Nearly all of the internet's international data — the messages, calls, and searches that cross continents — travels through fibre-optic cables laid on a seafloor that has never been fully surveyed, running over terrain no human eye has seen

The internet is often described as if it lives in the air. We talk about clouds, wireless networks, satellites, signals and invisible connections. The image is useful in daily life, but it hides the more physical truth. When a message, video call, financial transaction or search request crosses an ocean, it is usually not crossing by radio wave through space. It is travelling as light through glass, inside fibre-optic cables lying on the seabed.

That seabed is still one of the least directly known parts of Earth. The cables are planned, surveyed, engineered and maintained with extraordinary care. They are not simply dropped into the ocean at random. But they still run across a planet whose ocean floor has not been fully mapped in modern detail, and whose deep seafloor has almost never been seen directly by human eyes or cameras.

That is the quiet tension behind the modern internet. One of the most heavily used systems humans have ever built depends on long, thin strands of infrastructure crossing terrain that remains, in global terms, only partly known.

The internet is more physical than it feels

The International Cable Protection Committee, the industry body founded in 1958 to promote submarine cable protection, describes submarine telecommunications cables as the backbone of international digital connectivity. In a May 2026 note on global cable infrastructure, the ICPC said these cables carry the vast majority of international data traffic, with roughly 500 cable systems and about 1.8 million kilometres of cable infrastructure now forming the global network.

That figure is easy to read past. It means the internet’s international layer is not primarily an orbital system. Satellites matter enormously for remote regions, ships, aircraft, disaster response, military communications and places where cable connections are absent or impractical. But for the heavy, routine movement of intercontinental data, subsea fibre is the main route.

The reason is capacity. Fibre-optic cables can carry huge volumes of information with low latency and high reliability. Each cable contains glass fibres through which pulses of light encode data. Along the route, repeaters boost the signal so it can travel thousands of kilometres between landing stations. In deep water, the cable itself may be only a few centimetres across. Near shore, where anchors, fishing gear and human activity are more common, it is usually armoured and may be buried beneath the seabed.

So the familiar experience of digital life rests on a surprisingly narrow physical thread. A conversation between continents, a bank transfer, a business file, a search query and a video stream can all become light inside a cable sitting in cold darkness far below the surface.

A mapped route is not the same as a mapped ocean

It would be wrong to imagine cable companies laying these systems blindly. New submarine cables involve careful route planning. Operators examine charts, seabed data, existing cable routes, fishing activity, shipping lanes, volcanic and seismic risk, canyons, slope stability, environmental constraints and national permitting requirements. Route survey vessels can map the corridor in much greater detail than the broad global datasets available from satellites.

But a surveyed corridor is not the same thing as a fully known ocean floor. The corridor can be understood well enough for engineering decisions while the wider seafloor around it remains poorly resolved, unseen or only roughly inferred. That distinction matters because global cable maps can give a misleading impression of certainty. They show neat coloured lines between continents, but those lines simplify an uneven physical world of plains, ridges, trenches, seamounts, canyons and sediment flows.

NOAA Ocean Exploration explains the gap clearly. Satellite-derived maps provide a general picture of the entire seafloor, but their detail is limited. Important features can remain unseen in that kind of map. As of April 2026, NOAA says 28.7% of the global seafloor had been mapped with modern high-resolution technology such as ship-mounted multibeam sonar. NOAA also notes that explorers have directly seen less than 0.001% of the deep ocean seafloor.

Those two numbers sit behind the user’s startling sentence. The ocean floor is mapped in one broad sense, but not in the detailed sense that would make it familiar ground. Much of it has not been visually explored. A cable can cross an abyssal plain or descend near a slope that exists in charts and sonar returns, but has never been viewed directly by a person or by an exploration vehicle’s camera.

The seafloor is not a flat cable tray

From above, the ocean can look like open space. From below, it is terrain. Submarine cables have to deal with geology, sediment, biology and human use. They cross continental shelves, descend slopes, traverse deep basins, avoid some hazards and accept others where there is no perfect route.

The most common cable problems are not usually dramatic acts of sabotage. The ICPC says approximately 150 to 200 submarine telecommunications cable faults occur globally each year, and that around 70 to 80% are caused by accidental human activity such as fishing and ships’ anchors. That is why cables are often buried or armoured in shallower, busier waters. The deep ocean may be quieter in terms of shipping and fishing, but it is not inert. Earthquakes, submarine landslides, volcanic activity and fast-moving sediment flows can damage cables or reshape the ground around them.

There is a long history here. Telegraph cables crossed oceans long before the internet existed, and the basic idea has always been both simple and difficult: connect continents with a line through hostile water. The modern version is far more capable, but it inherits the same dependence on the seabed. The route has to reach land. It has to pass through zones of national jurisdiction. It has to avoid other infrastructure where possible. It has to survive pressure, corrosion, abrasion and occasional human error.

When faults do happen, repair is slow and material. A ship must reach the area, locate the damaged section, retrieve the cable from the seabed, splice in a replacement and lay it back down. In shallow water, where most faults occur, weather, permits and local marine traffic can complicate the work. In deep water, the distances and depths are the problem. The cloud, in other words, has a maintenance fleet.

The map is also a political document

The public Submarine Cable Map gives a useful sense of how dense this network has become. The lines concentrate along certain routes and landing points. That concentration is not accidental. Cables follow demand, geography, permitting, existing infrastructure, seabed conditions and commercial logic.

This makes the network resilient in some ways and exposed in others. Modern internet traffic can often be rerouted when a cable fails, and many regions are served by multiple systems rather than a single fragile line. Yet chokepoints still matter. Landing stations matter. Narrow seas matter. The seabed is not just a scientific or engineering setting. It is also a legal, commercial and strategic space where private companies, states, navies, fishers, environmental interests and coastal communities all overlap.

That is why the language around cables has become more serious in recent years. Governments increasingly describe them as critical infrastructure. The ICPC says its work includes promoting awareness of submarine cables to governments and other seabed users, establishing recommendations for installation and maintenance, and supporting protection under international law.

Still, the basic fact can get lost under the policy language. The world’s digital life depends on cables that most people will never see, crossing a seafloor that science itself has only partly resolved. This is not a flaw in the system so much as a reminder of its physical setting.

What we know, and what we do not

The unseen seabed should not be romanticised. Unknown does not mean unknowable, and darkness does not mean chaos. Engineers know enough to build and operate these systems at remarkable scale. Oceanographers know far more about the seabed than they did a generation ago. Multibeam sonar, autonomous vehicles, remotely operated vehicles and international mapping efforts are steadily improving the picture.

But it is equally misleading to treat the ocean floor as a finished map. NOAA’s figures make the scale of the gap plain. Less than a third of the global seafloor has been mapped with modern high-resolution methods, and only a minute fraction of the deep seafloor has been directly seen. For a planet whose societies depend on ocean routes for trade, energy, data and security, that is a striking imbalance.

Submarine cables sit at the centre of that imbalance. They are among the most advanced communication systems humans have built, yet they are installed in an environment we are still working to describe. Their routes are precise lines through incomplete knowledge.

That may be the most useful way to think about the hidden internet. It is not immaterial. It is not floating somewhere above us. It is glass, steel, copper, plastic and careful engineering, stretched across the seabed between continents. It is maintained by ships and crews, guided by surveys and charts, protected by law and procedure, and still exposed to a world below the surface that remains only partly measured and barely seen.

Every time data crosses an ocean, it passes through that arrangement. The ordinary act of sending something abroad depends on infrastructure laid through deep terrain that most of humanity will never look at, and that our instruments are still gradually bringing into view.