When you send a message from Chennai to a friend in New York, it doesn’t go up to a satellite and bounce back down. It almost certainly travels through a cable roughly as thick as a garden hose, lying on the floor of the ocean, carrying your data as pulses of light at close to the speed of light itself. This is true for over 95% of all intercontinental internet traffic — video calls, bank transfers, cloud backups, streaming shows — and most people never think about it.

There are more than 550 of these submarine cables in service today, stretching over 1.4 million kilometers, which is enough to circle the Earth about 35 times. They are, quietly, one of the most important pieces of infrastructure humanity has ever built. And they are far more fragile, contested, and fascinating than most people realize.

Not Satellites — Glass

It’s a common assumption that international internet traffic mostly travels via satellite. It doesn’t, and the physics explain why. A geostationary satellite sits about 36,000 kilometers above Earth, so a round trip signal has to cover roughly 144,000 kilometers — introducing noticeable lag. A cable across the Atlantic, by contrast, might run 6,000 kilometers, with light traveling through glass fiber at around two-thirds the speed of light in a vacuum. The latency difference is the reason video calls and stock trades feel instantaneous rather than sluggish.

The cables themselves are surprisingly modest in diameter — usually 17 to 25 millimeters near shore, tapering to about as thin as a marker pen in the deep ocean where they’re less exposed to fishing trawlers and ship anchors. Inside, a bundle of optical fibers, often no thicker than a human hair, is wrapped in layers of steel wire, copper, and waterproof insulation. Newer cables typically carry between 8 and 24 fiber pairs, and a single modern cable can move well over 20 terabits per second.

Building One Is a Massive Undertaking

Laying a transoceanic cable is closer to a naval engineering campaign than an IT project. Specialized cable-laying ships — there are only a few dozen in the world capable of this work — spool out cable at a rate of roughly 100 to 200 kilometers per day, guided by seabed surveys that map out mountain ranges, trenches, and fault lines to avoid.

Near coastlines, where water is shallow and human activity is dense, cables are buried under the seabed using a plow towed behind the ship, protecting them from anchors and fishing gear. In the deep ocean, they’re simply laid on the surface of the seafloor, sometimes drifting gently into canyons and settling there for decades.

Costs for a major transoceanic cable typically run from $200 million to $500 million, and construction can take two to three years from planning to activation. That expense used to be shouldered almost entirely by telecom consortiums — groups of national carriers pooling resources. Increasingly, though, cables are financed by a small number of large technology companies that have become some of the biggest owners of undersea infrastructure in the world, seeking direct control over the pipes carrying their own data between data centers.

Things Go Wrong More Often Than You’d Think

Cable breaks happen constantly — by some industry estimates, well over 100 times a year worldwide. Most are mundane: a fishing trawler drags a net across a cable, a ship anchor snags one, or shifting sediment on a slope triggers an underwater landslide that snaps a line. Earthquakes have taken out multiple cables at once, as happened off Taiwan in 2006, which disrupted internet and phone service across much of East Asia for weeks while repair ships raced to fix the damage.

Because there are only a limited number of specialized repair ships in the world, and rough weather or geopolitics can delay access to a break site, a single cable fault can sometimes take days or weeks to fix, even though the physical splice itself might only take a day of actual work once the ship is on-site with the right winter-diving and grappling equipment to haul a cable up from several kilometers down.

This fragility is also why cable routes matter geopolitically. Certain chokepoints — the Red Sea, the Luzon Strait, the English Channel — carry a disproportionate share of global traffic through a narrow physical corridor, making them attractive targets for both accidental damage and, in a few widely reported incidents in recent years, suspected deliberate interference. Governments have become increasingly attentive to submarine cable security as a form of critical infrastructure protection, similar to how they treat power grids or pipelines.

Redundancy Is the Real Hero

The internet doesn’t collapse every time a cable breaks because the system is designed with heavy redundancy. Most regions are served by multiple independent cables along different routes, and network operators can reroute traffic within seconds when one path goes down — usually invisible to end users apart from a brief spike in latency. It’s a bit like a city’s road network: closing one highway causes traffic to redistribute across side streets rather than grinding the whole city to a halt, as long as enough alternate routes exist.

This is also why new cable projects tend to prioritize geographic diversity over pure speed. A cable that shaves a few milliseconds off a route but shares a chokepoint with three existing cables is often a worse investment than a slightly longer cable through an underused corridor, because it actually reduces the risk of a regional outage.

Why This Still Matters

It’s easy to think of “the cloud” as something abstract and placeless, but every video call, every cloud save, every international wire transfer is, in a very literal sense, dependent on physical glass wires resting on the ocean floor, maintained by a small fleet of specialized ships and a global community of engineers who spend their careers thinking about seabed topology and fiber attenuation.

The next time a website loads a fraction slower than usual, or a headline mentions a cable being cut somewhere across the world, it’s worth remembering: behind the seamless experience of a borderless digital world is a surprisingly physical, surprisingly vulnerable network of wires — some as old as the 1980s, still quietly carrying pieces of everyone’s daily lives across the bottom of the sea.