Architecture of the Deep: Building the Undersea Backbone

A modern subsea cable system bears little resemblance to the wires in your walls. In deep water, these cables are roughly the diameter of a garden hose—4 to 5 centimeters thick. Yet inside that slender tube lies remarkably sophisticated engineering.

The anatomy of a submarine cable. At the core are optical fibers—strands of glass as thin as human hair. Lasers send pulses of light through these fibers at speeds approaching 186,000 miles per second. Surrounding the glass is a protective gel tube, followed by layers of galvanized steel strands, copper or aluminum tubing, polycarbonate, and, finally, multiple layers of polyethylene waterproofing.

Shallow-water cables wear additional armor. Near coastlines—where fishing trawls and ship anchors pose constant threats—operators add stranded steel wire for mechanical strength. Some cables near busy ports resemble industrial hawsers more than communications lines.

The wet plant and dry plant. Industry insiders divide undersea infrastructure into two distinct domains. The wet plant includes everything underwater: the cable itself, optical amplifiers called repeaters, branching units that split signals toward different countries, and equalizers that manage signal strength. The dry plant begins at the beach manhole—the physical point where cable meets land—and continues into the cable landing station, where terrestrial networks finally take over.

Optical amplifiers: the unsung heroes. Light attenuates over distance. Without intervention, a signal crossing the Atlantic would fade to nothing. The solution is the erbium-doped fiber amplifier (EDFA), placed at regular intervals along the cable route. These devices, powered by electricity sent along the cable itself, boost optical signals without converting them back to electrical form. This quiet miracle occurs every 50 to 100 kilometers along the seafloor.

Modern systems now carry 16 to 24 fiber pairs and employ wavelength division multiplexing, sending dozens of different light colors simultaneously through the same glass. Capacities now exceed 500 terabits per second across transoceanic distances—a six-hundred-thousand-fold improvement since the first fiber-optic submarine cable in 1988.

Routes and Reasons: The Geography of Connection

Submarine cable networks do not simply connect random points. Their paths reflect centuries of economic geography and, increasingly, the commercial strategies of a few technology giants.

The corridors of history. The busiest routes remain the transatlantic and transpacific corridors—legacy pathways established by nineteenth-century telegraph cables. London, New York, Tokyo, and Singapore remain central nodes, but the map is shifting.

The hyperscale revolution. Until recently, subsea cable systems were built by consortia of telecommunications carriers. That model has changed. Companies like Google, Meta, and Amazon now own or co-own major cable projects. These internet content providers need dedicated bandwidth between their global data centers, and they have decided to build it themselves.

The 6,600-kilometer MAREA cable, connecting Virginia to Spain, exemplifies this shift. Capable of carrying 160 terabits per second, it was built by a consortium including Microsoft and Meta. It is not a public utility; it is private infrastructure serving specific corporate interests.

Africa's digital transformation. The continent with the most to gain from subsea connectivity is Africa. Thirty-seven of Africa's thirty-eight coastal states are now connected to at least one submarine cable. Meta, Google, and Amazon all have existing or planned cables making landfall on African shores, drawn by the continent's rapidly digitizing economy.

Yet connection does not equal resilience. Several African nations—including Guinea, Liberia, and Mauritania—depend on a single cable. When that cable fails, they lose nearly all international bandwidth. And with only three cable repair vessels serving the entire African continent, outages can last for weeks or months.

Southeast Asia: crowded waters, fragile links. The region is a critical node in global subsea cable networks, linking the Indian and Pacific Oceans. It is also uniquely vulnerable. The world's busiest shipping lanes run directly over thousands of kilometers of cable. In February 2023, all five cables serving Vietnam simultaneously suffered partial or total damage, resulting in a 75% loss of the country's international data flow.

Vulnerabilities and Repairs: When the Internet Breaks

For all their sophistication, submarine cable systems remain surprisingly fragile. Worldwide, there are 150 to 200 accidental cable faults each year.

Who—or what—cuts the cables? The leading cause is fishing activity. Bottom trawling, particularly in waters up to 2,000 meters deep, routinely snags and severs cables. Ship anchors are the second most common culprit, especially in shallow waters near ports and straits.

The famous "shark bite" narrative is largely exaggerated. Sharks do occasionally bite cables—they are attracted to electromagnetic fields—but such incidents represent a tiny fraction of total faults.

Manufacturing defects, improper installation, and simple aging also contribute. Cables that are insufficiently buried or exposed to shifting currents suffer abrasion that eventually compromises their integrity.

The repair process: slow and expensive. Finding and fixing a deep-water cable fault is not like repairing a terrestrial line. When a cable fails, operators first take measurements from the landing station to estimate the fault's location. A cable-laying vessel or construction vessel is then dispatched—a process that can take weeks if the nearest suitable ship is on the other side of the ocean.

Once on site, a remotely operated vehicle (ROV) locates the exact extent of the damage. The faulty section is cut, the ends are sealed and brought to the surface, and a spare section is spliced in. The repaired cable is then carefully laid back on the seabed. The Malta-Sicily Interconnector suffered anchor damage in December 2019; the repair contract alone was valued at €11 million, and the vessel operations consumed 54 days.

The repair vessel shortage. The industry faces a looming capacity crisis. While at least eleven new cable-laying vessels are on order, demand is growing even faster—driven not only by telecommunications but by offshore wind farm export cables. A wind farm with 100 kilometers of export cable has a 24% chance of cable failure in any given year. Over 11 years, the cumulative probability reaches 95%.

The New Geopolitics: Infrastructure as Weapon

The weaponization of the seabed. In September 2022, the Nord Stream 2 pipeline was ruptured by explosions in the Baltic Sea. Whether state-sponsored sabotage or a rogue operation, the incident shattered the assumption that critical undersea infrastructure exists outside the domain of warfare.

Since then, a string of cable disruptions in the Baltic and the Red Sea—some accidental, others suspected sabotage—has triggered a policy reckoning. NATO has deployed a flotilla to monitor the Baltic. The European Union published its first Action Plan on Cable Security in February 2025.

The U.S.-China technology rivalry. The subsea cable industry is highly concentrated. Four companies—America's SubCom, France's Alcatel Submarine Networks, Japan's NEC, and China's HMN Technologies—control 98% of the global market. In August 2025, the U.S. Federal Communications Commission approved new regulations establishing a "presumption of denial" for cable projects involving companies linked to China or other designated adversaries.

Critics argue this approach fragments the global internet. By excluding major suppliers from entire markets, the policy raises costs and slows connectivity in precisely those regions—Africa, Southeast Asia, the Pacific Islands—that need bandwidth most. Yet supporters counter that undersea infrastructure is too vital to entrust to strategic competitors.

The resilience dilemma. Redundancy is the best way to protect against cable disruptions. This means having multiple routes, different landings, and real competition in the repair market. But redundancy costs a lot of money. In a time of geopolitical instability, picking cable partners has become a decision about foreign policy.

European leaders are now in a difficult situation. They are worried about the dominance of U.S. hyperscalers in cable ownership, even though they share American security concerns. Europe is the most connected continent in the world, with more cables landing there than anywhere else. However, it doesn't have its own big cloud provider. Alcatel Submarine Networks can install things at a world-class level, but it still has trouble fixing things.

Conclusion

They say that the internet lives in the cloud. The internet is composed of glass, steel, and polyethylene, and it sits two miles below the seas in the dark.

Subsea cable systems are the most critical infrastructure you have ever seen. They enable global commerce, sustain military alliances, and carry humanity's collective conversation across continents in milliseconds. Yet they operate on trust, reciprocity, and the tacit understanding that the ocean belongs to no single nation.

That comprehension is now being put to the test. Anchors, trawlers, and maybe even state actors exploring the limits of hybrid warfare are cutting cables. There aren't many repair ships. There aren't many suppliers. And the agreement that subsea infrastructure should stay a global public good is falling apart.