Cloud services have a seabed. More than 99 per cent of international data traffic is carried by submarine telecommunications cables, according to the International Telecommunication Union. These systems cross oceans, land at carefully chosen coastal sites and are repaired by specialised vessels when faults occur.
Eventually, a cable reaches the end of its commercial service. Perhaps traffic has moved to a newer system. Perhaps the route or capacity no longer fits the network. The signal stops—but the physical cable remains.
What follows is not governed by one universal rule. An out-of-service cable may be left in place, recovered in part, recovered over a longer section, repurposed or dealt with when a new project needs the route. The right choice depends on permits, seabed conditions, other users, the cable’s construction and the disturbance a recovery operation could cause.
A cable is both infrastructure and material
Modern telecommunications cables vary by depth and location. Deep-ocean sections can be relatively slim; sections near shore may be armoured for protection against anchors, fishing activity and other hazards. A cable can combine optical fibres, conductive metal, protective polymers and steel armour.
That mix creates potential value in recovery, especially for metals. It also creates work: a cable must be located, lifted, handled on a vessel, landed, transported and separated into useful fractions.
The material case cannot be separated from the marine operation.
Why leave a cable in place?
Once correctly installed, a telecommunications cable is generally immobile and produces no operational emissions into the surrounding environment. US NOAA guidance describes the environmental effects of fibre-optic cables as generally modest, with the most significant risks associated with installation activities such as trenching, vessel movement and disturbance of the seabed.
Recovery can create another period of physical disturbance. In a sensitive habitat, or where a cable has become buried, the impact of pulling it out may outweigh the benefit of removing it. Old cables can also lie near active infrastructure, making intervention more complex.
“Leave in place” should not mean “forget.” Accurate records matter for navigation, fishing, future construction and marine spatial planning.
Why recover one?
An out-of-service cable can obstruct a future route, create snagging risk if exposed, or contain materials worth recovering. The International Cable Protection Committee has long maintained recommendations on out-of-service cables and route clearance.
In 2025, the ICPC highlighted a peer-reviewed assessment of environmental considerations around cable recovery. Its central message was not that every cable should be removed. It was that responsible recovery can be environmentally proportionate, while more consistent monitoring and reporting are still needed.
This is a useful pattern for infrastructure at sea: evidence should decide the intervention, not a blanket preference for either removal or abandonment.
How recovery works
A recovery vessel first needs a reliable route position and a plan that accounts for water depth, burial and nearby assets. A section may be located and brought to the surface using specialist equipment. The operation can proceed in segments, and full recovery may not be possible or desirable across an entire route.
Once landed, the cable becomes a dismantling and materials problem. Steel, copper and polymers have different processing routes. Whether a fraction is genuinely recycled depends on its condition, separation and the available market—not simply on the fact that it reached shore.
The decision is local, even when the network is global
Environmental impact varies with substrate, habitat, burial method, water depth and the scale and duration of activity. NOAA’s account of cables in the Olympic Coast National Marine Sanctuary shows why route planning and post-installation monitoring matter: impacts and recovery differed with local seabed conditions, and granular habitats recovered relatively quickly in the observed areas.
That does not make all routes low-impact. It shows why generic claims fail. The same recovery method can have different consequences on mobile sand, rocky habitat or a biologically sensitive seabed.
Decommissioning is not the reverse button on installation. It is a new marine project with its own purpose, evidence and trade-offs.
What better end-of-service planning looks like
A credible plan begins before the cable goes quiet. It records materials and route condition, defines who remains responsible, anticipates conflicts with other seabed users and sets criteria for partial or full recovery. It also states what will happen to recovered material.
As the cable network grows, these questions move from the edge of digital infrastructure to its centre. The internet’s material footprint is not only the energy used by data centres. It also includes the long, engineered objects connecting them beneath the sea.
Sources & further reading
- ITU, Strengthening the world’s digital backbone, 2025.
- International Cable Protection Committee, Submarine Cable Protection and the Environment, including the April 2026 issue on cable recovery.
- NOAA, Submarine Cables — Domestic Regulation and environmental impacts.
- NOAA Office of National Marine Sanctuaries, Submarine Cables in Olympic Coast National Marine Sanctuary.
Editorial note: Telecommunications and power cables have different designs and operating effects. This article focuses on telecommunications cables.