Subsea cables. The ships that repair them, a small and ageing fleet of purpose-built vessels shared across every ocean, decide how quickly the fibre-optic lines that carry data between continents come back into service after a break. No commercial actor has reason to enlarge that fleet. Since 2021 the United States (US) and its allies have squeezed China's leading cable manufacturer out of Western projects, splitting the firms that build the global network into two technology camps. Exclusion has proved the cheap half of decoupling, the deliberate reduction of technological interdependence. Capacity is the dear half, and no one has been willing to pay for it.
Two industries sit beneath this argument, and only one has split. Manufacture is concentrated: America's SubCom, France's Alcatel Submarine Networks (ASN), and Japan's NEC held roughly 87% of the market in 2021, with China's HMN Technologies on a further 11%. Repair works differently, through cooperative zone agreements such as the Atlantic Cable Maintenance Agreement (ACMA) and its Asian equivalents, which pool ships across dozens of cable owners and contract them to a few specialist operators, among them Britain's Global Marine, France's Orange Marine, and the state-controlled Chinese firm S.B. Submarine Systems (SBSS). That pooling is where the ripple runs: an operator with no stake in the US-China contest still buys restoration from the shared fleet, and SBSS repairs cables owned by Google and Meta.
The Split Stops at the Waterline
Exclusion has proven both real and cumulative. On 17 December 2021, the US Commerce Department placed HMN Technologies (formerly Huawei Marine) and three related Chinese cable manufacturers on its Entity List, on the stated grounds of acquiring US-origin items in support of the People's Liberation Army's modernization initiative, one anchor of Washington's 'Clean Cable' agenda aimed at removing Chinese vendors from trusted digital infrastructure. As Winston Qiu has documented, the Federal Communications Commission (FCC) has revoked authorisations for Chinese carriers, established a presumption of denial for foreign-adversary affiliates, and barred covered equipment, such as HMN's, from any new system landing in the United States. The exclusion took effect quickly: by 2023, HMN had supplied 18% of the cable length that came online in the preceding four years but was due to build only 7% of the length then under development, according to TeleGeography data reported by Reuters in March 2023.
None of these instruments addresses repair: Entity Listing restricts the flow of US-origin goods and services to the listed firms, and FCC rules govern licensing and equipment for systems landing in the United States; neither touches maintenance of systems at sea. In a letter published on 21 July 2025, three House chairmen wrote to the chief executives of Google, Meta, Microsoft, and Amazon noting that Chinese entities “have continued to provide maintenance or servicing” to cable systems in which US firms hold interests. Decoupling has split manufacturing; at least until mid-2025, the repair layer remained vendor-agnostic, though there is now visible pressure in Washington to end that neutrality.
The Fleet Was the Weak Point First
The geoeconomic contestation over subsea cables predates US-China decoupling, and so does the constraint that now limits it. Roughly 62 cable-laying and maintenance vessels operate globally (only about 19 contracted to maintain existing lines), with maintenance ships averaging roughly 20 years of age, per industry analysis, and new construction lagging retirements. The network suffers roughly 150–200 faults a year, 70–80% of them accidental (fishing gear and dragged anchors), while data derived from the International Cable Protection Committee (ICPC) indicate that average repair times have more than doubled since 2012, from under 20 days to over 50, with permitting friction cited as the key culprit. Sustaining current service levels will require roughly $3bn of investment, entailing 15 replacement cable ships and five additional ones, per TeleGeography's The Future of Submarine Cable Maintenance.
The Baltic Sea incidents that pushed cable security onto the national-security agenda, the severing of the C-Lion1 and BCS East-West Interlink cables in November 2024 and the damage to Estlink 2 and four telecoms lines that December, illustrate exposure to this constraint, whatever their cause. According to the National Security Overview 2026 of Finland's Security and Intelligence Service (Supo), presented on 10 March 2026, cable damage has regularly occurred in the Baltic throughout the 2000s; the recurrence of incidents permits no conclusion about their intentionality, and poorly maintained shadow-fleet vessels raise the risk of accidents. Sabotage or accident, restoration runs through the same limited fleet.
Demand is now colliding with that limited capacity. The value of new cables planned to enter service between 2026 and 2029 exceeds $16bn, per TeleGeography's April 2026 update, with hyperscalers (the largest cloud and platform firms) driving the surge, including multiple Google- and Meta-led trans-Pacific systems.
Governments have begun to respond. France completed an 80% state acquisition of Alcatel Submarine Networks from Finland based Nokia on 31 December 2024, positioning itself as strategic owner of a critical cable-systems vendor. On 5 February 2026, the European Commission introduced its Cable Security Toolbox, accompanied by a €347m Connecting Europe Facility allocation to strategic cable projects. A €20m call earmarked for repair is intended to fund modules and equipment stationed at ports or shipyards, not the ageing fleet itself.
As new fronts open in the US-China confrontation, they sharpen a deficiency that had been tightening for a decade before the first entity listing. For practitioners, the implication is 'lock-in': vendor and route choices now carry geopolitical weight, and repair-timeline risk hinges on a common fleet that policy could yet divide. The signals to watch are whether new purpose-built cable-ship orders are placed (and by whom, sovereign or commercial), whether repair times in Western repair zones lengthen, and whether Chinese repair vessels are formally excluded from Western repair agreements. None of the three has yet turned. The asymmetry is the point: the network has been divided at the point of manufacture, where substitution is possible, and left undivided at the point of repair, where it is not.
