Nine Builders, Three Boats

Week 47 | November 2026

On 14 October, about forty people watched a 368-ton former US Navy test vessel work through three and a half hours off Long Beach under autonomous control. It navigated, took new tasking mid-mission, sailed in formation with a second hull and handled staged machinery failures, all of it relayed to a control station ashore. HD Hyundai supplied the ship autonomy, and Anduril the mission layer. The audience included the US Navy, the US Coast Guard, ABS, the Republic of Korea Navy and DAPA, Seoul’s procurement agency.

Nobody signed anything. HD Hyundai’s release says the partners “plan to pursue US unmanned surface vessel programs.” Korea’s Etoday was blunter about both Korean groups now building for the American market: starting fabrication “does not mean the US military has placed a production order” (my translation).

Before reading that as another slow-customer story, look at what the customer did in September.

In Week 32 I said to watch the October results from the Navy’s MUSV marketplace. They came early. On 29 September the Navy gave production agreements to Galliano Marine Service, HII and Saronic for ten vessels each, at an average of US$40 million a hull, with first deliveries before the end of FY2027. All three build in Louisiana. All six boats that finished testing earned US$15 million. Leidos, PacMar and Sea Machines got that and a catalogue listing, with no production order. Birdon got neither.

Six months from cancelling MASC to production awards. Whatever else ails US naval acquisition, this part moved.

Anduril bid for Phase I and missed the list, as I noted in July. Phase II’s Call for Solutions went out on 23 September with lighter requirements: no high-speed requirement, no refuelling at sea, at least ten days unmanned, and two 20-foot containers in place of two 40-foot. Chris Miller, the Navy’s acting portfolio manager for robotic and autonomous systems, explained why: “I’m trying to back off of those so that I can have a healthy competitive industrial base.” Both Korean-backed teams are lining up for that door.

HD Hyundai is selling the ship’s nervous system. Its VAone platform runs on Avikus’s HiNAS navigation software and adds machinery monitoring and damage control, with Lattice on top for missions. Kim Hyung-taek, who heads HD HHI’s naval research institute, made the pitch in Seoul in September: “On medium and large naval ships, automating navigation alone cannot replace the crew’s work” (my translation). The machinery-failure serial at Long Beach was that argument staged for a Navy audience. The release also claims weapons autonomy, which nobody independent has described.

Hanwha is selling what goes on deck. With Miami’s Magnet Defense it is building the H38, a 38-metre uncrewed version of Magnet’s M48 catamaran. The M48 design comes from Magnet’s crewed 48-metre research vessel, which logged 32,000 nautical miles of AI-assisted trials, including a 2024 round trip from Miami to American Samoa. Hanwha takes most of the construction and the weapons integration, and Magnet the navigation autonomy. At AUSA in October the pair showed containerised Chunmoo rockets and K-VLS missile cells. Kevin Schoonover, a Hanwha senior vice president, put the H38 at “about half” the M48’s roughly US$30 million.

Because everyone does. The marketplace only buys boats that already work at sea, so the cost of proving them falls on industry. Magnet says it has put more than US$50 million into the M48. Saronic is committing more than US$3 billion of private money to Port Alpha in Brownsville. The Koreans are playing by the same rules.

They do carry one extra constraint. In Week 41 I walked through 10 U.S.C. § 8679, the bar on building armed-forces vessels in foreign shipyards, and why it doesn’t reach a Korean-owned yard on American soil. The statute has no exception for uncrewed vessels: its only exception covers inflatable and rigid inflatable boats. And August’s presidential waiver memo names surface combatants, tankers and RoRo ships. USVs aren’t on that list.

So both plans route around it. HD Hyundai’s Ulsan hull is a company prototype built under contract to Anduril, which has named Edison Chouest Offshore as its US production partner. On my reading, a hull built for a company’s own trials isn’t “constructed for” the armed forces. A delivered production MUSV would be, and so would this hull if the Navy ever takes title to it. The partner comes with a twist, too. Edison Chouest, trading as Galliano Marine Service, already holds one of the three production awards, for a hull of its own. Hanwha builds in the US from day one.

The more telling move was quieter. In April, HD Hyundai, Anduril and ABS signed a trilateral MoU that HD Hyundai describes as covering “regulatory and certification frameworks” for autonomous surface vessels. That’s a builder paying to lay a certification path before any customer has asked for one.

Here’s the arithmetic that should worry everyone in this race. The Navy has ordered thirty boats, worth about US$1.2 billion. Its own budget says what comes next. After 36 hulls funded in FY2026, the FY2027 request buys three. Across the five-year plan it buys 47, about US$4.4 billion, at ten to twelve a year from FY2028.

Now count the credible suppliers chasing those hulls: three winners, three listed finishers, Birdon, and two Korean-backed teams. Nine in all, though two of them would build through Edison Chouest. Blue Water Autonomy and Saildrone, shut out of the first round, are suing over it.

The Korean safety net is thin too. MASGA, the US$150 billion shipbuilding package Seoul pledged last year, still has “no concrete projects,” according to Korea JoongAng Daily. A trade ministry official told Baird Maritime the headline number was “a figure that came out during negotiations.”

There is a home market. DAPA told the National Assembly on 13 October that combat USV acquisition would go into full swing, and an official says the program should start next year. Its long-time domestic USV developer, LIG D&A, faces a 24-month DAPA bidding ban, paused by a court on 27 October pending a 5 November ruling. If the ban holds, that leaves room for exactly these two groups.

Building on spec is rational. The first allied builder with a certified, US-built, armed hull can sell it in Washington and in Seoul, and Congress can still add money the Navy didn’t ask for. Being early costs far less than being late.

Fair. It still only pays if the follow-on buys are big enough to share, and on the Navy’s current plan they are smaller than the first.

For program offices, September’s lesson is that the marketplace works. Industry will fund the hulls if you buy what passes. For builders, the same month is a warning. On the Navy’s own numbers, the year that bought thirty boats was the peak. The queue for what follows is longer, and the orders are smaller.

Korea can build these ships, and has chosen to prove it at its own expense. Whether American demand is big enough to repay that is the question Long Beach couldn’t answer.

This one isn’t finished. The Ulsan prototype is due in the water as early as this month, after slipping from October, and the H38 starts testing in January. Watch which of them enters Phase II, whether Congress adds money to a profile that buys three boats next year, and what happens to the FY2026 money for six hulls that don’t yet appear to be on order.

Since you have been, thanks for reading.

Cheers,

Mick

Ocean Tech Intelligence provides informational analysis only. Nothing in this publication constitutes financial, investment, legal, or strategic advice. Readers act on this content at their own risk. For full details see our Disclaimer.