By Robbin Laird
When I sat down with then Chief of the Australian Navy, Vice Admiral Tim Barrett, in 2017, one of the quieter points he made has aged into one of the more prescient ones. Amid the launch of the largest Royal Australian Navy recapitalization since the Second World War, Barrett was already pushing his service away from a habit that has plagued navies for decades: designing a warship for the threat of the day and then rebuilding it, ship by ship, every time the threat changes.
His navy, he told me, was moving toward ships built from the outset to take on new payloads rather than be reconstructed around them. That is a modest-sounding idea. It is also one of the hardest things in shipbuilding to get right, and one of the easiest to get wrong.
The Bolt-On Failure and the Built-In Success
The U.S. Navy’s Littoral Combat Ship remains the cautionary tale. The LCS was sold on the promise that a single hull could swap between mine countermeasures, anti-submarine, and surface warfare missions by exchanging containerized modules in port. In practice, the Government Accountability Office found the Navy fell behind schedule developing the mission modules themselves and never built a comprehensive plan to close the resulting capability gaps. By 2016, the Navy had quietly abandoned the swapping concept altogether, assigning each hull a single permanent mission instead. Commentary since has been blunt: modularity, as practiced on LCS, was an idea that did not survive contact with the fleet.
The lesson usually drawn from that failure is that modularity itself does not work. That is the wrong lesson. The Royal Danish Navy has been running a version of the same idea since the 1980s, when its STANFLEX system let small patrol vessels and later the Absalon-class frigates swap standardized mission containers to punch above the weight a six-million-person country could otherwise afford. The concept has since matured into SH Defence’s Cube system, now offering several hundred qualified payloads, from torpedoes to unmanned systems to mine-laying packages, that can be loaded through side, stern, or top openings using a skidding system that keeps working in rough seas. SH Defence’s sales and marketing manager, Peter Liisberg, has described the point of the system directly: capability that is not fixed to a ship’s superstructure can keep pace with a threat environment that no navy can predict a decade out.
The difference between the American failure and the Danish success is not the presence or absence of modularity. It is where in the design process modularity gets decided. LCS treated modularity as a mission-generation feature bolted onto a hull that was otherwise built like any other warship, and the swapping infrastructure, the trained crews, and the logistics tail needed to make it real were never fully funded or matured. STANFLEX and the Cube system that grew from it treat modularity as a structural decision made before the keel is laid, embedded in standardized interfaces, frames, and handling systems that the rest of the ship is designed around.
Barrett’s Instinct, Playing Out in Steel
This is precisely the distinction Barrett was drawing in 2017, before most of the language above existed to describe it. He was not arguing for a rare, occasional refit cycle done more cleverly. He was arguing that a ship kept off patrol for a bespoke systems upgrade was a ship failing at the one job that mattered to him: being available. Designing in modularity from the start, rather than retrofitting it after the fact, was his answer to that problem for the submarine and future frigate programs then on his drawing board.
Japan’s Mogami-class frigate, and the enlarged New FFM now in construction, offers a cleaner test of that instinct than most, because the mothership logic was designed in from the keel rather than argued for after the fact. Tokyo’s shipbuilders describe the class in almost exactly Barrett’s terms: a major surface combatant conceived from the outset to host, launch, and recover autonomous systems as part of its core mission set rather than as an add-on.
The clearest case is mine warfare, where the Mogami’s OZZ-5 autonomous underwater vehicle searches for and classifies mines using synthetic aperture sonar and returns to the ship to offload its data, letting the crew build a minefield picture without sending the manned hull into it. The same forward-looking posture shows up in how the class has absorbed its own combat-system growth: the first six hulls were commissioned “fitted for but not with” a 16-cell Mk 41 vertical launch system that later ships received as built, and the follow-on New FFM design, stretched and widened for the purpose, doubles that capacity to 32 cells rather than reopening the original hull to fit it.
Australia has effectively adopted this logic as policy: in August 2025, Canberra selected the New FFM to replace its Anzac-class frigates, the same navy that inherited Barrett’s original bet on modularity now betting on a design built around the mothership concept rather than the bolt-on module.
The Danish Admiral Nils Wang has pushed the logic a step further, arguing in discussions I have had access to around the Cube system that sustained modularity eventually erodes the usefulness of legacy platform labels altogether. A ship built to swap propulsion modules, weapons modules, and sensor modules over its lifetime is not really a frigate or a corvette in the traditional sense; it is closer to what Wang calls a mother ship, defined less by its own fixed loadout than by the range of capability it can host and enable across a battlespace. That reframing matters because it changes what a navy is actually buying when it commissions a hull: not a fixed set of weapons and sensors locked in for thirty years, but the standardized frames, power, data architecture, and handling systems that let tomorrow’s payload replace today’s without touching the ship’s structure.
Why This Matters More Now Than It Did in 2017
Barrett made his case before the maturing autonomous systems mesh fleet gave modularity a second, larger job to do. It is no longer only about shortening the time a single manned hull spends offline for a combat-system upgrade. It is about whether that hull’s standardized interfaces can also host, launch, and recover the unmanned surface and underwater systems that increasingly carry payload the manned ship no longer needs to carry itself. A ship designed for bolt-on modularity, LCS-style, is poorly positioned for that second job.
A ship designed for structural modularity, STANFLEX-and-Cube-style, is already most of the way there, because the standardized frames and handling systems built to swap a mine-laying container for a decoy launcher are the same infrastructure needed to swap in an unmanned systems control module or a launch-and-recovery skid.
Barrett’s closing argument to me in 2017 was that none of this could be achieved without a national commitment treating shipbuilding as an industrial and national undertaking rather than a single-service acquisition program. The record since, from Denmark’s decades-long investment in a standardized payload ecosystem to the United States’ costly detour through LCS, suggests he had the causality right.
Modularity that is legislated, funded, and engineered in from the first design review tends to work.
Modularity announced as a feature and left to the acquisition system to sort out later tends not to.
The fleet you build today is, in that sense, still fighting the fleet you have.
Fighting with the Fleet You Have and Building Availability into the Autonomous Systems Age
Attacking the Drydock Problem: Robotics and the Compression of Repair Time