The hidden cables beneath the sea that keep the internet connected
A message sent from one continent to another may cross an ocean before its sender has finished the sentence. The route is not usually a satellite beam. It is a thin strand of glass on the seabed, carrying pulses of light through a cable designed to survive pressure, salt water and decades of use. This infrastructure is easy to overlook because it sits far from the screens that depend on it. Yet the International Telecommunication Union says submarine cables carry more than 99 percent of international data traffic.
That does not mean every internet connection travels under the ocean. Local networks, data centres, land lines, radio links and satellites all matter. But when information must cross a national border or an ocean, submarine fibre cables carry the overwhelming share. They support financial transfers, cloud services, video calls, government networks and the ordinary exchange of messages.
From a landing station to the seabed
A submarine cable begins and ends at a landing station, where it connects to terrestrial networks. Depending on the route, it may be protected near shore and more lightly armoured in deep water. Repeaters along long routes amplify the optical signal; the fibre itself carries data as light. The cables are not broad pipes through which the internet flows in one direction. They are shared transmission systems with multiple fibres, owners, capacity agreements and onward connections.
TeleGeography’s public map shows hundreds of active and planned systems linking landing points around the world. The map looks like a dense web across oceans, but lines on a global map can conceal dependence. Some island states and less connected regions rely on only a small number of routes. If a route is cut, traffic can sometimes be redirected through another cable, but that alternative may have less capacity, travel farther or be unavailable to some providers.
The physical threats are often ordinary. Fishing gear and ship anchors can damage cables, while earthquakes, underwater landslides and other seabed events pose natural risks. The ITU’s April 2026 backgrounder, citing the International Cable Protection Committee, reports more than 170 cable repairs worldwide in 2025—an average of nearly four a week. Most reported faults are not evidence of sabotage. A cable fault is an event; its cause must be established separately.
Why one break can matter more than another
A fault does not automatically disconnect a country. Network operators can reroute traffic if alternative paths exist and if they have enough capacity and commercial access to them. The impact depends on where the break occurs, how many cables share a corridor, whether landing stations or terrestrial links are also affected, and how much spare capacity is available. Redundancy on a map is useful only when the alternate route is working and can carry the load.
The ITU and the International Cable Protection Committee formed an international advisory body in 2024 to address resilience. Its final report, approved in July 2026, identifies physical exposure, geographic concentration, repair delays and reliance on a few systems as major concerns. Its recommendations call for better coordination between governments and operators, more route diversity, faster permitting and stronger monitoring and preparedness.
Repairing a cable is a specialised operation. A repair vessel must reach the fault, recover the cable, splice the damaged fibres and return it to the seabed. Weather, vessel availability, permissions and the depth and location of the break all affect the timeline. A fault in shallow coastal waters presents a different task from one in deep ocean. The system’s resilience therefore depends not only on the cable itself but also on repair ships, spare parts, trained crews and agreements among countries.
Connectivity is a matter of geography and policy
The cable network is privately financed and operated, but it has public consequences. Large technology companies have become important investors and users, while states control landing permits, coastal waters and security rules. A region can attract investment because it is a data hub; it can also be vulnerable if its international capacity depends on limited routes. New cables can bring lower latency and more capacity, but projects require capital, environmental review, coordination and years of work.
Satellites provide a complementary layer, particularly for remote areas and emergency services, but they do not simply replace the capacity of fibre routes carrying intercontinental traffic. The most resilient design is a layered one: submarine routes with geographic diversity, terrestrial backbones, local exchange points and satellite options where appropriate. Operators and governments also need clear plans for sharing information and prioritising repairs.
The quiet lesson beneath the ocean is that the internet is not weightless. It depends on physical lines, permits, ships and decisions about where to invest. Most users will never see a cable or its landing station. They notice the infrastructure only when a route fails, when traffic slows or when a community discovers that the map offered fewer alternatives than it appeared to show. The better measure of connectivity is not how many lines are drawn across the sea, but whether people have another route when one disappears.
Sources: ITU report on submarine-cable resilience, ITU submarine-cable backgrounder, ITU Global Connectivity Report 2025, ITU Advisory Body report, TeleGeography submarine cable map.
