What Is a Submarine Cable? Nearly All the Internet Lies on the Seabed

At the centre of a modern submarine cable sit a small number of optical fibres, each a strand of extremely pure glass thinner than a human hair, carrying light pulses that encode data. Around them is a steel and copper structure that provides strength and carries electrical power, then waterproofing and insulation, and in shallow water heavy armouring of steel wire. The copper matters because the cable must power repeaters, which are amplifiers spaced roughly every fifty to a hundred kilometres along the route that boost the optical signal to counter the loss over distance, and those repeaters are fed by a high voltage applied from the shore stations at each end. In deep water the cable is unarmoured and remarkably slim, because there is nothing there to damage it, whereas on the continental shelf it is heavily protected and frequently ploughed into the seabed, since that is where fishing gear and anchors are. Capacity has grown enormously without the cable changing much physically, because improvements in the light sources and in how many separate wavelengths a single fibre carries have multiplied throughput on existing glass.

The operation is specialised and depends on a small global fleet:

• •Route surveying, which maps the seabed to avoid trenches, seamounts, volcanic areas, coral and known fishing and anchoring grounds
• •Manufacture in continuous lengths, with the cable coiled into enormous tanks in the ship
• •Laying at a controlled speed and tension, paying out slightly more cable than the distance covered so it follows the contours without suspending across gaps
• •Burial near shore using a plough towed behind the ship that cuts a furrow and lays the cable into it
• •Fault location by measuring the electrical and optical characteristics from shore, which narrows a break to within a short stretch
• •Repair by a cable ship grappling the cable off the bottom, cutting it, raising each end, splicing in a new section and lowering it back, which takes days to weeks and depends on a ship being available

Roughly a hundred to two hundred faults occur worldwide each year, and the large majority are caused by human activity in shallow water, principally fishing trawls and ship anchors dragging across the cable. Natural causes include submarine landslides and turbidity currents, and an earthquake off Taiwan in 2006 severed multiple cables at once and disrupted communications across Asia for weeks, while a volcanic eruption cut Tonga off almost entirely in 2022 because the country depended on a single cable. Shark bites were a genuine problem for early fibre cables and are now rare. Deliberate damage is a live concern, with several incidents in the Baltic and elsewhere prompting investigations into anchors dragged across cables, and the vulnerability is real because the cables are unprotected, their positions are published for safety reasons and repair capacity is limited. Resilience comes from redundancy rather than from protection: a well-connected country has many cables on diverse routes so that traffic reroutes automatically, and the places that suffer outages are those served by one or two.

Ownership has shifted markedly. Cables were historically built by consortia of telecommunications carriers sharing the cost, and over the past decade the large content and cloud companies have become the dominant investors, financing and in some cases wholly owning transoceanic cables to carry their own traffic, which concentrates control of intercontinental capacity in a small number of firms. Landing stations are the other concentration point, since cables come ashore at a limited number of suitable locations and a handful of sites carry disproportionate traffic, which makes them both economically and strategically significant. Routing follows old patterns for physical reasons, so a message between two African countries may travel to Europe and back because the direct capacity does not exist, and building regional capacity has been a deliberate development priority. The historical continuity is striking, since the first transatlantic telegraph cable was laid in 1858 and failed within weeks, the durable one followed in 1866, and several modern cables come ashore near the same beaches for the same reasons of geography.

Optical fibres on the seabed carry nearly all intercontinental data, with satellites handling a small fraction. The cable is slim in deep water and heavily armoured near shore, and copper conductors power amplifiers spaced every fifty to a hundred kilometres. Most of the hundred-plus annual faults come from trawls and anchors in shallow water. Resilience comes from having many diverse routes, which is why single-cable countries lose everything at once.