Undersea internet cables carry 99 per cent of traffic, but repairs take weeks

More than 99 per cent of all intercontinental internet traffic travels through thin fiber-optic cables resting on the seabed, not through satellites. Around 500 of these subsea highways span the globe, covering hundreds of thousands of miles beneath the ocean surface. When a ship's anchor drags along the seafloor or a deep-sea seismic event snaps a line, the breakdown threatens to disconnect whole regions from the global network.

Yet when an undersea cable breaks, most users never notice a complete blackout. That resilience is built directly into the architecture of the web through route-level redundancy and specialized deep-sea maintenance operations.

What happens when an undersea cable breaks

Subsea cables are surprisingly vulnerable. While encased in protective steel wire and plastic near shorelines, out in the deep ocean they are often no thicker than a garden hose. Fishing trawlers and dragged anchors cause the vast majority of accidental damage, though underwater landslides and seismic shifts also split lines.

When a cable snaps, laser light passing through the microscopic glass fibers inside stops instantly. Automated monitoring systems detect the disruption within milliseconds. If a network relies on a single pipeline, bandwidth drops sharply or cuts off entirely. But modern telecom consortiums and cloud providers design their systems to split data across multiple distinct geographic paths.

If a cable across the Atlantic snaps, algorithms immediately redirect data packets through alternative cables running along different paths, such as via northern or southern transit corridors.

How engineers repair a cable on the ocean floor

Fixing a severed fiber-optic cable thousands of metres beneath the sea is a complex maritime operation. Telecommunication companies do not perform these repairs themselves; they rely on a small global fleet of specialized cable ships stationed at key ports worldwide.

Once a fault location is identified using optical time-domain reflectometers—devices that measure light reflections to calculate the exact distance along the glass fiber where the break occurred—a repair ship sets sail. Upon reaching the site, engineers drop a specialized remotely operated vehicle or a grapnel hook to grab the broken cable ends from the seafloor.

Crew members haul both severed ends up to the ship's climate-controlled clean room. There, specialized technicians strip back the heavy protective shielding and splice the microscopic glass strands back together, aligning them with sub-millimetre precision. Once testing confirms the signal is restored, the spliced section is wrapped in a waterproof joint housing and lowered back onto the seabed.

Why route-level redundancy is necessary

Repairing a broken cable is rarely fast. Depending on weather conditions, sea depths, and the availability of repair ships, a single fix can take anywhere from a few days to several weeks. In some cases, obtaining regulatory permits to work in national territorial waters adds further delays.

Because repairs take time, tech companies like Google, Meta, Microsoft, and Amazon increasingly invest directly in building their own subsea cable networks. Rather than relying on a single trunk line, these companies build mesh networks across multiple ocean routes.

By ensuring that data can flow along three or four alternative physical paths, network operators prevent single cable cuts from crippling international commerce, banking, and communications.