A submarine that never breaks the surface can now send out an underwater drone and bring it back through the same torpedo tube it left from. That is the result of a test completed at Seneca Lake in upstate New York, where a second-generation REMUS 620 unmanned underwater vehicle launched from and was recovered into a test fixture built to match the torpedo tubes aboard the Navy’s Virginia-class attack submarines. The trial, announced by the drone’s manufacturer HII, marked the first time this class of vehicle completed the full cycle of exit and reentry through a torpedo tube interface.
If the technology survives the jump from a freshwater lake to the open ocean, it could reshape how American submarines operate in contested waters. Crews would be able to deploy autonomous drones for surveillance, mine detection, or intelligence collection and then retrieve them, all while staying deep and undetected. In a potential conflict in the Western Pacific, where Chinese anti-submarine capabilities are expanding rapidly, that kind of stealth matters.
How the test worked
The REMUS 620 is a medium-class autonomous underwater vehicle roughly 12.75 inches in diameter, designed to fit inside a standard 21-inch torpedo tube. During the Seneca Lake trial, the drone was first “swum out” through a Virginia-class torpedo tube and shutterway test fixture, replicating the exact dimensions and mechanical interface of the tubes found on the Navy’s front-line attack boats. After operating independently in the lake, the vehicle navigated back to the fixture, acquired its docking target, and aligned itself for recovery.
The key enabling technology is a system called Yellow Moray, originally developed at the Woods Hole Oceanographic Institution. Yellow Moray is a torpedo tube launch and recovery (TTL&R) guidance package that steers an unmanned vehicle through the precise alignment needed to enter a torpedo tube underwater. During recovery, the REMUS 620 autonomously docked with a component called the Shock and Fire Enclosure Capsule, a protective housing that sits inside the tube and shields the drone during the high-force mechanical events associated with weapons handling aboard a submarine.
The practical appeal is that submarines would not need external payload modules bolted to the hull or separate dry-deck shelters to carry drones. No extra drag, no visible modifications that adversary sonar or satellite imagery could spot. The drone leaves through a standard tube, completes its mission, and returns for data offload or recharging. The submarine stays quiet the entire time.
Earlier work aboard USS Delaware
The Seneca Lake test did not come out of nowhere. HII and Navy budget documents reference earlier Yellow Moray operations conducted aboard the USS Delaware, a Virginia-class boat, establishing that the underlying TTL&R technology has already been tested inside an actual submarine hull. Those earlier trials provided the operational foundation: proving that the guidance system and mechanical interface could function in the cramped, weapon-packed environment of a real torpedo room.
The Seneca Lake event built on that work by pairing the proven Yellow Moray system with the newer, second-generation REMUS 620 vehicle and a full torpedo tube fixture, rather than a simplified stand or tank setup. It represents a deliberate progression from concept development to hull integration to a more mature pairing with an operationally relevant drone.
Why torpedo tubes are a hard engineering problem
Torpedo tubes are optimized for large, rigid weapons like the Mk 48 heavyweight torpedo and Tomahawk cruise missile. Fitting a relatively delicate sensor platform through the same space, repeatedly, without damaging its hull, sensors, or propulsion systems is a nontrivial challenge. Clearances are tight. The mechanical environment is harsh. And the torpedo rooms aboard Virginia-class boats are among the most densely packed spaces on the ship, with every cubic foot allocated to weapons storage, loading rails, and firing mechanisms.
That raises practical questions the public record does not yet answer. Whether the Shock and Fire Enclosure Capsule displaces a weapons slot, how crews manage the transition between drone operations and conventional weapons loading, and whether the drone can be recharged in place inside the tube or must be moved to a separate bay are all unknowns. Any system that competes with torpedo storage or slows reload drills will need to justify itself in terms of operational payoff.
What still needs to happen
The distance between Seneca Lake and a deployed capability is significant. Several important gaps remain.
First, the test used a fixture, not an actual submerged submarine. Ocean conditions introduce currents, salinity gradients, temperature layers that bend acoustic signals, biological fouling, and the challenge of a moving boat holding position while a small drone attempts to dock. No official timeline for at-sea integration testing has been disclosed.
Second, HII’s announcement included no quantitative performance data: no docking success rates, no count of launch-and-recovery cycles, no time-per-cycle figures. Without that information, outside analysts cannot gauge how close the system is to the reliability threshold the Navy would require for operational patrols.
Third, how the REMUS 620 transfers mission data back to the submarine after recovery remains unclear. Submarines operate under strict emissions control, and any data link between the drone and the boat’s combat systems would need to function through the tube interface without generating detectable signals. Whether the current design uses a physical connector inside the capsule, an acoustic link, or another method has not been described publicly.
Finally, the concept of operations for these vehicles once launched is undefined in the available material. How far from the host submarine the drone operates, how long its missions last, and what level of autonomy it uses to navigate and avoid detection will all shape how commanders assess the risk of using it.
Where this fits in the Navy’s unmanned undersea strategy
The REMUS 620 torpedo tube program is one piece of a broader push to integrate unmanned systems into the submarine fleet. The Navy is simultaneously developing larger platforms like the Orca Extra-Large Unmanned Underwater Vehicle, built by Boeing, which is designed to operate independently for months at a time. The Orca is too big for a torpedo tube; it launches from a pier or a surface ship. Torpedo-tube-compatible drones like the REMUS 620 fill a different role: shorter-range, recoverable missions that extend a submarine’s sensor reach without requiring the boat to expose itself.
That distinction matters strategically. A Virginia-class submarine carrying a mix of torpedoes, cruise missiles, and recoverable drones could conduct intelligence-gathering sorties in denied waters, map minefields ahead of an amphibious assault, or monitor chokepoints, all without surfacing or relying on external support. The submarine becomes not just a weapons platform but a forward-deployed sensor node that can refresh its picture of the battlespace autonomously.
As of June 2026, no primary source has published a target date for fleet-wide deployment or specified how many drone cycles a submarine could complete during a single patrol. The technical barriers are clearly falling, but the path from a controlled lake in New York to routine operations across the fleet still contains real engineering and operational unknowns. What the Seneca Lake test does confirm is that the basic mechanics work: an autonomous underwater vehicle can leave a torpedo tube, complete a mission, and come back, all while the submarine stays hidden below the surface.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
