Rivals Jointly Fund the Cable They All Need on the Seabed
Nearly all intercontinental data travels through fibre optic cables lying on the ocean floor. They cost hundreds of millions, take years to lay, and are usually financed by consortiums of companies that compete everywhere else.
The Physical Reality Behind the Cloud
Intercontinental internet traffic does not travel by satellite in any meaningful volume. It travels through fibre optic cables roughly the diameter of a garden hose, laid on the seabed by specialised ships, connecting landing stations on opposite coasts.
A major transoceanic system costs in the range of a few hundred million dollars and takes two to four years from planning to service, most of which is permitting, marine surveys, and manufacturing rather than laying.
The cable itself is passive glass. The capacity it delivers is determined by the equipment at each end, which means an existing cable can be substantially upgraded without touching the wet plant, a property with important economic consequences.
The Cost Structure Determines the Ownership
Once a cable is built and lit, the marginal cost of carrying an additional bit is close to zero. All the cost is in construction and in the periodic upgrade of terminal equipment.
That produces a familiar and dangerous shape. Capacity arrives in enormous discrete increments, marginal cost is negligible, and a single new system on a route can collapse pricing for everyone already there.
| Feature | Consequence |
|---|---|
| Very high fixed cost | Large capital at risk before any revenue |
| Near zero marginal cost | Price competition can go to nearly nothing |
| Lumpy capacity additions | One new cable can oversupply a route |
| Upgradeable terminal equipment | Capacity grows without new cable |
A route with two cables and demand for one has no stable price, because either owner can profitably undercut down to nearly zero. That is why capacity on competitive routes has fallen by orders of magnitude and why nobody builds one alone.
The Consortium Model
The traditional structure has a group of telecommunications carriers jointly funding a system, each contributing capital and receiving capacity in proportion.
The logic is straightforward risk sharing. No single carrier bears the full construction cost, each obtains the capacity it needs, and because the participants are the natural buyers, the risk of building capacity nobody wants is reduced.
The arrangement is unusual in that direct competitors jointly own an asset and then compete using it, which requires careful governance and has generally been accepted by competition authorities because the alternative is that fewer cables get built.
What the Content Companies Changed
The most significant development in the sector is who is now funding construction.
Large technology companies moving enormous volumes of traffic between their own data centres concluded that buying capacity from carriers was more expensive than building their own systems. Several now own cables outright or hold majority stakes in consortiums, rather than participating as customers.
Their economics differ fundamentally from a carrier. A carrier builds a cable to sell capacity and needs the route to be profitable on its own. A content company builds it to reduce its own network costs and to control latency and routing, and the return appears as avoided cost elsewhere in its business.
That means they can rationally build capacity that would not make sense for a carrier, which has expanded supply on major routes and further pressured the merchant capacity market.
The Routes Nobody Wants to Build
The reverse problem exists on thin routes. Small island nations and some developing markets generate insufficient traffic to justify a dedicated system, and they end up dependent on a single cable or on expensive satellite backhaul.
A single cable means a single point of failure. Several countries have experienced multi day national internet outages from one cable break, since restoration requires a repair ship to reach the location, retrieve the cable from the seabed, splice it, and return it, which takes weeks depending on distance and weather.
Development finance institutions have funded systems on exactly these routes for this reason, treating connectivity as infrastructure rather than as a commercial project.
The Vulnerability
Cables are broken regularly, mostly by fishing gear and ship anchors in shallow water, with damage concentrated near shore where the seabed is busiest.
Redundancy is the normal defence, since traffic reroutes automatically over other systems if enough alternative capacity exists. That works on well served routes and fails on thin ones.
Deliberate damage has become a live policy concern, and cables have attracted attention as strategic infrastructure with limited physical protection. The practical response has been diversifying routes and landing points rather than protecting the cables themselves, which is not realistically possible across thousands of kilometres of seabed.
How to Read the Business
For anyone analysing the sector, the useful distinctions are between owning wet plant, which is a capital intensive infrastructure business, and selling capacity, which is a commodity business with brutal pricing; and between routes with multiple systems, where pricing is competitive, and thin routes, where a single owner has genuine pricing power.
Landing stations and the terrestrial backhaul from them are frequently better businesses than the cable, because they are local chokepoints with fewer alternatives, which is a recurring pattern in infrastructure generally.
The Bottom Line
Submarine cables are the physical layer under everything described as the cloud, and their economics are dominated by enormous fixed costs meeting near zero marginal costs, which is why competitors build them together and why capacity prices collapse whenever a new one lands. The most consequential recent change is that the largest buyers became the largest builders, which altered the return requirement on new systems and left carriers competing against participants who do not need the route to be profitable at all.