Internet still relies almost entirely on the seabed
The idea that the Internet has become largely wireless is misleading. Satellites, mobile networks and Wi-Fi are highly visible in the way we use the Internet, but communications between major regions of the world still rely primarily on fiber-optic cables installed on the seabed.
More than 99% of international data traffic currently travels through this infrastructure. The global network includes more than 500 operational submarine cable systems, representing around 1.4 million kilometers of connections, according to data compiled by the International Telecommunication Union.
These cables do not simply carry web pages, videos or messages exchanged between individuals. They also support the international operation of cloud services, financial transactions, business and government communications, as well as countless digital services that have become dependent on the continuous exchange of data between countries and continents.
Their importance is increasing as data volumes grow. The International Telecommunication Union estimates that demand for bandwidth across this infrastructure has been growing by around 22% per year since 2021. The growth of cloud computing, artificial intelligence and digital services is therefore not reducing this physical dependency. On the contrary, it is increasing the amount of data the network must carry.
Submarine cables are already damaged several times a week
This dependency does not mean that every cable break becomes a crisis. Damage to submarine cables is already part of the normal operation of the global network. More than 170 repairs were reported in 2025, an average of nearly four interventions per week.
Most of these incidents are not acts of sabotage. Ordinary human activities remain the main cause, including ship anchors, fishing and other operations carried out near cables. Earthquakes, underwater landslides, currents, storms and equipment failures can also cause damage.
This distinction matters. Several recent incidents have attracted attention because of the geopolitical context in which they occurred, but a damaged cable does not, by itself, prove that the damage was intentional. The cause of each incident must be established separately.
Most cable breaks go unnoticed by users
If several cables can be damaged each week without Internet access disappearing, it is because the network is not built around a single connection. Multiple routes connect the world's major regions, and operators can reroute traffic when a cable becomes unavailable.
This redundancy explains why an isolated cable break can be almost invisible to the end user. Data takes another route while the damaged cable is located, recovered and repaired. Depending on the region affected and the alternative capacity available, however, an incident can still cause slower connections, increased latency or more significant disruptions.
Not all regions have the same level of protection. Some territories depend on a limited number of cables or landing points, while certain maritime areas concentrate several major international routes. A cable break therefore does not have the same consequences everywhere.
The problem is no longer simply the loss of a single cable
This concentration is what can turn an ordinary technical incident into a strategic risk. When an alternative route exists, the network can absorb the failure. When several connections pass through the same area, are damaged simultaneously, or the remaining capacity becomes insufficient, the situation changes.
Organizations responsible for cable resilience are now paying as much attention to route diversity as to the physical protection of the cables themselves. The goal is to prevent a single event from eliminating several routes at once and to preserve enough capacity elsewhere to maintain communications.
The Internet therefore does not depend on a single cable that could simply be cut to bring the network down. Its resilience comes precisely from the number of available connections. But that resilience has limits: the more disruptions are concentrated along the same routes, the more the network loses the alternative paths that normally allow it to absorb failures.
What would happen if several cables were cut at the same time?
Internet access would not disappear everywhere at once
Even if several cables were disrupted simultaneously, the first effect would probably not be an immediate, worldwide Internet outage. As long as alternative routes remain available, operators can shift part of the traffic onto cables that are still operational.
But this ability to reroute traffic is not unlimited. The remaining connections also have finite capacity. As more routes become unavailable, more traffic has to be concentrated onto fewer paths.
The consequences would therefore depend heavily on which cables were affected, their capacity, their location and the routes still available. Several disruptions spread across very different routes could be absorbed more easily than the simultaneous loss of several cables crossing the same strategic corridor.
Multiple cable disruptions have already shown how this works
This scenario is not entirely theoretical. In 2024, several incidents simultaneously affected cables serving Africa, the Middle East and connections between Asia and Europe. In the Red Sea, the loss of several systems affected a significant share of the capacity used along these routes.
The Internet did not stop working. Some traffic was rerouted through infrastructure that remained available, but the disruptions showed the difference between losing a single cable and losing several connections within the same corridor. Operators had to modify their routes, and some connections experienced degraded performance or increased latency.
Other multiple cable disruptions off the African coast that same year produced very different effects depending on the country. Networks with more alternative routes recovered more quickly, while territories relying on a limited number of connections were more exposed.
The location of the disruptions matters as much as their number
There is therefore no universal number of cables that could simply be cut to cause a widespread outage. The global network does not consist of ten, twenty or fifty interchangeable connections of equal importance.
Some areas concentrate large numbers of cables connecting major economic regions. Straits, narrow seas and areas close to certain landing points can therefore become particularly sensitive: several physically separate systems may follow relatively similar routes.
A series of disruptions concentrated in such an area could therefore have greater consequences than a larger number of cable breaks spread across different parts of the world. This is why geographic route diversity is now considered a key element of network resilience.
Repair capacity would itself become a critical issue
Another limitation appears after a cable is damaged: a submarine cable cannot be repaired remotely. The fault must be located, a specialized vessel must be mobilized and sent to the area, the ends of the cable must be recovered, the repair completed, and the connection returned to service.
Under normal conditions, these operations gradually restore lost capacity. But if several cables were damaged within a short period of time, the human, technical and maritime resources required for repairs would be needed simultaneously.
The issue would then no longer be simply whether traffic could be rerouted, but how long the remaining infrastructure could handle the additional load before the damaged connections were restored.
A massive disruption would first cause an uneven degradation of the network
The most credible scenario is therefore not an on-off switch causing the Internet to disappear everywhere in the world at the same moment. A sufficiently severe disruption would instead produce a sequence of effects: traffic rerouting, saturation of some connections, increased latency, degradation of certain services and, in the most dependent territories, partial or longer-lasting losses of international connectivity.
As available submarine capacity decreased, the way remaining capacity was allocated would also become increasingly important. Not all uses could necessarily maintain the same level of performance at the same time if the remaining routes no longer had sufficient capacity.
In an extreme scenario, the loss of enough strategic routes could temporarily fragment parts of the global network or leave some territories severely isolated. But setting a universal threshold would be misleading: the severity would depend primarily on the routes affected, the capacity still available and the time required to restore the connections.
France and Europe: where are the key dependencies?
France lies at the crossroads of several major digital routes
France is not located at the edge of the global network. Its Atlantic coast connects it directly to transatlantic routes, while its Mediterranean coast provides a gateway to Africa, the Middle East and Asia.
This position gives France several possible routes for its international communications. This is an advantage in terms of resilience: when one connection becomes unavailable, the existence of alternative routes and multiple landing points increases the possibilities for rerouting data.
But the geography of the network also means that some areas are far more important than others. Cables do not reach the entire French coastline evenly. They converge at landing points that are then connected to major terrestrial networks and European data centers.
Marseille has become one of the world's major digital hubs
This is particularly visible in Marseille. Its location on the Mediterranean makes it a natural gateway for cables connecting Europe to North Africa, the Middle East and Asia. With the arrival of the Medusa system, 17 submarine cables now have a landing point there.
This concentration has transformed Marseille into a major interconnection hub. Data arriving through the Mediterranean can then reach local data centers and major European hubs, including Paris, London and Frankfurt.
In June 2026, a new section of the Medusa system between Marseille and Bizerte, Tunisia, became operational. Its purpose is not simply to add capacity: this new route is also intended to strengthen the diversity and resilience of connections between Europe and North Africa.
Adding more cables is not enough if their routes remain concentrated
The number of available cables alone does not determine the resilience of a territory. Several connections may follow nearby maritime corridors, reach the same landing area or subsequently depend on shared terrestrial infrastructure.
Real resilience therefore also depends on geographic diversity. Two cables following very different routes provide more options for bypassing a localized disruption than two connections exposed to the same maritime passage or concentration point.
This is one reason why European policies no longer focus solely on adding new capacity. They also aim to diversify routes, identify shared dependencies and strengthen infrastructure whose loss could affect several communications simultaneously.
Europe now treats submarine cables as critical infrastructure
This dependency has gradually changed the way the European Union views the submarine network. Cables are no longer treated solely as telecommunications infrastructure: their security is now integrated into European policies for the resilience of critical infrastructure.
The European Union has since strengthened its strategy for submarine cables, with new projects designed to diversify routes, improve monitoring and increase repair capacity. Regional mechanisms are also being developed in the Baltic and Mediterranean seas to improve coordination when submarine infrastructure is affected.
France is well connected, but still depends on the global network
The number of connections and France's geographic position therefore provide significant resilience. However, this does not make the country independent from the rest of the network. Some French communications travel through infrastructure located far beyond national waters before reaching their destination.
A cable therefore does not need to be damaged off the French coast for a route used from France to be affected. A disruption in the eastern Mediterranean, the Red Sea or another part of an international cable system may force operators to change the path taken by data.
This is what makes submarine cables a global issue: a country can protect its landing points and expand its own connections, but it still depends on the continuity of the international routes to which those connections are linked.
Could satellites realistically take over?
Satellites can maintain connectivity when terrestrial networks are disrupted
Satellite networks have an obvious advantage when a cable is disrupted: they do not need to follow the same physical route under the sea to connect a user to the network. They can therefore maintain or restore certain communications when terrestrial or submarine infrastructure becomes unavailable.
This capability makes them a genuine resilience tool. Satellites are already used to connect isolated areas, restore communications after disasters or provide connectivity when conventional networks are degraded.
Low Earth orbit constellations have strengthened this capability. By deploying large numbers of satellites and operating closer to Earth than geostationary systems, they can provide high-speed connections with latency low enough for many everyday uses.
But maintaining connectivity is not the same as replacing the submarine network
The key difference lies in the scale of the problem. Submarine cables are currently the main infrastructure carrying international data traffic. More than 99% of these flows travel through the submarine network, which directly connects major telecommunications networks and data centers across continents.
Satellite networks can support certain uses or provide additional routes, but they do not currently have enough capacity to seamlessly take over all the traffic that normally travels through submarine fiber-optic cables.
The larger the disruption, the more visible this difference would become. Satellites could maintain certain communications even while the performance of the international network was severely degraded, but this would not mean that all users and services could immediately return to their usual level of connectivity.
Satellite connectivity also depends on infrastructure on the ground
The idea that the Internet could simply be moved into space is also misleading. A satellite constellation is part of a broader network that includes user terminals, ground stations, telecommunications infrastructure and connections to the rest of the Internet.
Some constellations can transmit data between satellites before sending it back to the ground elsewhere, increasing their ability to bypass unavailable local infrastructure. But the traffic must ultimately reach the users, operators, data centers and services it is intended to connect.
Satellites therefore provide a different physical route, which can be particularly valuable when other infrastructure is disrupted, but they do not form a network completely independent of the terrestrial ecosystem.
Europe is developing satellite backup capacity
This complementary role is now reflected in European strategy. Since January 2026, GOVSATCOM has allowed EU Member States and certain European authorities to access pooled secure satellite communication capacity for critical missions.
The European Union is also developing IRIS², its future sovereign constellation for secure communications. The program was strengthened in 2026 and now plans for a core constellation of 348 satellites in low and medium Earth orbit, with the first launches scheduled from 2029.
The goal is not to replace submarine cables, but to add infrastructure that is independent of certain terrestrial and maritime routes, allowing government, security and emergency communications to be maintained when conventional networks are unavailable or insufficient.
In a major crisis, available capacity would become the key issue
A massive disruption to the submarine network would therefore create a very different situation from a conventional outage. Cables still in operation would continue to carry and reroute part of the traffic, while satellite networks could provide additional capacity or maintain certain communications.
But if the remaining capacity fell below the amount of data normally exchanged, simply switching technologies would not be enough to maintain normal network operations everywhere.
The real value of satellite networks would then lie in what they could still keep connected while other routes were degraded. This situation would also reveal another dependency: reliance on the actors that own, operate and control the remaining satellite capacity.
Who would become more important if submarine cables became insufficient?
Dependency would not disappear: it would shift
If a significant part of the submarine network became unavailable, the Internet would not simply switch to a single infrastructure capable of replacing the cables. Instead, the capacity that remained available would become more valuable: cables still in operation, alternative terrestrial routes and satellite networks.
The issue would then become as much a strategic question as a technical one. Actors capable of providing an alternative route or additional capacity would play a more important role in maintaining communications.
This does not mean that they would be responsible for the situation or would necessarily benefit from it economically. But a major crisis would immediately reveal which operators, companies and states still have infrastructure capable of carrying data when conventional routes are degraded.
Satellite operators would become far more strategic
In such a scenario, low Earth orbit satellite constellations would naturally become more important. Their capacity would remain far below what would be required to replace the entire submarine network, but they could maintain certain communications independently of the affected maritime routes.
This capability would give particular importance to operators that already have a large constellation, deployed terminals and a network operating at scale. Starlink currently holds a particularly strong position in this area, while Europe is seeking to develop its own alternatives.
Recent events have already shown how this dependency can become political. In October 2026, restrictions imposed on certain Starshield services used by Ukraine immediately renewed interest in European alternatives, particularly Eutelsat and its OneWeb constellation.
But switching from one network to another is not immediate. These networks use different terminals and architectures and do not offer the same capacity. Having a competing constellation therefore does not mean being able to instantly take over another operator's users and traffic.
Tech giants already control a growing share of the physical routes
Infrastructure concentration does not begin in space. For several years, major cloud and technology companies have invested directly in submarine cables to connect their data centers and secure capacity suited to their own needs.
Google, Meta, Microsoft and Amazon are now owners, co-owners or major users of numerous international cable systems. According to TeleGeography, the networks of major content and cloud providers accounted for around three quarters of international bandwidth used worldwide in 2025.
This development is gradually changing the architecture of the Internet. Infrastructure that was once developed primarily by telecommunications operators working in consortia now also depends on private companies whose networks directly connect their own data centers across multiple continents.
A cable crisis would therefore strengthen actors that are already central
A major disruption would not create this concentration: it would make it more visible. Companies with capacity across multiple routes could continue to shift part of their traffic, while satellite operators would control complementary infrastructure that would become particularly valuable.
States with diversified routes, satellite capacity, repair capabilities and multiple landing points would also be in a stronger position than territories dependent on a limited number of infrastructures or providers.
The risk posed by submarine cables therefore goes beyond the possibility of losing connectivity. It also raises a question of sovereignty: who owns the routes, who has alternative capacity, and which actors does a country depend on when part of the global network stops functioning normally?
From technical incidents to infrastructure under increasing surveillance
A series of incidents has changed how risk is perceived in the Baltic Sea
Submarine cable disruptions are not new and are still mostly caused by accidents, particularly ship anchors and fishing activities. But since 2023, a series of incidents affecting submarine infrastructure in the Baltic Sea has gradually changed how European states view this risk.
Several cases have attracted particular attention. In October 2023, the Balticconnector gas pipeline and telecommunications cables were damaged after the passage of the Newnew Polar Bear, whose anchor was linked to the damage. In November 2024, two additional telecommunications links were cut less than twenty-four hours apart, with investigations focusing on the Yi Peng 3. One month later, Estlink 2 and four telecommunications cables were damaged, leading Finnish authorities to intercept the Eagle S.
The accumulation of incidents became significant enough to go beyond a simple succession of isolated cases. In May 2025, NATO reported that at least eleven submarine communication or energy cables had been damaged in the Baltic region over the previous fifteen months.
This figure does not prove that these incidents were part of a single coordinated campaign. It does, however, show why their recurrence has become a security concern. In several cases, a vessel or an anchor could be physically linked to the damage, without this necessarily being sufficient to establish an intent to sabotage or, even more so, the involvement of a state.
The Baltic Sea also concentrates numerous telecommunications, energy and transport infrastructures within a geographically limited area with significant maritime traffic. The recurrence of these incidents has therefore revealed a broader risk: several critical infrastructures could be affected within the same region, whether the cause is accidental or intentional.
NATO has deployed military assets to monitor submarine infrastructure
The most visible response came in January 2025. Following further damage between Finland and Estonia, NATO launched Baltic Sentry, an activity designed to strengthen the protection of critical infrastructure in the Baltic Sea.
The initiative includes frigates, maritime patrol aircraft and naval drones. The objective is not to place a military vessel above every cable, which would be impossible, but to improve awareness of maritime traffic, detect suspicious behavior and strengthen the ability to respond quickly when an incident occurs.
This development is significant. Infrastructure installed to carry commercial and civilian data is now monitored in some areas using military assets. The physical protection of the Internet has therefore become a matter of collective security for the Atlantic Alliance.
The European Union has moved from cable protection to a broader resilience strategy
The European Union has followed a similar path. In February 2025, it adopted an action plan dedicated to the security of submarine cables. The plan does not focus solely on preventing attacks: it covers prevention, detection, response, recovery and deterrence.
A European risk assessment was subsequently carried out in 2025, together with mapping work and stress testing. In February 2026, the Commission published a dedicated toolbox and identified cable projects considered priorities for strengthening European resilience.
This change in approach is significant. The objective is no longer simply to prevent a cable from being damaged, but to assume that an incident may still occur and prepare the network to continue operating, detect damage quickly, provide alternative routes and restore connections.
Sabotage is now a recognized risk, but attribution remains the greatest challenge
The events in the Baltic Sea ultimately show why submarine cables are particularly vulnerable to so-called grey-zone situations. An anchor can destroy a fiber-optic cable without the use of explosives or military weapons, and the same type of damage can result from negligence, an accident or a deliberate act.
This ambiguity makes attribution considerably more difficult. Even when a vessel is identified and its anchor is physically linked to the damage, proving intent and then establishing that an operation was ordered by a state requires an additional level of evidence.
This is probably one of the most important lessons from the current situation: the threat lies not only in the technical possibility of cutting a cable. It also lies in the difficulty of quickly determining whether a disruption resulted from a maritime accident, serious negligence or a deliberate hostile operation.
How the situation has evolved
This timeline tracks the main incidents and decisions that could change the assessment of submarine cable security and resilience in Europe. It is updated when a new event introduces a significant development: a major disruption, investigation findings, a newly identified threat, stronger surveillance measures, or changes in repair and redundancy capabilities.
October 2023: Balticconnector and telecommunications cables are damaged
During the night of October 7 to 8, 2023, the Balticconnector gas pipeline between Finland and Estonia was damaged, along with telecommunications cables in the region. The Finnish investigation linked the damage to the Newnew Polar Bear, a Chinese-flagged container ship, and to an anchor that had been dragged along the seabed.
The incident became one of the first major warnings about the vulnerability of different types of infrastructure concentrated in the same area. The physical damage was established, but whether it was intentional was not.
November 17 and 18, 2024: two cables are cut in the Baltic Sea
Two fiber-optic connections were damaged less than twenty-four hours apart: the BCS East-West Interlink between Sweden and Lithuania, followed by C-Lion1 between Finland and Germany.
Investigators focused their attention on the Yi Peng 3, a Chinese bulk carrier that had crossed the affected areas. Swedish authorities requested its cooperation, and several countries subsequently took part in the investigations. The possibility of deliberate action was examined, but no state attribution could be established at that stage.
December 25, 2024: Estlink 2 and four telecommunications cables are affected
The Estlink 2 electricity cable between Finland and Estonia went offline, while four telecommunications cables were also damaged. Finnish authorities intercepted the Eagle S, an oil tanker associated with Russia's "shadow fleet", suspected of having dragged its anchor along the seabed.
The incident triggered a much faster response than previous episodes. It directly contributed to increased maritime surveillance in the Baltic and turned the succession of cable disruptions into a collective security issue.
January 14, 2025: NATO launches Baltic Sentry
NATO announced Baltic Sentry to strengthen the protection of critical submarine infrastructure in the Baltic Sea. Frigates, maritime patrol aircraft and naval drones were deployed to improve surveillance and response capabilities.
This decision marked a significant shift: submarine telecommunications and energy infrastructure was no longer a concern only for operators and civilian authorities. Its protection had become part of NATO's operational security priorities.
February 21, 2025: the European Union adopts its cable security action plan
The European Commission and the EU High Representative presented an action plan covering the entire resilience cycle: prevention, detection, response, recovery and deterrence.
The European approach therefore no longer focuses solely on protecting cables against disruption. It also aims to identify critical infrastructure, detect incidents more quickly, provide alternative routes and reduce the time required to restore damaged connections.
October 2025: the European Union completes its first joint risk assessment
The European expert group on submarine cable infrastructure published a risk assessment, accompanied by mapping work and guidance for conducting stress tests.
The objective is now to identify not only exposed cables, but also shared dependencies and areas where several infrastructures could be affected by the same event.
February 5, 2026: Europe commits €347 million and adopts a dedicated toolbox
The European Commission published its Cable Security Toolbox and identified Cable Projects of European Interest considered priorities. At the same time, it announced an additional €347 million for 2026 and 2027 to strengthen backbone infrastructure, cable resilience and European repair capabilities.
The announced measures also include the development of capabilities to restore damaged cables more quickly. Repair capacity therefore becomes an explicit element of strategic security, alongside surveillance and redundancy.
June 23, 2026: first regional hubs and new repair capabilities
The European Commission provided €5.8 million in funding for the first two Regional Cable Hubs, one focused on the Baltic Sea and the other on the Mediterranean. They are intended to strengthen cooperation, surveillance, threat detection and incident response.
A new €40 million call was launched at the same time to increase European repair capacity for submarine communication cables. The objective is to reduce the amount of time a connection remains unavailable after a disruption.
October 17, 2026: where does the situation stand?
As of this update, the disruptions observed in recent years have not caused a collapse of the European network. Redundancy and rerouting have allowed communications to continue despite several significant incidents.
What has changed most is the way the risk is being addressed. Repeated damage in certain areas, the difficulty of attribution and the possibility of multiple simultaneous disruptions have led the European Union and NATO to treat submarine infrastructure as a strategic issue requiring surveillance, alternative routes and additional repair capacity.
The main point of concern is now the emergence of incidents capable of affecting several important routes simultaneously, or a succession of disruptions occurring quickly enough to exceed available rerouting and repair capacity. This is the type of development that would materially change the situation described in this report.
Going further: data and monitoring
Some of the developments discussed in this report can also be observed through public data, including maritime traffic, infrastructure, energy and other indicators that may help track how the situation evolves.
These data are monitored separately to keep this report focused on a consolidated analysis of the subject, while still making it possible to observe measurable changes over time.
Explore the trackers and public data →
Follow developments more closely
To follow important developments more directly, you can also find L'Actu sans le Bruit on Telegram. The channel follows new events as they unfold, while this report is updated only when those developments materially change the situation.
The channel is published in French, but Telegram's translation feature can be used to translate posts into English.