countries · sectors · sub-national hubs · trade bodies · FTAs · tools · academy · essays
Approximately 99% of intercontinental data moves over submarine fibre — including every cross-border trade transaction, every banking message, every customs filing.
The popular mental model of intercontinental data flow is satellite-mediated. The actual mental model is fibre-mediated: roughly ninety-nine per cent of intercontinental data — including the data underlying every cross-border trade transaction, every SWIFT message, every customs filing, every B2B video call between buyer and seller, every container-tracking telemetry stream, every electronic letter-of-credit document, and almost everything else that the modern trading system runs on — moves through submarine fibre cables that lie on the seabed of the world's oceans. Satellite capacity exists, but at orders-of-magnitude lower throughput, higher latency, and higher cost per gigabit. For practical purposes, intercontinental commerce runs on the submarine-cable system.
The structural relevance to a trade-intelligence audience is that the submarine-cable system is geographically concentrated, partially redundant, and exposed to a small set of physical and geopolitical risks that have produced material commercial disruptions in recent years. A consignment of pharmaceuticals from Hyderabad to Düsseldorf cannot move if the customs filing cannot be transmitted from the Indian customs broker to the German Atlas system. A letter of credit cannot present if the SWIFT message cannot route from the issuing bank to the advising bank. The cable system is therefore a load-bearing rail of the trade economy, even though it never appears on a bill of lading.
There are roughly five hundred named submarine cables in commercial service worldwide, with cumulative installed length on the order of 1.5 million kilometres. The system has expanded substantially since 2015 as cloud-provider capital — Google, Meta, Microsoft, and Amazon all now own or co-own significant cable assets — has supplemented the traditional carrier consortium model. Approximately seventy per cent of new cable capacity since 2020 has been built or co-financed by the four major hyperscaler cloud companies, which has reshaped the system's commercial geometry.
For trade purposes, however, the relevant subset is much smaller. Cross-border trade transactions concentrate on a few dozen high-capacity systems serving the principal commercial axes: trans-Atlantic, trans-Pacific, Europe-to-Asia (the Mediterranean / Red Sea / Indian Ocean route), Europe-to-Africa, and the various intra-regional Asia, intra-regional Europe, and intra-regional Americas systems. The major systems below are the ones whose disruption would be felt commercially.
MAREA is one of the highest-capacity trans-Atlantic cables in service, connecting Virginia Beach in the United States to Bilbao in Spain. The cable was originally a Microsoft, Meta (then Facebook), and Telxius (Telefónica) consortium and entered commercial service in 2018 with eight fibre pairs. The southerly trans-Atlantic routing — landing in Iberia rather than Northern Europe — provides geographic diversity from the dominant trans-Atlantic flows that route between the US northeast and the UK / France / Netherlands.
For trade purposes, MAREA's relevance is the European-Iberian-American commercial corridor: Spanish, Portuguese, and Latin American businesses with US counterparties route through MAREA along with the US east-coast trade with Iberia. The cable's onward connection in Spain feeds the European backbone via terrestrial fibre.
Dunant is Google's privately-owned trans-Atlantic cable, in service since 2021, running between Virginia Beach and the French Atlantic coast. Dunant uses space-division multiplexing — twelve fibre pairs against MAREA's eight — to deliver substantially higher per-cable capacity and serves Google's intra-cloud connectivity between US and European data centres. The cable's trade relevance is indirect but substantial: a large share of business-to-business commerce now relies on Google Cloud, Workspace, or Search infrastructure that depends on Dunant for its trans-Atlantic data flow.
Grace Hopper is a second Google trans-Atlantic cable, in service since 2022, with branching units connecting the United States to both the United Kingdom (Bude, Cornwall) and Spain (Bilbao). The dual-landing design provides Google with onward European reach via two distinct national networks. The cable's commercial trade relevance follows the same pattern as Dunant — every Google-hosted business application that relies on trans-Atlantic data flow now has redundant capacity across Dunant and Grace Hopper.
The SEA-ME-WE series — currently progressing from the in-service SEA-ME-WE-3, 4, and 5 toward the under-construction SEA-ME-WE-6 — is the principal Europe-to-Asia submarine cable backbone. The route runs from Marseille on the French Mediterranean coast, through the Suez Canal land-route (the cables run terrestrially across the Egyptian isthmus rather than transiting the Suez Canal itself), down the Red Sea, around the Bab-el-Mandeb, into the Indian Ocean, with branching landings in the Gulf, India, Sri Lanka, and onwards to Malaysia, Singapore, and beyond.
SEA-ME-WE-6, which is in late-construction phase as of 2026, will be the highest-capacity Europe-to-Asia cable in service, with new landings, additional fibre pair count, and modern coherent optical technology. The cable's structural significance is the dependence of Indian, Bangladeshi, Sri Lankan, and South-East Asian internet traffic on the Red Sea passage. A cable cut in the Red Sea — which has happened multiple times in the last decade, including the high-profile 2024 incidents associated with the Yemeni civil war — produces measurable degradation of internet performance across South Asia and South-East Asia.
AAE-1 is a parallel Europe-to-Asia cable to the SEA-ME-WE series, in service since 2017, with a similar Marseille-to-Hong Kong corridor and branching landings throughout the Middle East, East Africa (Djibouti), Pakistan, India, Myanmar, Cambodia, Vietnam, and Singapore. The cable provides the principal redundancy to SEA-ME-WE-5 and -6 on the Europe-to-Asia route, and was the system that absorbed substantial traffic during the 2024 Red Sea cable disruptions.
2Africa is the longest submarine cable in service, completed progressively from 2023 onward, encircling the African continent with branching landings in approximately twenty African countries plus Europe (Spain, France, Italy) and Middle East / Asia (Saudi Arabia, Pakistan, India). The cable's structural significance is its first-of-its-kind African continental coverage — many African coastal countries had previously been served only by lower-capacity cables, and 2Africa materially increased the bandwidth available to African internet markets.
For trade purposes, 2Africa is the data-rail equivalent of the AfCFTA free-trade area: the cable provides the digital infrastructure to support the increasing intra-African trade volumes that AfCFTA is designed to enable. The trade-finance, customs-clearance, and B2B-commerce capacity of African economies is meaningfully shaped by 2Africa's coverage.
JUPITER is a high-capacity trans-Pacific cable, in service since 2020, connecting the US west coast to Japan with onward branching to the Philippines. The cable carries a substantial share of US-Japan commercial traffic and is one of the principal trans-Pacific redundancy systems alongside FASTER, PLCN, and the older NCP, TPE, and Unity systems.
FASTER, in service since 2016, is the trans-Pacific cable system that established the modern multi-carrier-plus-hyperscaler consortium model. The cable connects Oregon to Japan with onward connections via the regional Asian cable network. FASTER's commercial relevance is structural — it carries a meaningful share of US-Japan and US-Asia broader internet traffic, and its capacity remains in active use even as newer cables (NCP, JGA-N, JGA-S, the Topaz cable) are added to the trans-Pacific corridor.
Three geographic chokepoints concentrate disproportionate cable density, creating concentrated infrastructure-disruption risk that interacts directly with the maritime-chokepoint axis treated elsewhere in this atlas.
The Red Sea is the structural cable chokepoint of the Asia-Europe corridor. Approximately seventeen submarine cables transit the Red Sea, including the entire SEA-ME-WE series (3, 4, 5, 6), AAE-1, IMEWE, EIG, FALCON, and others. The cables enter the Red Sea at the Bab-el-Mandeb, run northward, and exit via the Egyptian isthmus where they cross to the Mediterranean as terrestrial fibre. The 2024 multiple-cable cuts associated with the Yemeni civil war and Houthi-area maritime activity demonstrated that simultaneous disruption of multiple Red Sea cables is operationally feasible and produces measurable internet degradation across South Asia and South-East Asia.
The mitigation is partial. Some Asia-Europe traffic can be re-routed via Pacific cables to North America and onward to Europe — adding latency and consuming capacity on systems sized for direct trans-Pacific traffic. The Asia-Pacific-to-Europe-via-Russia (TEA-NEXT) terrestrial route is geopolitically constrained. The northern Arctic-route cables — the Polar Connect and similar projects — are partially built but not yet at scale.
The Luzon Strait between Taiwan and the Philippines is the second major cable chokepoint, where the dominant trans-Pacific cable systems converge before diverging into the Asian regional network. The 2006 Hengchun earthquake near Taiwan severed multiple cables in the Luzon Strait simultaneously, producing weeks of degraded service across Asia. The geographic concentration has not changed since, although newer cables have added capacity and some routing diversity.
The English Channel and approaches concentrate trans-Atlantic cables landing in the UK, Ireland, France, and the Netherlands together with the intra-European cable network. Cable density on the Channel seabed is high, and fishing-anchor and shipping-anchor incidents are a recurring source of disruption. The 2022–2024 period saw a series of cable disruptions in northern European waters — some of them likely accidental, some of them possibly deliberate — that highlighted the geopolitical-risk dimension of cable infrastructure.
The submarine-cable system is structurally resilient at the global level — no single cable cut produces a system-wide outage — but is locally vulnerable, with cable-density chokepoints creating concentrated risk. The resilience model relies on a small fleet of cable-repair ships (approximately sixty in operation worldwide) that respond to cuts via a tiered service-level-agreement system among consortium members. A typical mid-ocean cable cut takes one to three weeks to repair; a Red Sea or Luzon Strait cut may take longer because of seabed conditions and political authorisation requirements.
For trade-purposes, the implications are practical. Banks running SWIFT operations have geographically-redundant network paths and typically do not lose international correspondent capability through a single cable cut. Customs systems (India's ICEGATE, Germany's Atlas, China's customs single window, the EU's NCTS) are similarly geographically-redundant. B2B video calls and routine email may degrade visibly during a multi-cable disruption but do not stop. The systems most exposed to cable-disruption are the ones with low geographic redundancy, narrow latency tolerances, or single-region cloud-provider deployments — and these include some bilateral trade-finance applications, some real-time supply-chain telemetry, and some high-frequency commodity-trading flows.
Explore
Every page in the AJG platform cross-links to these primary entities. Click any pill to explore that branch of the knowledge graph.