Every underwater drone story ends up being a story about the battery. The autonomy can be brilliant, the sonar can be brilliant, and then somebody asks how long the machine can stay down and the honest answer gets measured in days.<\/p>\n
Northrop Grumman’s answer was to build something that spends most of its working life parked on the ocean floor with nearly everything switched off. It is called Manta Ray, it was built for DARPA, and the last time anyone publicly watched it swim was March 2024, off the coast of Southern California.<\/p>\n
Plenty has happened in the extra-large drone business since. The Navy’s May 2026 shipbuilding plan moved Boeing’s Orca into real procurement. The Pentagon’s Defense Innovation Unit has named at least three separate industry teams to design a new extra-large uncrewed undersea vehicle, the most recent of them in July. Manta Ray has been quiet through all of it, which is an odd outcome for the one American machine built specifically around the endurance problem everybody else is now spending money on.<\/p>\n
Manta Ray is a glider, and gliders have been in the water since the 1960s doing survey work for the Navy and NOAA. The mechanism is old and mildly counterintuitive.<\/p>\n
Rather than turning a propeller, the vehicle changes its own weight. It pumps seawater in to get heavy and sink, pumps it back out to get light and rise, and the wings convert all that vertical falling into forward travel. Brian Theobald, chief engineer for Manta Ray at Northrop Grumman<\/a>, has described the propulsion as “falling forward [with purpose] through the water all the time,” going up as well as down.<\/p>\n The energy bill is the reason anyone bothers. A glider only draws real power during the few minutes at the top and bottom of each arc, when it is changing buoyancy. The rest of the run costs almost nothing. Teledyne’s Slocum Sentinel, part of a fresh Royal Navy glider order this year<\/a>, uses the same physics to stay under for more than two years on a single load of batteries.<\/p>\n Manta Ray is not a purist about it. DARPA’s 2024 testing exercised buoyancy, propellers and control surfaces, so the vehicle can push when it needs to. It just does not need to very often.<\/p>\n The part that got DARPA’s attention had nothing to do with the water. Northrop assembled the vehicle in Annapolis, Maryland, broke it back down into subsections, shipped it across the country in five standard shipping containers and put it together again in California.<\/p>\n Kyle Woerner, then DARPA’s program manager for Manta Ray, said the test validated the vehicle’s “readiness to advance toward real-world operations after being rapidly assembled in the field”<\/a>. He called the combination of cross-country transport, in-field assembly and deployment a first for an extra-large UUV.<\/p>\n The logic underneath is unglamorous and sound. Every mile an underwater drone swims to reach its patrol area is a mile of battery it does not get to spend on the actual job. Ship it in crates and the energy budget starts where the mission does. It also keeps the thing off a naval pier, which matters more than it sounds when pier space is the bottleneck.<\/p>\n Neither Northrop nor DARPA has published a length, a wingspan or a weight for the vehicle. The clearest public sense of scale is a DARPA photograph of Woerner standing on top of it, mid-conversation with a member of the Northrop team.<\/p>\n Northrop’s own description<\/a> of Manta Ray runs to four points: it carries payloads, it runs autonomously with nobody on site, it ships in five containers, and it saves energy by anchoring to the seafloor and hibernating in a low-power state.<\/p>\n That last one is doing the heavy lifting, because hibernating is not the same as being switched off. A vehicle sitting on the bottom in a low-power state can keep a clock running and a receiver listening, then wait for a mission command without eating the reserve it will need later.<\/p>\n Where the power comes back from is the harder question, and DARPA has never hidden the goal. The program’s public description<\/a> asks for energy management and undersea energy harvesting at operationally relevant depths, low-power propulsion, and ways to stop biofouling and corrosion from chewing up a machine meant to be down there for a very long time.<\/p>\n Northrop went at that with Seatrec, a renewable-energy company whose hardware pulls power out of the ocean’s thermal gradient, the temperature gap between the warmer mixed layer near the surface and the colder water below it. The two later pushed the idea into a concept called the Mission Unlimited UUV Station<\/a>, an entry in a NOAA and Department of Energy ocean observing prize, built around Seatrec’s thermal energy pod and a Northrop connector that can be mated while submerged.<\/p>\n The second Manta Ray performer took a different road. PacMar Technologies of Honolulu announced a $54.8 million Phase 2 contract in February 2022 and spent 2024 testing a full-scale energy harvesting system. When fishermen off Waikiki came across a bright yellow object in the water that year, DARPA confirmed to The Debrief that it was a PacMar prototype built to test exactly that.<\/p>\n The May 2026 shipbuilding plan is where the money turns up. The Navy plans to buy two Boeing Orca XLUUVs in fiscal 2027 for $135.8 million, then another 14 across the five-year defense program, for $1.13 billion across the whole XLUUV line, according to Naval News<\/a>.<\/p>\n Orca is not a glider. It is a hybrid diesel-electric vehicle about 51 feet long that stretches to roughly 85 feet with its modular payload section fitted, launched from a pier because it is too big to come off a ship or out of a submarine. It is also the program the Government Accountability Office found in September 2022<\/a> to be at least $242 million and 64 percent over its original cost estimate, and at least three years late.<\/p>\n Running alongside it is a separate Defense Innovation Unit effort called the Combat Autonomous Maritime Platform, or CAMP, meant to prototype extra-large drones faster than a conventional program can manage.<\/p>\n Metron announced a CAMP prototype agreement in December 2025 for a system called Lancet, built around Cellula Robotics’ fuel-cell Guardian vehicle. Anduril was selected in March 2026<\/a> for Dive-XL, the American relative of the Ghost Sharks it builds for the Royal Australian Navy, and says its undersea vehicles have logged 42,355 kilometers and 6,752 hours of mission time. Kongsberg and Oceaneering came in this July<\/a> for concept definition, system architecture and design trade studies, with a design due in the third quarter.<\/p>\n None of those is Manta Ray, and DIU has not publicly announced a down-select. Anduril, for its part, has been busy on the seabed in other ways this year, putting sonar tracks onto a Navy screen from an untethered node during a six-week AUKUS exercise<\/a>.<\/p>\n The program page has not come down. DARPA still describes Manta Ray as a multi-phase effort aimed at a new class of long-duration, long-range, payload-capable UUVs, and currently lists Kevin Sloan as its program manager. Woerner’s name is on the 2024 test release. Sloan’s is on the program page now.<\/p>\n What has not appeared since May 2024 is another milestone. DARPA said then that it was engaging with the Navy on next steps for testing and transition. Neither has announced what came of that, and Northrop’s page still describes what the vehicle will do rather than what it has done.<\/p>\n That is not evidence a program has been killed, and it would be sloppy to read it that way. A lot of DARPA work goes quiet because the interesting part moved somewhere that does not issue press releases. But the public record is the public record, and on the public record the last thing Manta Ray did was get bolted together on a California dock and swim for two months.<\/p>\n There is a reading where the silence is the whole point. A machine designed to sit on the seabed unnoticed for months is not one you would expect to see much of, and the Navy has never been talkative about what it wants extra-large drones doing out there. A manta-shaped machine did make news this year, incidentally, just not this one: Chinese state television showed a flapping-fin drone running a mine hunt in near-zero visibility<\/a> in March.<\/p>\n The less flattering reading is simpler. Manta Ray answered the hard part of undersea endurance on paper and in one two-month test, and the 2026 money went to vehicles that are easier to buy, easier to schedule and easier to defend in front of a budget committee. The drone the Navy actually keeps ordering in volume is a 10-foot tube called Lionfish<\/a>, not anything with wings.<\/p>\n Manta Ray was built to do the thing none of them can do yet, which is stay. Two years on, that is still the most interesting sentence anyone can write about it, and it is still written in the future tense.<\/p>\n","protected":false},"excerpt":{"rendered":" Every underwater drone story ends up being a story about the battery. The autonomy can be brilliant, the sonar can … <\/p>\nFive shipping containers and a build job at the far end<\/h2>\n
The hibernation is where the endurance actually comes from<\/h2>\n
The Navy spent this year writing checks for other people’s drones<\/h2>\n
DARPA still lists Manta Ray as a program<\/h2>\n