Bulwark Dynamics: A $6.8M Seed and the Bet on Disposable Landing Craft

$6.8 million buys a mid-sized construction crane, a single city bus, or — if the buyer is a Menlo Park startup called Bulwark Dynamics — the first stretch of a production line for landing craft that beach themselves without a crew.

Bulwark closed a $6.8 million seed round on 9 September, and the cap table says more than the headline number. Alongside SBI US Gateway Fund, Archetype Ventures and Nēnē Ventures sits Onomichi Dockyard, a commercial shipbuilder that is not merely writing a cheque. It is signing up to build the boats.

The company's own framing is blunt: distributed operations need a logistics model that does not yet exist, and the hardware to serve it has to be cheap enough to lose.

The vessel nobody was building

Bulwark's product is the Caravel, an unmanned landing craft that carries cargo from ship to shore and drives itself up the beach. The prototype, a Caravel 15, has already been running on San Francisco Bay. The numbers on the company's specification sheet are the interesting part: 23,000 pounds of payload, a 1,100-nautical-mile range, and native compatibility with standard ISO and military container footprints.

Those three figures describe a specific design philosophy. A landing craft that takes standard containers needs no repackaging dock, no crane on the shore, and no crew standing on an exposed beach to catch a line.

Autonomy is the least differentiated layer here. Waypoint navigation, sensor fusion and GPS-denied operation are now table stakes for a dozen marine-robotics teams. What is scarce is a hull that can be produced by the hundred, and a yard willing to tool up for it before the orders exist.

Why the shipyard, not the software, is the bet

Defence procurement has a habit of rewarding the wrong line item. Software demos are cheap and fast; steel is slow and capital-hungry. The bottleneck in unmanned logistics is usually the second one, which is why the Onomichi relationship is the part of this deal worth underwriting.

Bulwark describes the arrangement as a route to series production rather than a one-off fabrication order. The company opened a prototype production facility in Menlo Park in January, and has said the seed proceeds will fund development, manufacturing stand-up and testing in operationally representative conditions.

The company states its intent plainly on its own site:

Logistics win wars. Yet logistics are often overlooked.— Bulwark Dynamics, company website

Strip the rhetoric and the operating model is familiar from other hardware categories: keep the high-margin autonomy and integration work in-house, push the low-margin, capital-intensive hull work to a partner who already has the slipways, the welders and the certifications.

The economics of expendable hulls

The Caravel's range figure is where the strategy becomes legible. At 1,100 nautical miles, the craft is not a harbour shuttle. It is a shore-to-shore delivery system that can stage from a ship well over the horizon, land, discharge, and return without a port.

The thesis rests on a brutal arithmetic: if a hull is cheap enough, losing one is an inconvenience rather than a crisis. A logistics network built from many modest vessels can absorb attrition that would sideline a single high-value ship. Whether that arithmetic closes depends on a manufacturing cost the company has not disclosed.

The same specs make the Caravel legible to civilian buyers. Shore-to-shore delivery without a port is useful for disaster response, remote construction and offshore energy maintenance — markets that do not depend on a procurement cycle to place an order.

The crowded water is the point

Marine autonomy is not an empty field. Saronic Technologies has taken its Marauder unmanned surface vessel from design to on-water trials inside a year, and a long tail of maritime-robotics firms is chasing the same shift from crewed to uncrewed hulls. Bulwark's differentiation is narrower than "autonomy" and more specific than "defence": it is betting on the unglamorous last mile of coastal delivery, where the craft has to touch sand.

That is a harder technical problem than open-water transit. Beaching imposes shallow draft, a bow ramp, sea-state tolerance and self-discharge — constraints that push against the speed and payload figures most marine-robotics platforms optimise for. It also means the closest analogues are not other autonomous boats but the ageing fleet of crewed landing craft and connector vessels they would replace.

Investors should read the seed round as a financing of a manufacturing question rather than an autonomy one. The autonomy stack is necessary but not sufficient; the winner in this niche will be whichever team proves it can build, certify and deliver hulls at a cost that makes exchanging one for a resupply run a rational decision.

What has to be true by 2028

Three things have to fall into place for this to become more than a well-angled seed round.

First, cadence. A co-production agreement is a promise; a monthly delivery schedule is evidence. Until Onomichi's yard is turning out hulls on a published rhythm, the manufacturing thesis remains a slide.

Second, integration. The payload is the product. Bulwark wins if customers treat the Caravel as a container that happens to float and drives itself, and loses if every deployment becomes a bespoke integration project.

Third, a buyer who commits to volume. One-off demonstrations do not pay back tooling. The economics only work for a customer — commercial or public — willing to order in batches.

Can disposable autonomy beat capital ships?

Scenarios to 2028

Sources