By Robbin Laird
Marcus Jones, writing for CIMSEC on October 2, has made a case that deserves a wider audience than the shipbuilding specialists who will read it first. His argument is that Korean shipbuilding will matter to American naval power only if American institutions change how they buy, schedule, repair, and build. Money and goodwill are not enough.
I agree with nearly all of it. The article stops one step short, though, and that step is where the more interesting strategic question sits.
Jones writes about the crewed fleet: destroyers, auxiliaries, submarines, and the repair of hulls the Navy already owns. That fleet needs the fix he describes.
But the fleet being built around it is increasingly a mesh of autonomous and lower-cost systems, and that fleet is made differently.
The question I want to raise is whether the production disciplines America is about to import from Korea suit that second fleet, or whether we risk solving the old problem with habits that will constrain the new one.
The Test Jones Sets
Jones starts from a hard fact. The Navy’s May 2026 Shipbuilding Plan describes the failure of American shipbuilding as structural. The fleet counts 291 battle force ships against a statutory objective of 355. Virginia-class production runs at roughly one and a quarter boats a year against a stated goal of two. Shipbuilding employment has fallen by tens of thousands since 1980. The plan’s new “Total Naval Vessel Force” of 450 reaches that figure largely by counting auxiliaries and unmanned vessels alongside the 299 battle force ships projected for 2031, and Jones is right to warn against reading that as restoration of the manned fleet.
Against this background he treats South Korea as the first serious test of whether the Navy’s declared reforms can become industrial practice. The Korea-U.S. Shipbuilding Partnership Initiative, the new partnership center that opened in Washington in July, and the $150 billion investment commitment tied to MASGA give the effort institutional form. Hanwha owns Philly Shipyard and plans an expansion that would lift output from fewer than two ships a year toward as many as twenty. HD Hyundai has signed a broad agreement with Huntington Ingalls covering distributed shipbuilding, automation, and auxiliary teaming. Samsung is working with Vigor, with General Dynamics NASSCO, and with Hanwha Ocean.
Jones’s most useful contribution is a distinction between three things that are easily conflated: Korean work performed abroad, Korean capital invested in the United States, and Korean methods adopted by American yards. Work abroad relieves bottlenecks. Capital rebuilds physical capacity and sustains commercial workload. Only the third changes how American yards perform. The first two, he argues, should be judged by how well they enable the third.
I would adopt that test as it stands.
It is rigorous, and it cuts through the announcement-driven character of most allied industrial cooperation. My question is what exactly is being absorbed.
What Korean Discipline Buys
Jones describes the Korean advantage precisely. Korean commercial yards practice front-loaded production engineering, high design maturity, extensive design reuse, controlled change, complete work packages, and protected production sequences. They do not push unfinished work downstream for later correction. The Navy’s own emerging acquisition language now echoes these principles: mature designs before construction, modular fabrication, and sensitive integration concentrated at designated yards.
The early repair cases show both the value and the friction. Hanwha’s overhaul of USNS Wally Schirra was first described as a three-month job. It involved more than 300 work items and extensive hull deterioration, and it required a rudder that Hanwha engineers had to reverse-engineer because usable blueprints did not exist. The ship left after seven months. HD Hyundai’s work on USNS Alan Shepard started at about sixty items and grew by more than a hundred.
Jones reads these as a procedural mismatch and not a Korean failure, and I think that reading is correct. Commercial shipbuilding assumes a mature work package. Naval repair finds out what is wrong with the ship after the ship is in the dock. The lesson is that Korean collaboration earns its keep when it forces the Navy to improve technical data, configuration control, and pre-availability planning. An efficient allied yard that inherits American uncertainty is merely efficient at being uncertain.
All of this concerns hulls that are designed once, built to a mature plan, and expected to serve for decades. That is the logic of the traditional yard.
Two Industrial Models, Not One
My argument about the future of naval shipbuilding is that the industrial base is bifurcating. One side is the traditional yard and dry dock: large facilities, heavy capital, mature designs, long cycle times, and workforce pipelines measured in years. That side will keep building the ships that cannot be replaced by anything else, and Jones’s reforms are aimed squarely at it.
The other side looks different. Autonomous surface and undersea systems are increasingly built through dispersed, mold-based manufacturing, closer to production lines than to yards. Companies such as MARTAC, with its T-82, and C-2 Robotics in Melbourne illustrate it. These hulls are cheaper, produced in larger numbers, and often built in places that have never built a warship. Their value comes less from the hull than from the payloads and software that ride on it, and those change quickly.
Here the Korean discipline Jones praises becomes a double-edged tool. Design freeze, controlled change, and suppression of late modification are virtues when the product is a destroyer or a logistics ship. The August 13 presidential memorandum Jones cites directs that mature parent designs not be subjected to iterative Navy redesign without senior approval, which is sound for that class of vessel. For an autonomous mesh fleet, though, the capability is expected to change continuously. A fleet node that cannot take a new sensor, a new communications package, or a new autonomy stack within weeks has missed the point of being a node.
The reconciliation, I think, is to freeze the interfaces and leave the capability free. Korean production discipline applied to the hull, the power system, the payload bays, and the standards by which modules connect would make a mesh fleet easier to build at scale. The same discipline applied to the payloads and software would work against everything that makes the fleet adaptive. Controlled change is a good rule for a hull and a poor one for a mission system.
This distinction matters in practice. When the Navy and its allies write the contracting rules that will govern Korean participation, they will either encode the difference or flatten it. If the working assumption is that Korean methods mean maximum design stability everywhere, we will have imported the right habit for half the fleet and the wrong one for the other half.
The Sentence Jones Doesn’t Develop
Jones mentions in passing that Samsung signed an agreement in July 2026 with Saronic, extending its engagement with the American maritime system into autonomous vessels and production automation. He lists it with the other pathways, ownership, investment, repair, design support, and workforce development, and moves on.
I would not make too much of a single agreement, and neither does Jones.
But it points at something the larger discussion has missed. Jones observes that Korean firms are moving toward higher-value vessels, naval work, and smart-yard technologies as Chinese competition erodes their position in standard commercial ships. Automation and production engineering are therefore part of how Korean yards are repositioning themselves, not side interests.
If those capabilities carry over to dispersed autonomous production, Korea could be useful to both halves of the American fleet. It could teach the traditional side how to build disciplined, high-throughput blocks, and it could help the autonomous side industrialize without losing the speed that makes it valuable.
That possibility turns the question around. The usual framing asks what Korea can supply that America lacks. A better framing asks which part of the American maritime enterprise should be learning which Korean lesson, and in what order. The traditional yards need production discipline. The autonomous builders need scale and automation. Treating those as the same problem will produce one program that serves neither.
Commercial Hulls and the Mothership Path
There is also a hull-type connection. Jones argues that commercial workload sustains the industrial substrate on which naval construction depends, and he points to Newport News and NASSCO, which grew through merchant and commercial work before becoming major naval builders. He identifies auxiliaries as one of the most plausible near-term lanes for Korean participation, because they are industrially demanding without the political and technical sensitivity of submarines, carriers, or major combatants.
This is the lane in which commercial and autonomous shipbuilding overlap.
Korea’s strengths are in LNG carriers, very large containerships, and other technically demanding commercial hulls. Those are the kinds of hulls from which motherships and conversion platforms for autonomous systems are most likely to come, whether purpose-built, as in Singapore’s and China’s programs, or converted from commercial and support hulls, as in several American efforts.
A Korean-American partnership that builds auxiliaries and commercial-derived hulls in American yards could serve the conventional fleet and the autonomous fleet at once. Hanwha’s commercial orders at Philadelphia, intended to create workload for the yard, point in the same direction.
What Seoul Wants, and Why It Matters
Jones is unusually clear about Korean motives, and that clarity is the most valuable part of his article. Korean participation is not a simple matter of Washington identifying a deficiency and Seoul filling it. South Korea has long sought freedom of action despite permanent exposure to larger powers. Shipbuilding investment in the United States offers a protected market, diversification away from Chinese competitive pressure, and a chance to turn commercial expertise into strategic relevance, in effect becoming a producer of alliance capability and not only a consumer of American security.
He also notes the limits. Korean firms will be cautious about becoming indispensable to the production of front-line combat power for a Taiwan or South China Sea contingency. Beijing has already shown it can reach into the partnership: in October 2025 it sanctioned five U.S.-linked Hanwha Ocean entities, including Hanwha Philly Shipyard, then suspended the measures for a year the following month. By my reading, that suspension comes up for decision about a year after it was granted, so the partnership’s exposure to Chinese leverage is a live issue now and not a historical footnote.
This is the middle-power problem, and it is not unique to Korea. Australia faces a version of it, with a hull gap in its surface fleet, alliance obligations, and an economy deeply tied to China. Japan faces another.
What these three have in common is that each wants to contribute to allied capability on terms that preserve its own agency and manage its exposure to Beijing. A coalition shipbuilding model that treats them as interchangeable pools of capacity will fail on their incentives.
One that fits participation to what each wants, with a Korean lane in disciplined commercial and auxiliary production, an Australian lane in autonomous systems and sustainment, and a Japanese lane in proven hull designs, has a chance of lasting past the next election in any of the capitals.
Autonomous systems also help with the sensitivity problem. Firms wary of building crewed warships for a Taiwan contingency may be more willing to build commercial-derived motherships, uncrewed hulls, and production tooling.
This is not a way around Seoul’s caution. It is a way of matching the work to what Seoul can accept.
What America Must Change, Revisited
Jones lists eight requirements: durable law and governance, a cleared production-engineering exchange, governed distributed shipbuilding, bounded participation by vessel class, a binding workload compact, a serious workforce compact, a security architecture suited to allied work, and output-based measurement. I would not change the list. I would add four refinements that reflect the bifurcation.
First, write the standards before the contracts. If the aim is to freeze interfaces while leaving capability free, the Navy needs open, enforceable interface standards for hulls, power, and payload bays. Those standards should be set early and should bind any yard, American or allied, that builds to them.
Second, run two contracting tracks. Crewed hulls and auxiliaries belong under the mature-design, controlled-change regime Jones describes. Autonomous platforms need acquisition rules that allow frequent payload and software updates without reopening the hull contract. The failure mode to avoid is applying the slower track to everything because it is the one the system already knows.
Third, extend the workload compact to cover both models. Jones is right that no investment survives if contracted work disappears or moves unpredictably, and his figure for Norfolk’s Colonna’s Shipyard is striking: of 46 solicited 2026 availabilities the firm examined, 21 had reportedly been moved, canceled, delayed, or put on hold. Dispersed autonomous builders face the same instability. A start-up manufacturer cannot hire and tool up against orders that may not materialize.
Fourth, treat workforce as two problems. The traditional yards need welders, fitters, and the white-collar production engineers Jones rightly says are routinely left out. Dispersed manufacturers need people trained in composites, automation, software integration, and rapid production. Korean training partnerships can contribute to both, but only if somebody designs them that way.
The Measure That Matters
Jones closes with a set of output questions. Is usable capacity rising? Is throughput up? Is rework down? Are work packages more mature? Are workers being retained?
I would add one more, specific to the argument here: can the American maritime system now run two industrial models at once?
That is a stricter test than it sounds.
A system that has learned Korean discipline for destroyers and applies it to a mesh fleet will have improved its old fleet and slowed its new one. A system that, in its enthusiasm for autonomous production, abandons design maturity for crewed ships will repeat the delays and redesigns Jones documents. The capacity to hold both logics, and to know which belongs where, is the real organizational change on offer.
If KUSPI helps American yards become more productive, the supplier base deeper, and workload more predictable, Jones is right that it will have mattered. It will matter more if it also helps the United States build a fleet in which a disciplined hull carries a changing payload, built in a mix of yards and factories, with the rules of each fitted to the work it does.
Korean shipbuilders have spent decades learning how to build ships well. The open question is whether American institutions can learn which of those lessons apply to which of their ships.