A major AI-focused network project between Qatar’s GBI and Nokia highlights the critical role of subsea and terrestrial fibre in global connectivity. With a new terrestrial route through Iraq, GBI is enhancing east-west internet capacity, resilience, and route diversity. It tells us a lot about the anatomy of the whole global setup – and specifically the hidden connections between the building blocks of the internet.
In sum – what to know:
Middle East connectivity – a new high-capacity terrestrial route through Iraq complements existing subsea cables, enhancing east-west internet links and providing alternative paths for enterprise AI.
Physical AI networks – subsea cables, landing stations, and inland fibre routes form the backbone of global connectivity, enabling low-latency, reliable access to hyperscale cloud platforms.
Diversify for resilience – by combining subsea and terrestrial infra, enterprises get predictable performance, redundancy against outages, and reduced latency for high-end AI apps.
News yesterday about a major AI network infrastructure project between Qatar-based telco Gulf Bridge International (GBI) and Nokia to enhance east-west internet connectivity through the Middle East has prompted RCR Wireless / RCR Tech to put together a quick primer on the anatomy of the global internet via sub-sea cables, landing stations, and terrestrial backhaul routes – with a timely view to discuss the importance of geographic diversity (GBI’s Iraq route) to protect against outages and geopolitical risks.
GBI, founded in 2008, owns and operates subsea and terrestrial fibre infrastructure linking the Gulf region to Europe, Asia, and other global hubs. It offers enterprise customers and regional operators dedicated network links into the big data centres and cloud ‘on-ramps’ operated by the major hyperscalers (AWS, Azure, Google, plus the rest). In other words, customers along GBI’s fibre-optic paths in the Middle East can connect securely to those cloud platforms, rather than sending data over the public internet.
The firm wholesales high-capacity connectivity, including public internet access, and also private networks (MPLS, SD-WAN), high-speed long-distance connections (DWDM, IP Transit), and direct links to major cloud providers (Cloud Connect). An advantage is its route diversity, including a new terrestrial ‘north route’ via Iraq to Europe, which bypasses traditional subsea corridors, offering greater diversity and resilience for regional and international data traffic. But this discussion is not really about GBI, even if the GBI/Nokia deal helps to frame it.
So let us consider it only in that context. Nokia is supplying an optical network solution (1830 GX) for the new inter-Iraq terrestrial deployment, which also provides subsea backhaul. This will enable GBI to deliver more than 50 terabits (TB) of capacity to its customers. The project is significant on the grounds it provides strategic infrastructure in a region that historically has had limited high-capacity, diverse connectivity options, and there is lots in the release, and elsewhere on the GBI network, about growing the digital economy in the Middle East .
The bigger point is its new terrestrial route complements traditional subsea corridors between Europe and Asia, and represents a strategic regional extension to a global patchwork of subsea cables, landing stations, and terrestrial backhaul routes that keeps the cloud-based internet ticking (most of the time), and is being rapidly beefed-up to support new enterprise AI access. In the end, the cloud depends on a very physical network of cables and fibres – and backhauling is the link that keeps it all connected. So let’s break this down…
Anatomy of the global internet
As above, the internet is a physical network. Subsea cables carry the bulk of international traffic, linking continents via fibre-optic lines that run thousands of kilometers under oceans. These subsea cables terminate at landing stations on the shore, which serve as gateways to connect to inland fibre routes into cities, industrial hubs, and cloud data centres. GBI, in this example, operates both systems – subsea and terrestrial – for enterprises of various sorts in the Middle East to connect directly on private and secure paths with major cloud platforms.
Inland fibre routes, often overlooked, are the real arteries in this global internet body – which (to confuse metaphors) deliver long-haul ‘backbone’ transport between major regions, and regional backhaul ‘feeder’ links to bring traffic from local networks onto the backbone fibre. For AI workloads (and any ‘real-time’ enterprise application), every millisecond matters, of course. So by controlling both subsea and sur-terra infrastructure, the likes of GBI can offer end-to-end connectivity that balances bandwidth, reliability, and latency for new cloud access in the region.
Why subsea backhauling matters
Subsea backhauling, as discussed by GBI and Nokia, refers to the transport of data from subsea landing stations into the wider terrestrial network, ensuring that undersea capacity reaches enterprise and cloud data centres efficiently. Without effective backhaul, high-capacity submarine cables cannot deliver low-latency, reliable connections that modern applications – including AI – require. Subsea backhaul can be provided via leased circuits, dark fibre, or IP-layer transport – depending on the supplier service and the customer need.
For GBI, subsea backhauling is a differentiator. By coupling subsea capacity with its terrestrial network, it can manage performance with a higher degree of predictability across the Middle East, and into Europe or Asia – bypassing congested public internet paths. It means better latency for enterprise workloads that go via hyperscale clouds, and a degree of redundancy against potential outages in public-access subsea corridors. It provides enterprises with a degree of confidence to run AI analytics, automation, or trading applications over these paths.
Terrestrial network resiliency and latency
Terrestrial routes – like the Iraq overland (tunnelled-fibre) bypass – provide a way for service providers like GBI to dodge undersea chokepoints or politically sensitive regions. Its ‘north route’ is a clear example, providing a secure and direct alternative path from the Gulf to Europe. For enterprises, this translates to more predictable latency, higher availability, and resilience against outages caused by cable faults, natural disasters, or regional disruptions. Beyond resilience, terrestrial routes can materially improve performance for latency-sensitive applications.
AI workloads, video streaming, and industrial control systems benefit from even modest reductions in transit time. By integrating the Iraq route with its existing subsea network, GBI offers enterprise customers multiple options – optimized for latency, redundancy, or overall bandwidth. Such flexibility is especially important in regions where single points of failure in digital infrastructure remain a concern – which is basically everywhere, actually, as the US-originated global AWS outage has recently shown.
Critical chokepoints and route diversity
Just because of geography, physics, and politics, trans-international fibre routes tend to be concentrated in certain areas: through narrow straits and channels, avoiding deep-sea trenches and ridges, navigating territorial waters, where possible following the shortest path. But this over-concentration creates chokepoints, which disrupt large volumes of traffic when things go south, whether because of natural events, accidents, or geopolitical disruptions. Hence why route diversity is so important. Redundant terrestrial and subsea paths provide alternative routes.
Which allow enterprises to maintain their operations, when the main thoroughfares come to a standstill.
Terrestrial-subsea integration for AI
AI increases the demand for bandwidth, speed, and predictable network performance. Integrated terrestrial–subsea networks, like GBI’s model, allow operators to scale capacity end-to-end while maintaining predictable performance for demanding workloads. These networks support hybrid architectures where data is processed across multiple sites, in real time, or in the cloud. Such integration will be key to enabling AI-driven apps such as industrial automation, real-time analytics, and multi-cloud orchestration.
In the end, the GBI strategy – hybrid infrastructure, capacity fibre, route diversity – is just what AI ordered.
