Global subsea cable network

What it is

Subsea cables are fiber-optic cables laid on the ocean floor, connecting continents and countries with strands of glass that carry light-encoded data pulses. As of early 2026, TeleGeography counts more than 600 active or planned systems spanning roughly 1.8 million kilometers. They carry over 99% of international internet traffic; satellites handle real-time voice and critical military links but remain too slow and expensive for bulk data at scale.

The main players are cable-owning consortiums of telecommunications companies, national carriers, and, increasingly, US hyperscalers. Google, Meta, Microsoft and Amazon became major investors in new cable from roughly 2016, pursuing dedicated capacity and route control rather than sharing bandwidth in traditional consortiums. On the supply side, a small number of manufacturers build most new systems: Alcatel Submarine Networks (a Nokia subsidiary based in France) and SubCom (a US firm). A separate, very thin fleet of purpose-built cable repair ships, roughly 40 to 50 vessels worldwide, performs all maintenance.

Governance sits with the International Cable Protection Committee (ICPC), founded in 1958, which has 254 members from 70 countries representing 98% of the world's subsea cables. In November 2024, the ITU and ICPC jointly established the International Advisory Body for Submarine Cable Resilience to coordinate government and industry responses to a growing threat environment.

History

The first transatlantic telegraph cable was completed in 1858, though it failed within weeks; a durable copper connection followed in 1866. The first transatlantic telephone cable, TAT-1, went live in 1956. Fiber-optic submarine cables began replacing copper in 1988 with TAT-8 across the Atlantic, offering orders-of-magnitude more capacity. The commercial internet boom of the 1990s triggered a major construction wave, followed by a glut and bust in the early 2000s. Investment resumed steadily from 2010, and the hyperscaler-led era accelerated from around 2016, with capacity doubling roughly every three to four years on high-demand routes. The MAREA transatlantic cable, a Microsoft-Meta project commissioned in 2017, carries 224 terabits per second, illustrating how far capacity has advanced since the kilobit-per-second era.

Current state

As of mid-2026, the system faces escalating physical risk. The ICPC tracks 150 to 200 cable faults per year, requiring roughly three repairs per week worldwide. Fishing gear and ship anchors cause about 80% of faults; natural events such as earthquakes cause most of the remainder, and deliberately caused cuts remain rare but are growing as a documented concern. The Red Sea corridor concentrates east-west capacity in waters where Houthi attacks since November 2023 have left abandoned ships that can drag anchors, and contested access slows repair permitting. The global cable repair fleet is already stretched thin, meaning each incident now takes materially longer to fix than it would have a decade ago.

Relationships

The Red Sea is the system's most acute current chokepoint. The PEACE cable cut in June 2026, and a simultaneous four-cable anchor-drag at Bab el-Mandeb, disrupted internet access across South Asia, the Gulf, and parts of Africa. The Aletar cable cut in the eastern Mediterranean in June 2026, repaired within 48 hours, showed how thin national redundancy can be for states reconnecting after conflict; Syria's entire internet depended on a single 1997-vintage 5 Gbps cable. At the Gulf's eastern end, Iran's moves to threaten cables under the Strait of Hormuz in 2026 extend coercion from shipping to data, directly threatening Gulf states' AI infrastructure investments. Near Taiwan, China-crewed vessels have repeatedly cut Matsu island cables, and a Taiwanese court convicted a Chinese captain in June 2026 for intentional damage, the first criminal verdict for cable sabotage in the strait.

What to watch

Whether the ITU-ICPC advisory body produces binding rules on cable-zone protection and repair-ship access, particularly for the Red Sea and Taiwan Strait. Whether hyperscalers accelerate investment in alternative routes, via the Arctic or around southern Africa, to reduce exposure to the Red Sea-Hormuz-Malacca corridor. Progress on expanding the global repair fleet, where lead times for new purpose-built vessels run to five years or more. Any formal legal framework for states to restrict or toll cable transit, which Iran has signaled and which would set a precedent with global implications.