Attacking Undersea Internet Cables
The cables are already being mauled in peacetime, usually by a ship that claims it was careless. In a war the same trick is available on purpose, and the West cannot put a guard on every mile of glass on the seabed.
Elisabeth Braw's essay in the 26 September issue of The Spectator is the civilian version of that problem. A Chinese icebreaker, the Ji Di, spent fourteen hours last spring inside Palau's exclusive economic zone. Palau has two fibre-optic cables and no navy that can see a research ship off. Chinese survey vessels have turned up uninvited off Taiwan, Vietnam and the Philippines. In Europe the Russian ship Yantar, fitted with submersibles and in the habit of appearing near cables, loitered north of Scotland last November, outside territorial waters, and shone lasers at the RAF pilots sent to watch it. John Healey's line was "we see you." The ship still finished its visit. Braw's point is that the seabed is mostly nobody's, the visitors do not have to declare a mission, and a rule-of-law navy cannot seize a hull because it dislikes the itinerary.
The attack that has actually been used is not a submarine duel. It is a merchant ship, an anchor, and a cable chart. Since 2021, northern European governments have been looking at cuts and asking whether they are looking at unusually bad seamanship. Anchors dragged across a marked cable do the damage; the ship is gone, or it says the drag was an accident, before a repair vessel can even be tasked. In February 2025 the Chinese-crewed Hong Tai 58 dragged an anchor across the cable between Taiwan and Penghu inside a no-anchoring zone. Taiwan charged the captain and a court later sent him to prison, a rare case in which the ambiguity broke. On 31 December 2025 Finland boarded the cargo ship Fitburg, suspecting damage to a Helsinki–Estonia telecoms cable. On 30 March 2026 the Taiwan–Matsu No. 3 cable was partly cut while a Chinese vessel was working a stranded fishing boat. Iran, once the war started, talked about charging fees on internet cables in the Strait of Hormuz, and a repair firm paused regional work. The method is the deniability. A torpedo is an act of war. A dragged anchor is a court case.
Mapping is the other half, and it can be done in daylight. The ships Whipps and Healey worry about can survey the bottom, note where a cable is buried and where it is exposed, and hand the chart to someone else. Braw treats the Yantar as possibly a diversion: send the famous spy ship, draw the frigates, and do the quieter work elsewhere. China has taken delivery of 24 submarines in five years. Russia treats submarines as its naval advantage. Britain has just test-fired a torpedo from the unmanned submarine Excalibur. None of that guards a cable in the Philippine Sea or the mid-Atlantic. A patrol can escort a suspicious hull for a day. It cannot escort a network.
Protection, such as it is, is redundancy, burial, and a few more ships. NATO's Baltic Sentry patrols, the EU's cable-security plan and a €347 million toolbox announced in February 2026, and Taiwan's own cable initiative are watches and prosecutions, not a shield. Cables can be armoured and buried in the shallows, where anchors reach. In deep water they lie on the bottom, because burial is impractical, and that is most of the route. Landing stations can be hardened. The spare capacity on other routes is the real defence: one cut is an outage measured in hours if traffic fails over. Several cuts on the same day, or any cut to an island with two cables, is a different event. Repair ships are few, the specialised ones are booked months ahead, and a hostile coast can refuse them entry. Palau cannot replace a cut pair. Matsu has already learned what a single damaged link does to an outlying island.
If the network is actually destroyed, not nicked, the first loss is not the army. It is everything that assumed the glass would be there. Finance, aviation, logistics, hospital records, and the civilian internet fail over until the spare routes congest and then fail. Military command does have satellite and radio paths; they are slower, easier to jam, and not sized for a country's ordinary traffic. A coordinated cut of the transatlantic and Indo-Pacific trunks would not need to be permanent. Weeks are enough. The repair clock starts only when a ship can sail, and in a war the ship is a target on a known course to a known break. Braw's forecast is modest and worse for that: more mauled cables, submarines shadowing each other, robots on both sides, and a mineral dispute on the same seabed. The West is buying submarines. It is not ready for a war whose first shot is an anchor in international water, and whose evidence is a broken strand nobody owns.
https://spectator.com/article/is-the-west-ready-for-the-undersea-war/
