For two years, the cheapest way to damage a stretch of Europe’s seabed has been a tired tanker and a dropped anchor. On Christmas Day 2024, a Cook Islands-flagged ship called the Eagle S dragged its anchor for roughly 90 kilometers across the Gulf of Finland, cut five cables including the EstLink 2 power link between Finland and Estonia, and kept sailing. Finnish prosecutors spent half a year trying to prove the crew did it deliberately. Last October a Helsinki court found no evidence the anchor had been dropped on purpose, ruled it had no jurisdiction over acts committed in international waters, and threw the criminal case out. That is the crude way to fight a war on the ocean floor.<\/p>\n
Russia’s Rubin Design Bureau has been drawing up the clean one. The bureau that designs the country’s ballistic-missile submarines is now working on a family of large underwater drones built to carry objects down to the seabed and set them there, presented in official materials as instruments for science. The same machines, according to the open-source analysts who first took the renderings apart, would be very good at putting a mine on a cable.<\/p>\n
The drone at the center of this is the Argus-D, where the D stands for Delivery. It weighs 5.5 tons, runs on lithium-ion batteries, and exists for one task: hauling a payload to the bottom of the sea and dropping it exactly where it is told to. According to independent naval analyst H. I. Sutton, who published the first detailed breakdown<\/a> of Rubin’s marketing material at his site Covert Shores, the Argus-D is 8.9 meters long, carries up to 300 kilograms in a modular bay, and can work at 1,000 meters, with a future version penciled in for 3,000.<\/p>\n It does not operate alone. A smaller sibling, the Argus-I (for Inspector), carries laser scanners and cameras and is officially meant to crawl along pipelines hunting for leaks. Both are designed to be serviced on the seabed by a third machine: a docking station that Rubin renderings show as a sailfish-shaped pod the drones nose into through two funnels. The two outlets that first broke the story down, the Ukrainian defense site Defense Express<\/a> and United24 Media<\/a>, call this station Octavia. It carries its own batteries, recharges the drones, swaps their mission files, stores their survey data, and relays it to shore. The point of it is that the drones never have to surface. A machine that recharges on the bottom is a machine that can sit there and wait.<\/p>\n Argus-D, by the numbers<\/strong><\/p>\n Officially, the payload is science. Rubin describes the Argus-D’s cargo as oceanographic gear: sensors for acoustics, seismic activity, and biological processes, the sort of equipment a research institute leaves on the seabed to listen to the ocean for a year. None of that is invented. Civilian seabed science uses exactly these instruments, dropped in exactly this way.<\/p>\n The trouble is that a modular bay does not care what you load into it. Defense Express, citing Sutton’s assessment, put the second use plainly: the same platform would be “well suited for deploying naval mines or explosive devices on undersea cables.” A payload section rated for 300 kilograms and measured in liters is a payload section whether it carries a hydrophone or a shaped charge, and a delivery drone that can set a beacon down to the centimeter can set something far less friendly with the same precision. The scientific label is not exactly false. It is a cover story, and giving dual-use hardware a civilian description is as old as naval engineering itself.<\/p>\n None of this means an Argus is about to crawl out of the dark and cut Berlin off from the internet. The current design is, by its own numbers, a limited weapon. Its range is roughly 100 kilometers. Its top speed is 6 knots, and it cannot hold that for long: the 20-hour endurance figure assumes a crawl of about 3 knots. There is little foreign infrastructure within 100 kilometers of Russia’s own coastline, which is the only area an Argus could reach under its own power.<\/p>\n To put one near the cables that actually matter, the links running between Finland, Estonia, Sweden and Germany, Russia would have to carry the drone out on a surface ship or a submarine and launch it closer in. That hands back the exact problem the drone was meant to solve. A vessel loitering over someone else’s cables can be photographed, tracked, and, as the Eagle S found out, boarded and seized. Depth is not the obstacle: the Baltic bottoms out near 459 meters, comfortably inside the Argus-D’s rated 1,000. Sutton rates the seabed docking station the least likely part of the whole system to ever get built, which would leave the drone leaning on precisely the kind of host ship it was supposed to replace. Reaching the floor of the Baltic is easy. Getting to the right patch of that floor unseen is the part Rubin has not solved.<\/p>\n So why pay any attention to an unfinished drone? Because the crude version of its mission already works, and works often enough that NATO built a standing patrol to stop it. The Baltic has been losing cables and a pipeline in a steady drip since 2022: the Balticconnector gas line in 2023, telecom links between Sweden and Lithuania in 2024, the EstLink 2 the same December the Eagle S came through. After that one, NATO launched Baltic Sentry<\/a> in January 2025, a mission of frigates, patrol aircraft and naval drones whose entire job is to watch the seabed and the ships above it.<\/p>\n For the better part of a year it held. Then, on the last day of 2025, Finnish authorities boarded a cargo ship called the Fitburg on suspicion of dragging its anchor across the Helsinki-to-Tallinn telecom cable. Two weeks ago, the other side of that contest was on display off the Latvian coast, where the US Navy ran its own seabed drones during the BALTOPS 2026 exercise, the defensive mirror image of everything Rubin is selling and a story worth reading on its own<\/a>. Every one of these incidents has so far needed a ship on the surface to make it happen. That is the single thing an autonomous delivery drone is designed to take away. It is also the same seabed where Denmark is lowering 73,500-ton tunnel sections by the dozen<\/a>, a reminder of how much expensive, immovable infrastructure now sits on the Baltic floor waiting to be either protected or pried at.<\/p>\n The Eagle S is the cautionary tale here, though not in the direction you might expect. Everyone could see that ship. It was tracked, escorted into port, and its crew was charged, and the case still collapsed, because the cables were cut in international waters and a court decided that was someone else’s jurisdiction to prosecute. That is what happens when the evidence is a laden tanker sitting in plain view. Now take the tanker away. Take away the crew, the anchor, the position track, the photographs. What Rubin is offering, if it ever matures past the prototype renderings, is the same result with nothing left on the surface to board. The drone does not have to be good. It only has to be deniable, and on that score the engineering is already most of the way there.<\/p>\n Image credit: Rubin<\/em><\/strong><\/p>\n","protected":false},"excerpt":{"rendered":" For two years, the cheapest way to damage a stretch of Europe’s seabed has been a tired tanker and a … <\/p>\nThe same cargo bay holds a seismograph or a sea mine<\/h2>\n
The drone cannot reach most of what Russia would want to hit<\/h2>\n
The crude version already works, and that is the uncomfortable part<\/h2>\n