The Navy's $40 Million Robot Warship Sets a Price for Autonomy at Sea

How do you price a warship nobody sails?

Dino Mavrookas had already flipped the third Marauder hull before most of the defense press finished reading the Navy's September 29 announcement. The Saronic co-founder said so in the company's own release, in the flat tone of a man describing a production line rather than a breakthrough. Marauder is 180 feet long, carries up to 150 metric tons of modular payload, and lists a range of 5,400 nautical miles. The Navy agreed to buy ten of them, along with ten each from Huntington Ingalls Industries and Galliano Marine Services, at an average procurement cost of $40 million per vessel.

That number is the story. Not the autonomy demo, not the eight days of round-the-clock trials, not the 140-plus scenarios Saronic says it ran without failure. The number. For the first time, the U.S. Navy has published a unit price for a medium unmanned surface vessel bought at production scale, and it did so through Other Transaction agreements rather than a traditional program of record. Thirty hulls. Roughly $1.2 billion. First deliveries before the end of fiscal 2027.

The Pentagon has spent a decade arguing about whether autonomous ships were ready. The MUSV Marketplace answered a narrower and more useful question: what does one cost, who builds it, and what does the buyer get for the money?

The price is the program

The award structure matters more than the platform specs. Each of the three winners signed a production Other Transaction Authority agreement covering ten vessels, with an average cost of $40 million per hull. The Navy did not disclose how the money splits among the three. It also did not run a conventional competition with a single downselect, a multi-year budget line, and a requirements document frozen years before first steel.

Instead, the service ran a marketplace. Seven companies were picked in May 2026 to bring vessels to an at-sea demonstration. Six met the Phase I bar. Three won production. The three that passed but did not win — Leidos, PacMar Technologies, and Sea Machines — each received $15 million and a listing on the Drones.mil Maritime Marketplace, where other authorized acquisition offices can place follow-on orders. Birdon was the only participant that did not meet the requirements, and the Navy's robotics portfolio manager, Chris Miller, said the company still has a viable path into the marketplace later.

The funding came from the reconciliation legislation known as the One Big Beautiful Bill Act. That detail is easy to skip and important to keep. This is not a line item that survived a normal budget cycle. It is money appropriated through a one-time vehicle, spent through a transaction authority designed to move faster than the Federal Acquisition Regulation, and tied to a delivery date less than two years out.

What the Navy actually bought

Phase I was not a beauty contest. The requirements called for a vessel able to travel 2,500 nautical miles at 25 knots in sea state 4 while carrying up to 25 metric tons across at least two 40-foot shipping containers. The pass test was roughly 26 hours and more than 360 nautical miles with a single command-and-control input allowed during the run. Miller's summary was blunt: the vessel had to demonstrate it could operate beyond 24 hours at extended range with very little operator involvement.

Those numbers describe a logistics truck that happens to float, not a stealthy strike platform. Two 40-foot containers is the interface language of commercial freight. Twenty-five tonnes is a payload budget that fits sensors, a small weapon station, a mine-countermeasures kit, or a resupply module without redesigning the hull. The 2,500-mile range at 25 knots is enough to move between theater ports without a tender.

Phase II, released as a call for solutions on September 23, moves the goalposts in the other direction. The draft requirements reviewed by USNI News cap hull length at 295 feet, ask for 10-day endurance, and require survival in sea state 6. The high-speed mandate is gone. So is the requirement to refuel at sea. Evaluations begin as early as November 16 and run on a rolling basis for five years.

Phase I requirements and Phase II draft parameters, per U.S. Navy statements and USNI News reporting, September 2026.

The relaxation is deliberate. Miller said the Phase I speed mandate created difficulties and that he wanted to back off the harder requirements to keep a healthy, competitive industrial base. Translation: the Navy learned that a demanding first round produces a short list of vendors capable of funding their own prototypes. A looser second round produces more bidders, lower prices, and a marketplace that can absorb whatever commercial shipyards happen to be building.

That is a procurement strategy, not an engineering one. It says the constraint on autonomous naval mass is not autonomy software. It is shipyard throughput and willingness to self-fund a demonstrator.

Three shipyards, three bets

The three winners represent three different theories about how to build a robot ship, and the difference will shape who survives the next five years.

Saronic is the venture-backed vertical integrator. Marauder is built at a yard in Franklin, Louisiana — the former Gulf Craft facility, where the company has put hundreds of millions into expansion. The longer-term plan is Port Alpha in Brownsville, Texas: a project the company values at more than $3 billion, with initial capacity around 150,000 gross tons a year and a path to two million gross tons, which Saronic says would equal about 15% of China State Shipbuilding Corporation's output. The yard broke ground in late September, aims to begin operations in 2028, and is designed from the start to build both crewed and uncrewed hulls. Its first contract there is the Navy's Landing Craft Utility, a manned vessel.

Huntington Ingalls is the incumbent that rented commercial capacity. Romulus 151 is a crewboat-shaped hull with a long aft deck for modular payloads, a published speed of 25 knots and a range of 2,500 miles. HII did not build the prototype in its own steel yards. It turned to Breaux Brothers, a Gulf Coast commercial yard with decades of aluminum work, and let that partner influence the design for manufacturability. The autonomy comes from HII's Odyssey control suite, developed with Applied Intuition, Greenroom Robotics, and Rodi Marine. HII says the system completed 13 days of Navy-monitored testing, including a 420-nautical-mile, 28-hour mission simulation.

Galliano Marine Services, which does business as Edison Chouest, is the offshore-supply builder. It is based in Houma, Louisiana, and its core business is the commercial vessel market that already knows how to produce aluminum workboats at a predictable cost. The Navy did not publish a platform name or detailed specs for the Galliano entry in the award announcement. That absence is itself a signal: the service is buying a production capability, not a branded product.

Platform data from company releases and Navy statements; Galliano's platform details were not included in the award announcement.

Notice what Saronic did with the specifications. Phase I asked for 25 tonnes and 2,500 miles. Marauder advertises 150 tonnes and 5,400 miles, six times the payload and more than twice the range. That is not a vendor meeting a requirement. It is a vendor building a commercial product and letting the Navy buy a version of it. The bet is that a hull with excess capacity can be sold into other missions, other agencies, and other navies without a redesign.

HII's bet is the opposite. Build close to the requirement, use a commercial yard to control cost, and make the autonomy suite the reusable asset. Romulus is not trying to be a better boat than Marauder. It is trying to be a cheaper way for the Navy to get the same container interface from a shipbuilder it already trusts.

Galliano's bet is the simplest. If the Navy wants volume, the cheapest place to get it is a yard that already builds workboats on commercial schedules. No venture capital, no new factory, no autonomy brand. Just a hull, a price, and a delivery date.

The losers got paid

The most interesting design choice in the MUSV Marketplace is what happened to the companies that lost. Leidos, PacMar Technologies, and Sea Machines each received $15 million for completing the at-sea phase. All three, plus the winners, are now listed on the Drones.mil Maritime Marketplace, where authorized acquisition authorities across the Department of War can place follow-on orders.

That is a deliberate attempt to stop the winner-take-all dynamic that usually follows a defense competition. In a conventional program, the losing bidders absorb their prototype costs and wait for the next competition, often years later. Here, the Navy paid for the demonstration, kept the losing designs alive, and created a storefront where a different program office can buy them without rerunning the whole process.

It also creates a price reference. If a program manager can see three qualified MUSV vendors on a marketplace and a published average of $40 million per hull, the negotiation starts from a different place. The Navy has effectively published a benchmark and invited other buyers to use it.

The losers' $15 million is not charity. It is the price of keeping the industrial base warm. Six qualified vendors is a healthier supplier pool than three, especially when the Phase II requirements are looser and new entrants are explicitly invited to submit updated product blocks and modular payload configurations.

Where the margin sits

A $40 million robot ship sounds cheap next to a $2 billion destroyer. It is not cheap next to the thing it replaces. A commercial crewboat of similar size costs a fraction of that, and the autonomy kit is the difference. The question for investors is who captures the gap between steel and software.

The Navy's own framing points at manning. USNI Proceedings has put crew costs at roughly 70% of total ownership cost for a military system, and unmanned vessels are attractive precisely because they remove that line item. But a $40 million hull only pays off if the Navy can operate it with a small shore-based crew and a maintenance model that does not require a full shipyard period for every software update. That is an operating-cost argument, not a capital-cost argument, and it is where the real savings or the real disappointment will show up.

The second place to look is the payload interface. Phase I specified two 40-foot containers. That is a commercial standard, and it means the mission module can be bought separately from the hull. A mine-countermeasures package, a sensor mast, a small strike module, or a logistics deck all fit the same footprint. If the interface holds, the hull becomes a truck and the payload becomes the product. If every mission requires a bespoke integration, the $40 million price is only the beginning of the bill.

The third is the autonomy stack. HII's Odyssey is marketed as open architecture. Saronic builds its own. Galliano's entry is not publicly specified. Open architecture is supposed to let the Navy swap software without swapping the hull, which would turn autonomy into a competitive layer rather than a lock-in. The Pentagon has been here before with open-architecture mandates that produced neither openness nor competition. The MUSV program will be an early test of whether a transaction-authority program can enforce the interface more effectively than a traditional one.

MUSVs give commanders a Tailored Offset — additional capacity and capability that can be rapidly adapted to the threat, distributed across the battlespace, and placed at risk without placing Sailors at unnecessary risk.— Chief of Naval Operations, U.S. Navy release, September 29, 2026

As we wrote in October, the maritime autonomy market is already splitting into two businesses: hulls and brains. The hull builders compete on shipyard capacity and labor. The brains compete on how many vehicles one operator can supervise and how gracefully the software degrades when communications fail. The MUSV award prices both in one number, which makes it hard to tell which half the Navy is really paying for.

Saronic's Port Alpha plan is the clearest expression of the vertical-integration thesis. If the company controls the yard, the hull design, and the autonomy, it can price the whole package and keep the integration margin. The risk is capital intensity. A $3 billion shipyard needs a lot of hulls to pay for itself, and the Navy has not committed to buying enough vessels to fill it.

HII's model is asset-light by comparison. It uses a commercial partner's yard and focuses on the control system. That is a software-and-integration margin, not a shipbuilding margin. It is also more scalable if the Navy spreads orders across multiple yards.

Galliano's model is the pure manufacturing play. If the Navy wants 30 hulls now and more later, the yard with spare commercial capacity can underbid a startup that is still building its factory. The question is whether a commercial workboat builder can absorb the integration burden that comes with a military payload.

What could break the $40 million

Four things could move the price, and three of them are already visible.

The first is the funding vehicle. The MUSV buy is backed by reconciliation money, not a standing program of record. One-time funding can buy a first batch quickly. It does not guarantee a second batch, and it does not create the multi-year certainty that shipyards need to invest in capacity. If the next appropriation does not include MUSV money, the marketplace becomes a one-off demonstration with a price tag attached.

The second is Phase II. Looser requirements are good for competition and bad for price discipline. If new entrants can qualify with smaller, slower, cheaper hulls, the $40 million average becomes a ceiling rather than a floor. That is good for the Navy and uncomfortable for the three incumbents. It also raises a question the first round did not answer: how much of the $40 million was the autonomy and how much was the urgency.

The third is the test standard. Phase I's pass condition was a single-input, roughly 26-hour, 360-mile run. That is a meaningful autonomy test, and it is not a combat-survivability test. Electronic warfare, denied navigation, and degraded communications are the conditions that separate a demo from a deployable system. HII's 13-day trial and Saronic's eight-day trial are more demanding than the minimum, but neither is an operational deployment. If the Navy tightens the standard in Phase II, costs rise. If it does not, the fleet gets a vessel that performs well in the Gulf of Mexico and remains unproven where it matters.

The fourth is industrial concentration. Three yards, two of them in Louisiana and one building a new Louisiana-adjacent supply chain, is a thin base for a 30-hull order and a thinner one for a sustained program. The Navy's own Phase II language about expanding the pipeline is an admission that three is not enough. A single-yard failure, a labor dispute, or a hurricane season could delay the whole schedule.

How to read the next twelve months

The signal to watch is not another demo video. It is the first contract modification. When the Navy changes a payload requirement, adds a mission module, or exercises an option, the price of that change tells you whether the marketplace produced a real commercial product or a bespoke government ship with a commercial label.

The second signal is the Phase II award list. If new vendors qualify at lower prices, the incumbents will have to defend their cost base. If the same three win again, the marketplace is a round-robin with extra steps.

The third is the delivery date. First vessels are due before the end of fiscal 2027, which is a short window for a new hull, a new autonomy stack, and a new production line. A slip of six months is normal. A slip of a year would suggest the price was set before the production risk was understood.

The Navy has done the hard part of price discovery. It has told the market what a medium unmanned surface vessel costs when bought thirty at a time, from three different kinds of shipyard, with a container interface and a published autonomy suite. The next twelve months will show whether that price is a benchmark or a one-time clearing price for a program that never became a program.

For now, the useful number is $40 million. Not because it is low or high, but because it is specific. Specific numbers force specific arguments. The argument ahead is whether autonomy at sea is a shipbuilding business, a software business, or a logistics business that happens to float. The MUSV award suggests the Navy is betting on all three at once, and paying each vendor to prove which one scales.