature Technologies Could Enable Shift to Rapid Delivery (GAO-26-108457)
- US Government Accountability Office — Full report, 249 pages (PDF)
- US Government Accountability Office — Weapon Systems Acquisition: Beyond Business as Usual (GAO-26-109135)
- Navy Times — Pentagon’s top new weapons programs are 12 years behind schedule
- Breaking Defense — Pentagon continues to struggle with key weapons development timelines
- USNI News — GAO Weapon Systems Annual Assessment to Congress
