Does investment buy pace in UK naval shipbuilding?
The UK has already invested heavily with BAE Systems, Babcock and Navantia UK, although the ships those investments were for are yet to enter service. Speed of build, rather than technology, is what politics now demands.
Anyone who has walked the Clyde in the last couple of years will have seen the Janet Harvey Hall at Govan, standing at 170 metres long and 80 metres wide, and big enough to build two Type 26 frigates side by side under cover. This is the visible part of a £300 million programme to modernise the yards.[1] At Rosyth, Babcock built Venturer Hall for the Type 31 and rolled the first ship out of it in May last year.[2] In Belfast, Navantia UK has put new cranes, robotic plasma cutting and a mechanised panel line into the old Harland & Wolff yard as part of a £77 million commitment, after the yard's collapse and rescue.[3] Three primes, three yards, and a great deal of new capability and capacity representing the most investment UK surface shipbuilding has seen in a generation, and something to celebrate against a backdrop of industrial reinvigoration.
However, none of the ships that those investments were made for are yet in service. As the investment lands and capability strengthens, the debate has now turned away from capability toward pace, but pace cannot be solved by this investment alone.
Where the time went
The first steel for HMS Glasgow was cut in July 2017. She was floated off in December 2022, named in May 2025, and is now at Scotstoun finishing equipment commissioning and combat system integration, with sea trials expected around the turn of the year and initial operating capability for the class in 2028.[4] In 2021 the original expectation was handover around 2024,[5] meaning on paper there is a four-year slip on a first-of-class ship. BAE's own original target was to shorten the build cycle from about 96 months on ship one to about 60 months on ship eight.[6]
The Type 31 was sold on pace, with the 2021 plan for a ship laid down every twelve months, each completed in about 28 months, the first in service in 2027 and the last in early 2030.[5] HMS Venturer's steel was cut in September 2021, she was rolled out in May 2025 and floated off that June, and she is now fitting out at Rosyth with service entry expected late this year or early next, 2-years behind the original intent.[2] The Fleet Solid Support ships were envisaged for 2028 when the contract was being competed. The contract was signed in January 2023, the first ship's steel was cut in December 2025, and the first ship is now expected in 2031.[7]
None of the delivery delays have a singular cause. Glasgow is a first-of-class ship whose design was being matured whilst steel was being cut, and whose gearboxes arrived late enough that the hull had to be opened to fit them.[5] The pandemic took the best part of a year out of every yard in the country. Barrow is building Dreadnought and Astute concurrently, and the two programmes compete for the same facilities and the same qualified people, though the public record does not show how much delay that has cost.[8] The FSS competition was paused and relaunched before anyone cut anything. Challenges and issues tend to compound, and those close to the programmes will understand that there is no singular cause. Underneath any frustration, industry and the customer both recognise that the plan now needs ships to be delivered at a pace that was never in the plan.
The submarine drumbeat
The challenge isn't just contained to the recapitalisation of the Surface Fleet. The Strategic Defence Review of June 2025 committed to up to twelve SSN-AUKUS boats, built at Barrow at a rate of one every eighteen months.[9] The current measurable rate is HMS Agamemnon, whose keel was laid in July 2013, commissioned in September 2025, and she sailed from Barrow for the first time ln 24 September 2026, 13 years and two months after keel-laying.[8] The United States, with two submarine yards and a programme that has run since the 1990s, has never reached its own target of two Virginia-class boats a year; it has delivered between 1.1 and 1.2 a year since 2022 against a target now raised to 2.33 to cover the boats promised to Australia.[10] The Defence Investment Plan published in June says steel will be cut on the first SSN-AUKUS in 2027, before the last Astute is finished, which is the right sequence and an encouraging one,[11] but an 18-month drumbeat is an aspiration. Lord Case, who chairs Team Barrow, told the Defence Committee last December that it is 'a very demanding target'.[12]
Building the record with the ship
The Type 23 life extension replaced the heart of the ship. The Power Generation and Machinery Upgrade programme meant gutting the engine room and rebuilding it around a new power and propulsion system, with everything that 'hangs off' it. We had 3D models of the compartment, had sent people to verify the ship against them, and we had laser-scanned the space before the design was finalised. I recall standing in that engine room after the lagging had come off and the old equipment was out, with the drawings on a toolbox and the design team on the phone, because what had been uncovered did not match what had been scanned. Pipework where the drawing said there was none, a penetration a few inches from where the model put it, a bracket nobody had recorded. In a compartment that small, with that density of machinery, pipework, cabling and exhaust trunking, every one of those discrepancies was a design change.
The cost wasn't just felt in the delays, but was counted in continuity of work. A team stopping, waiting for a query to be answered, moving on to something else in a piecemeal way, and losing the thread. What was needed was a proper record of how the ship had been built. The grade and tolerance of the parent metal, the QA and QC records from fabrication and painting, the dimensional inspection records against design tolerance, the actual routing of every cable and pipe and every hole cut through a bulkhead. That record exists for a ship that is built with through-life support information in mind.
Is the Asian market benchmark fair?
In the heat of debate, everyone reaches for what is being achieved in Asia. However, it is worth being precise about what the comparison means. Japan could be considered as the 'clean' case: JS Nagara, the tenth Mogami-class frigate, was laid down in July 2023, launched in December 2024 and delivered in June 2026, under three years from keel to delivery, from a yard producing two frigates a year and completing twelve hulls in about eight years.[13] What the Japanese record does not immediately show is the learning curve. The time from keel to commissioning across the first nine hulls sits between 29 and 35 months and, if anything, lengthened slightly as the class progressed; the two-a-year output came from building several hulls in parallel on a settled design, not from each ship being faster than the last.[13] Korea is now a dated case too. ROKS Chungnam, a 4,300-tonne frigate, went from keel to commissioning in 32 months, and HD Hyundai delivered a 10,000-tonne Aegis destroyer about three years after construction began, 'on schedule'.[14]
The benchmark can be considered 'real', but whether it is a fair comparison depends on what actually sits against it. A 6,200-tonne general-purpose frigate with a crew of 90, built as the tenth of twelve to a settled design, is not an 8,000-tonne anti-submarine frigate built first of class in a yard that had not built a frigate for a decade. The Japanese and Korean advantage is design maturity and continuous series production as much as it is yard speed, and both sit on a commercial shipbuilding base the UK gave up long ago. The United States has just made the same point. The Constellation frigate, an adapted European design whose changes kept growing, was cut from six ships to two last November after unit cost reached about $1.5 billion and delivery slipped three years, and the US Navy is now looking at the Japanese, Korean and Turkish designs instead.
Yard capacity cannot compensate for an immature design.[15] That is the uncomfortable lesson for the UK, because the Defence Investment Plan's Hybrid Navy rests on five new classes of vessel, four of them uncrewed, none yet designed, built or proven, and asks for them at a pace that one analyst has called one of the fastest periods of naval transformation in modern British history.[11]
Design maturity is not the only part of the picture either. The other significant constraint is people. Barrow employed around 13,000 in shipbuilding in the 1980s, fell to 3,000 in the order gap of the late 1990s, and had rebuilt to about 8,000 by 2019; the apprentices who were not recruited in that gap are the generational hole in the skills base that the yards are still working around.[16] Stop-start demand and short contract cover have thinned the industrial base or lost it altogether, and the current programmes of record are rebuilding it, with automation and technology taking some of the strain. But experience takes time to grow, and it grows slowly in a workforce whose history gives it little confidence that the jobs will last. Shipbuilding isn't alone either, with the nuclear sector looking for 40,000 new people by the end of the decade and half as many again highly skilled workers over ten years. Construction, shipbuilding and nuclear are three industries all drawing on the same pool.[17] The present political leadership is supportive of rebuilding Industrial Britain; the difficulty is that the political cycle now moves a great deal faster than the industrial one.
What the neo-primes really add
All of this brings us to the recent announcements that prompted this series. In September the US Navy placed production agreements for30 medium uncrewed surface vessels with three companies, ten each at about USD$40 million a vessel, following a phase of at-sea testing that each firm had reached after years of its own development and that culminated, in one case, in eight days of continuous operation; the first are due before the end of next year.[18] One of the three winners was HII, the largest traditional shipbuilder in America. And earlier this month Anduril and the US Navy announced Arsenal-2, a USD$6.6 billion yard in Maryland to make complex components and large assemblies for Virginia-class submarines, delivered to Electric Boat and Newport News for final assembly, with initial operations in 2030.[19] In Australia, Anduril's Ghost Shark factory in Sydney completed its first production uncrewed submarine seven weeks after the fleet contract was announced, although that followed several years of prototype development.[20]
Read together, these do not overturn the industrial base, rather they build capacity inside it. Arsenal-2 feeds the existing submarine builders; it does not replace them, and nor does it promise a boat before 2030. It will be interesting to see whether the increased capacity actually relieves the bottlenecks inside the primes, or whether constraints elsewhere mean that fabricated steel simply stacks up at the front of it.
The MUSV awards are for small, simple hulls bought through a marketplace rather than a programme of record, and the traditional prime won alongside the newcomers. The speed of delivery might be real, but the examples are still relatively narrow. A 180-foot vessel from concept to water in under a year, and a first production submarine seven weeks after contract, are both outcomes that come on top of years of quiet development. Nobody has yet shown that pace on a complex vessel, and the speed of delivery will be proven long before the support case for autonomous vessels of this size and beyond has been made, let alone proven. In shipbuilding, the neo-primes are joining the traditional programme, not upending it, and the sensible reading of their arrival is that the customer wants more capacity and more competition in the parts of the system where the incumbents are slowest.
Anduril calls its new yard a software-defined shipyard, and the test I proposed in the last article was not whether the phrase sounds modern but whether it changes how the work is done. On build pace alone, the evidence so far says that what changes the rate is a mature design, built in series, in a yard that never stops, and that the UK has none of those three at the moment on any surface combatant. The software-defined approach to manufacture is not really a claim about speed, more a claim about the flow of information.
Closed-loop feedback from the field into the production line, so that a fault found at sea is corrected in the next hull rather than the next class;
Design change incorporated into production in days, rather than through a change process measured in months or years; and, most of all
A baseline configuration for each ship as it rolls off the line. Not a generic class set of drawings, but the steel batch, the paint applied, the precise build state and the fingerprint of the hardware, software and firmware actually embodied in that hull.
No stack of paper on the toolbox, and nothing taped to the bulkhead. That is compelling, and it is where the software-defined shipyard will make its real difference, not in how fast the ship is built but in what the people who keep it going know about it afterwards. Which is the subject of the next article.
