“So Ingrid,” Sam asked, “how exactly will you convince your editor that submarine cables are relevant to The Cloud?”

We were maybe still in New Mexico or somewhere in Kansas. It was a night drive. Time moves weirdly during night drives. All roads basically become the closing moments of Terminator 2. Whenever it was and wherever it was, it was apparently a good time for my driving partner to pose questions about some of the stories I had lined up for this series, including one exceptionally long piece about a submarine cable (which, dear reader, will run later this week).

“Well,” I replied, “clouds are just evaporated molecules of water that emerge from larger bodies of water. Oceans are bodies of water. It’s...it’s relevant.”

Look, I was pretty tired. But if there is a case to be made for placing submarine cables within the landscape of The Cloud, it's more a case for historical continuity and resonance. Cloud infrastructure is a landscape of interdependent systems, submarine cables among them.

Submarine cables don't come up in the news that often, but if they do it seems to be in two forms: short articles reminding everyone that the Telegeography Submarine Cable Map exists, and short articles of hand-wavey reminders that submarine cables are vulnerable to harm (from tectonic plates, ship anchors, sharks, and terrorists, among others).

While these are totally valid topics to explore, I often find these stories lacking in context about the various systems, geographies, and politics that shape submarine networks. While there are lots of other super compelling aspects of submarine-cable law and policy (says the person who owns a copy of Submarine Cables: The Handbook of Law and Policy), here are two questions that might help readers take in Telegeography takes with a little less gee-whiz and a little more clarity.

Who Owns Submarine Cables?

For the most part, Internet submarine cables are built by consortia of international companies. The consortium method has the advantage of distributing risk on submarine-cable projects, which can range in cost from the hundreds of millions to the low billions. When they first emerged, during the postcolonial Cold War era, these consortia were made up of telecoms from different countries forming agreements with each other. Ownership and maintenance of the cables was aligned along these landing sites—the parts of the cable that ran through the U.S. might belong to AT&T; the parts running through the U.K., maybe to BT.

In the late 1990s as the Internet became increasingly privately held, venture capital began participating in these consortia as well. Perhaps the most famous early example of this shift was the FLAG network, the subject of Neal Stephenson's essay Mother Earth Mother Board” (arguably the Book of Genesis in the Internet-infrastructure nerd bible). Today, many kinds of non-telecoms invest in submarine cables, including companies like Google and Facebook, who both own stakes in Pacific submarine cables.

Military agencies also build submarine cables, which do not appear on public maps. While these cables have recently been the subject of much kerfuffle over whether they are vulnerable to Russian sabotage (which, while probably a possibility, featured such so much over-the-top panic in its speculations that it basically was only missing trained Russian spy sharks), military telecommunications networks have played a crucial but not extensively reported role in legal challenges to U.S. and U.K. drone-strike operations in Yemen, Somalia, and Pakistan.

In 2014, the U.K.-based organization Reprieve filed a complaint with the British government against BT for its role in facilitating drone strikes based on a submarine cable. In 2012, BT accepted a contract from the Defense Information Systems Agency (DISA) to connect a military base in the U.K. to a military base in Djibouti. The court threw out the complaint, ruling that Repreive presented insufficient evidence that the cable was explicitly used in drone strikes.

More recently, a Somali man brought a legal complaint to the German government over the role that the Ramstein Air Force base played in a drone strike that killed his father. Ramstein isn't a landing point for a submarine cable, but for a satellite uplink that connects directly to drones in Yemen, Somalia, and elsewhere. A combination of submarine and terrestrial cables connects the uplink to Creech Air Force Base in Nevada. While the Ramstein complaint is more about the uplink than the fiber, it and the Reprieve complaint are reminders of the scale of military network infrastructure, and the significance of that infrastructure in contemporary warfare.

Who Builds Submarine Cables?

For much of the history of submarine cables, many telecommunications companies were responsible for their own cable design and manufacturing, as well as fleets of cable ships (as seen in this AT&T video about their submarine-cable operations). After the first tech bubble collapsed in the late 1990s, many of the large telecoms divested themselves of their submarine-cable operations, spinning them off into separate companies. A number of these formerly telecom-owned companies act as “turn-key” suppliers—they manufacture the cable, survey routes and obtain relevant permits, maintain fleets of cable ships, and perform maintenance and repairs. TE SubCom, the current incarnation of what used to be AT&T Submarine Systems, explains what they do in this thrilling corporate video (from that video, I strongly encourage you to go down the rabbit hole that is corporate submarine-cable videos, it will probably transform your life forever).

How Do Submarine Cables End Up Where They End Up?

The design of submarine-cable routes and the decision of where to place them is a lengthy, complex calculation of bathymetry, geography, local land-use laws, international law, and the law of the sea. For a long time it was also a calculation heavily shaped by legacy—a lot of Internet cables run along similar routes and to similar sites as their telegraph predecessors, to major coastal cities throughout the world. There is an assumption that these routes are reliable and secure, as they've been tried and maintained in the past.

The history of modern telecommunications development traces the expansion of empire. The British telegraph submarine-cable system the All-Red Line, for example, is kind of just a cool map of Places Britain Used To Own. Today, a number of the landing sites on the All-Red Line remain landing sites for submarine cables.

As Nicole Starosielski documents in her book The Undersea Network (if we're being cute, the Book of Jonah in the Internet-infrastructure nerd bible), this pattern of submarine cables following empire came to reflect the new mode of empire: globalization. During the Cold War, landing sites began to shift away from central port cities to more rural areas, recognizing that decentralizing communication networks was a far more viable security strategy in the face of nuclear war. These more remote sites were sometimes chosen purely for logistical reasons—such as the Manchester cable station, which was chosen as the site for the first continental U.S.-Hawaii cable because it's literally the shortest distance away from Hawaii.

As ownership models moved away from strictly telecom consortia, new landing sites were built and selected because the incumbent telecom companies didn't want the new consortia to build in their territory, as it were. Today's submarine-cable networks are increasingly governed by logistics, bottom lines, and latency, rather than colonial control or Cold War paranoia. But they remain useful objects for understanding some of the public and private power dynamics that play across the network at the infrastructural level, literally across oceans.