An American seabed drone with no cable running back to shore heard submarine drones come through a channel and put them on a Navy screen within seconds, from an eight-foot tube that lets go of its own anchor and floats back up when the job is done

If you’d never heard of an Anduril Dive-LD before last week, you’re in good company. Then Iran’s Revolutionary Guard announced on September 8 that it had captured an American uncrewed submarine at the entrance to the Strait of Hormuz, Iranian state media named the model, and a lot more people found out what a 19-foot drone sub looks like.

The Pentagon tells a less exciting version. U.S. Central Command says it was an older model that malfunctioned while surveying the area, and spokesperson Navy Capt. Tim Hawkins told DefenseScoop it was “essentially dead in the water for more than a day” before the IRGC got to it. Anduril called the Dive-LD attritable, which is defense-speak for a machine you build expecting to lose a few of.

I get it. Losing a drone sub in a war zone isn’t exactly shocking, and CENTCOM says the drone the IRGC picked up wasn’t carrying any classified sonar or radar gear.

But the Anduril hardware I’ve been keeping an eye on doesn’t swim at all. It’s called Seabed Sentry, its job is to sit on the ocean floor and listen, and a few of the companies around it have been pretty busy since its first public test with the Navy this summer.

So what is Seabed Sentry?

Each node is basically a fat cylinder about 8 feet long and 21 inches across, with the computer and communications gear up front, a swappable payload bay in the tail and its own anchor. Anduril rates it for depths beyond 1,640 feet and says it can stay down for months or even years depending on how it’s configured, though those are the company’s figures.

Seabed Sentry doesn’t have a cable running back to shore, and that’s pretty much the whole point. Traditional seabed listening arrays do, and laying one usually means a cable ship, years of work and a construction job that’s pretty hard to hide. The node crunches its sonar data on board and only sends out the result, which is basically a short message saying what it heard and where it’s heading. Underwater acoustic modems are often slower than the dial-up internet some of you might remember, so streaming raw sonar through the water isn’t really an option.

When a node is finished, it cuts its own anchor loose and floats to the surface. Anduril engineering executive Shane Arnott told The War Zone when the system was unveiled that a crew can then clean the growth off it, charge it back up and send it down again, or fit a new payload if the mission has changed.

The ears come from Ultra Maritime. Its Sea Spear sonar unrolls from a 21-inch drum into a long listening array held out on a lightweight truss, and Ultra’s CEO has said the design borrows from extendable arms NASA worked on for spacecraft. It’s a passive array, meaning it listens without making a sound of its own.

At Lanternfish, a U.S. Navy exercise with British and Australian crews in the water too, the pair apparently did what it was built to do. Sea Spear picked up medium and large underwater drones moving through a channel, and the node pushed the track through Anduril’s Lattice software into the Navy’s own command system, where an operator at its Unmanned Operations Center had it on screen within seconds. Nobody had to relay it by hand. The Royal Navy says the exercise ran for six weeks across Virginia, Washington state and the waters off southwestern Australia. Still, a port protection scenario means shallow water, and both companies only reported catching medium and large drones, while Anduril’s own pitch says the bigger worry for ports is small ones. Both accounts also come from the vendors, and I’d want to see the Navy’s own scorecard before calling it a slam dunk.

Ultra paired its sonar with a sonobuoy that pings

Here’s where it gets a little more interesting. On August 13, Ultra Maritime said it had also paired Sea Spear with its new SSQ-125B active sonobuoy during Lanternfish, using the array to detect underwater drones passively and the sonobuoy to pin down where they were.

If you’re wondering what the difference is, passive sonar is basically listening in the dark. You can tell something’s out there and roughly which direction it’s in, but figuring out exactly how far away it is gets tricky. Active sonar is more like shouting and timing the echo, which tells you where the target is and also tells everyone else where you are.

That’s a pretty sensible split. Sonobuoys are more or less disposable, so it doesn’t much matter if one gives itself away, and a seabed node you’d like to keep hidden for months is better off keeping quiet.

Ultra calls the SSQ-125B “the only active sonobuoy of its kind,” and the Navy awarded the company a low-rate initial production contract for it in April. I can’t tell you whether a Seabed Sentry node was part of that particular run, though, because Ultra’s August release doesn’t mention Anduril at all.

Lockheed Martin is buying the sonar maker

Ultra is also about to change hands. Lockheed Martin agreed on July 6 to buy it from private equity firm Advent for $3.45 billion, and Ultra will become part of Lockheed’s Rotary and Mission Systems business once the deal closes. Lockheed didn’t give a closing date when it announced the deal, and as far as I can tell it hasn’t closed yet.

That could get a little awkward for Anduril. The two companies called their partnership exclusive when they announced it last year, and I haven’t seen either one say what happens to that arrangement once Lockheed, which already sells sonar of its own, owns the ears on Anduril’s seabed node.

Anduril’s drone subs come out of Rhode Island

Anduril, meanwhile, is working on volume. The company told Defense Daily in late August that it’s pushing the Dive-LD line at its 150,000-square-foot plant in Quonset Point, Rhode Island, to full capacity by the end of 2027. Anduril designed Seabed Sentry to be carried out and dropped by those Dive subs, the extra-large Dive-XL included.

Anduril also gave up its Seattle shipyard. The company confirmed to GeekWire in August that it had moved out of the old Foss shipyard site after the Navy canceled the surface drone program that space was supposed to support.

So what happens when one of these seabed nodes breaks down somewhere it shouldn’t?

Honestly, I don’t know. CENTCOM was quick to point out that the Dive-LD near Hormuz carried no classified sonar, and I’d argue a node whose whole job is running sonar software on the seabed would be a lot harder to shrug off. The recover-and-reuse pitch also assumes the node floats up where you’re waiting for it, and last week was a pretty good reminder that uncrewed hardware doesn’t always end up where its crew expected.

There’s still no announced customer

Seabed Sentry is still waiting on its first announced buyer, roughly a year and a half after Anduril unveiled it. The most obvious opening is in Britain. Janes reported in April that the Ministry of Defence was weeks away from picking a commercial partner for Atlantic Net, the first phase of the Royal Navy’s push to watch the North Atlantic with uncrewed sensors, and that consortia led by Anduril and Helsing were among five still in the running. Helsing is the German company behind the SG-1 Fathom glider, and from what I can find, the MoD still hasn’t named a winner.

Britain has put money behind the idea, at least on paper. The UK’s Defence Investment Plan, published in July, sets aside an extra £1.5 billion over four years for Atlantic Bastion, the wider program Atlantic Net belongs to. AUKUS defense ministers also agreed in May to jointly develop payloads for the three countries’ underwater drone fleets, with deliveries due to start in 2027.

Anduril also told Defense Daily it hopes to finish a second line at Quonset Point for the Dive-XL, the extra-large drone sub that can carry and drop these nodes, by the end of this year.