The Vulnerable Cable Beneath the Connected World

Infrastructure and restoration note, 24 October 2026.

When four submarine cables failed off West Africa in March 2024, the damage did not stay on the ocean floor. Thirteen African countries saw degraded service or near-total internet outages. Côte d’Ivoire lost almost all internet connectivity. In other places, providers rerouted traffic onto surviving cable systems and overland fiber. The physical break was shared; the interruption each household felt depended on where its provider could send traffic next. Internet Society’s incident report tracks the uneven effects.

That distinction is the starting point for understanding submarine cables. They carry more than 99 per cent of international data flows, according to the International Telecommunication Union, but a cable is one section of a much larger network. The service at a home or business also needs power, a local internet provider, terrestrial fiber, a landing station, data centers and functioning routes to the service being used. A cable fault can slow or interrupt that chain; it does not automatically unplug a continent. ITU’s 2026 resilience report describes the cables’ role in banking, education, health, government and industry.

The map of this infrastructure is large, but it is not evenly redundant. ITU’s 2025 State of Broadband in Africa showed that by 2021 South Africa was connected to eleven submarine systems, while Kenya, Nigeria and Cameroon each had six. Landlocked countries depend on terrestrial connections to reach coastal landing points. Since then, new projects have added capacity, but the same structural questions remain: how many independent paths reach a country, where those paths converge, and whether the domestic networks beyond the landing station can carry the traffic if a route fails. ITU’s Africa report lays out those differences without treating the continent as one network.

Capacity is a route, not a headline number

Cable capacity is often quoted in terabits per second, but installed capacity is not the same as what a local operator can buy, reach or reroute in an outage. ITU counted more than 500 operating systems and about 1.4 million kilometres of submarine cable in 2025; bandwidth was growing 22 per cent a year. A provider may still have little spare capacity on a surviving route. If several systems share the same coast or terrestrial backhaul, counting them as separate cables can exaggerate resilience. ITU’s 2025 network summary gives the global scale.

The March 2024 West Africa outage showed this concentration problem. The affected systems were ACE, SAT-3, WACS and MainOne. Internet Society reported that the four converged near Côte d’Ivoire, creating a common exposure; the cause was described as a suspected underwater rockslide, not an attributed attack. Google’s Equiano cable and a separate MoovAfrica route remained available, and operators rerouted traffic, including over terrestrial fiber. The Equiano route saw a fourfold traffic increase. That preserved service for some networks but did not prevent a near-total outage in Côte d’Ivoire. The incident report describes the redundancy and its limits.

The sequence also shows why household and business loss is hard to reduce to one outage number. A family may be able to call but not load a school portal; a bank’s local app may still work while its cloud authentication service fails; an exporter may reach local customers but lose access to a foreign order system. In March 2024, Microsoft reported disruptions to its EMEA services and restored them after rerouting. For firms, payment approvals, cloud tools, customer support and logistics can fail independently. These are the practical paths from cable capacity to lost working time, sales and access—not proof that all digital services stop at once.

Damage has causes, but damage is not attribution

Submarine cable faults are routine maintenance events. The International Cable Protection Committee estimates roughly 150 to 200 telecommunications cable faults worldwide each year. It says 70 to 80 per cent arise from accidental human activity, most often fishing and ship anchors; natural hazards and technical failures make up much of the rest. Sabotage is a risk, but ICPC says it accounts for less than one per cent of global cable damage annually. The committee also says it does not investigate or confirm the cause of an individual fault. Its public FAQ and statistics are a useful brake on speculation.

In East Africa, two systems—SEACOM and EASSy—were damaged off KwaZulu-Natal on 12 May 2024. Internet Society described a dragging anchor as the suspected cause, not a proven finding. It reported that the cables were repaired by 3 June, roughly three weeks after the damage. Several East African countries experienced reduced connectivity. The outage was serious, but the account supports a suspected accidental cause, not a finding of deliberate interference. The East Africa outage report records that distinction.

The Red Sea case the previous February was harder to interpret. HGC Global Communications said four of more than fifteen cables had been damaged and that it began rerouting affected traffic. Its statement did not identify the cause. The damage occurred during a period of attacks on commercial shipping, and public claims connected it to the conflict, but timing alone cannot prove who or what caused a cable fault. HGC’s 29 February 2024 notice confirms damage and mitigation, not an attacker.

That caution matters again this week. On 21 October, Iran’s Press TV quoted an IRGC deputy commander threatening to target undersea cables in the Strait of Hormuz if Iranian infrastructure were attacked. The statement confirms a conditional threat was made; it is not evidence that a cable has been attacked or that any particular service has failed. Nor does it prove that the speaker’s estimate of regional cable numbers is accurate. The ICPC describes Hormuz as a regional connectivity corridor and says bandwidth passing through it is less than one per cent of international bandwidth worldwide. The interview was carried by Press TV; ICPC’s clarification places the global-scale claim in context.

Repair time is part of resilience

After a cable breaks, an operator has to identify the fault, secure access, mobilize a specialized repair vessel, locate and retrieve the damaged section, splice in replacement cable and test the restored system. This is not a dive-and-patch job. Repair vessels are limited, and a ship may be on the other side of an ocean or already committed to another repair. Weather, seabed conditions, customs clearance, coastal permits and cabotage rules can add delay. ITU’s July 2026 advisory-body report says repair times have increased in some regions and identifies permitting, repair-ship availability and geographic concentration as resilience challenges. The ITU report recommends quicker repair approvals, better coordination and more route diversity.

Tonga’s 2022 experience puts a human scale on the delay. A volcanic eruption severed the country’s international link and damaged a separate domestic cable. International internet service returned after a repair vessel reconnected the main cable in February, while the domestic island link took until July 2023 to be repaired. One shared cause therefore produced different restoration timelines for the capital and outer islands. The National Oceanic and Atmospheric Administration’s Global Volcanism Program records the international repair and restoration date; Tonga Cable and local reporting document the longer domestic repair. The Smithsonian/USGS Global Volcanism Program timeline shows the international repair period.

The African data economy depends on what survives

Africa’s cable buildout has brought more capacity and lower-latency links than satellite-only connectivity could provide. It has also made redundancy a development question. A bank, hospital, school or small online business can have an excellent local network and still rely on international transit for software updates, identity checks, cloud hosting, payment authorization or customers abroad. A fault can raise latency and cost, interrupt services or force a provider to ration the capacity that remains.

In the March 2024 West Africa outage, some IXPs remained operational, local content stayed reachable and traffic shifted to alternatives. In East Africa, multiple cable systems and cross-border terrestrial links gave countries different options. Resilience is therefore not just “more cable.” It includes inland fiber, IXPs and locally hosted services, spare capacity on independent paths, reliable power, clear repair permissions, transparent service contracts and public information about restoration. The right household question is not how many cables appear on a map; it is whether a service can still reach its users if one route fails.

That is also a question of contracts and accountability. Governments can shorten repair delays by pre-agreeing permits, coordinating customs and designating a single point of contact. Operators can publish outage notices, affected services, rerouting status and realistic repair windows without exposing sensitive infrastructure details. Firms can test business continuity against loss of a landing region or cloud zone, rather than assume a backup route exists. Households need accurate information about which services are degraded and where alternatives remain. The ITU’s 2026 advisory body calls for coordination between governments and cable operators, faster deployment and repair, risk monitoring and geographic diversity.

A cable is vital, but it is not a magic switch for all connectivity. Most faults are accidental; some causes remain unknown; intentional sabotage is possible but cannot be inferred from a headline or a military threat. Redundancy can keep some services running, while shared routes and slow repairs leave others disconnected. For Africa’s digital economy, resilience means building alternate paths and making repair work possible before the next cut—not predicting an attack from the fact that the cable is important.

Sources

• ITU, International Advisory Body final report, 10 July 2026
• ITU, State of Broadband in Africa 2025
• ITU, Global Connectivity Report 2025: submarine cable statistics
• ICPC, cable fault causes, repair and attribution
• Internet Society, West Africa cable outage, 14 March 2024
• Internet Society, East Africa cable outage, 12 May 2024
• Nigeria Communications Commission statement, March 2024
• HGC statement on Red Sea cable damage, 29 February 2024
• Smithsonian/USGS Global Volcanism Program: Tonga eruption and cable repair timeline
• Press TV, IRGC deputy statement, 21 October 2026

The cable-laying vessel Île de Bréhat at Ogden Point, Victoria, on 2 October 2024. The dated harbor photograph identifies the specialist ship type; it does not show a current repair mission. Credit: Gordon Leggett / Wikimedia Commons. Image source · CC BY-SA 4.0 International