Why AI Growth in the Gulf Depends on Fragile Undersea Cables

Gulf nations are building vast AI computing centers to supply processing power to markets across Europe, Asia, and Africa. But there's a hidden problem: nearly all the data flowing in and out of these centers travels through undersea cables that pass through some of the world's most unstable maritime zones. When Iran recently suggested it might take control of seven cables running through the Strait of Hormuz, it exposed a vulnerability that could undermine the region's entire AI strategy.

The issue is geography. Most Gulf connectivity to Europe and the United States depends on routes through both the Red Sea and the Strait of Hormuz—two narrow passages where a conflict, sabotage, or even an accident could cut multiple cables at once. The existing cable network was built for a quieter world. It's now carrying the hopes of a multi-billion-dollar AI buildout.

The AI Infrastructure Bet

Saudi Arabia and the UAE have committed billions of dollars to become regional AI hubs, betting they can compete globally in compute services. The geography is appealing: the Gulf sits at a natural crossroads between Europe, Asia, and Africa, making it a logical place to host AI workloads that serve all three regions.

Microsoft's partnership with G42, announced in April 2024, is the highest-profile sign of this trend. Microsoft will supply cloud computing through Azure; G42 will build and sell AI products like its Jais language model to customers across the Middle East, Central Asia, and Africa. The deal signals confidence that Gulf-based compute can work at scale.

Ooredoo, a major telecom operator in the region, is betting heavily on this vision too. The company is building data centers—facilities that house computing equipment—and trying to become the go-to GPU provider (GPUs are chips that run AI calculations) for the entire Gulf Cooperation Council, or GCC. Ooredoo operates over 26 data centers today and is planning to expand to 120 megawatts of capacity—a significant commitment requiring roughly 4 billion Qatari riyals.

The Cable Problem

Here's where the ambition runs into reality. All that computing power is only useful if it can reliably connect to the rest of the world. Right now, seven undersea cables funnel data through the Strait of Hormuz, one of the world's busiest shipping lanes and a perennial geopolitical flashpoint. More cables run through the Red Sea, where ongoing regional conflicts have already disrupted shipping routes.

If multiple cables fail at the same time—which can happen if they're clustered in the same narrow passage—whole regions lose connectivity. This has happened before. In 2008, cable cuts in the Mediterranean knocked out service across multiple countries simultaneously. The Gulf today faces a similar concentration of risk, but with higher stakes: the infrastructure is meant not just for local use but for serving paying customers globally.

Regional tensions are making the problem worse. According to TeleGeography, an infrastructure research firm, geopolitical instability is delaying investment in new cable routes that could provide backup pathways. It's a vicious cycle: conflict threatens existing cables and also prevents building alternatives.

Plans for Better Routes

The industry is aware of the risk and working on solutions. Ooredoo is building something called the FIG subsea cable, a new route designed to move data more efficiently within the Gulf region itself. The cable is expected to be ready by late 2027 and will carry 720 terabits per second—about the volume you'd need for thousands of high-definition video streams, scaled to data center traffic.

More ambitiously, Ooredoo is planning to partner with another operator, stc (SONIC), to build terrestrial routes—fiber-optic cables running overland—that would bypass the Gulf of Aden and Red Sea altogether. That could reduce dependence on maritime chokepoints, though overland routes bring their own geopolitical and engineering challenges.

Ooredoo is also rapidly expanding tower infrastructure across the region. The company operates over 30,000 towers across six countries and is assembling what it describes as the largest regional tower operator in the Middle East and North Africa through partnerships and asset purchases.

Why This Matters

There's a fundamental tension at play here. The Gulf's location is attractive for AI because it's central to three continents. But that same location puts critical infrastructure within striking distance of multiple potential conflict zones. Geographic advantage and geopolitical risk come bundled together.

The broader point is worth flagging: when enterprise customers—financial firms, healthcare providers, governments—consider storing and processing AI workloads in the Gulf, they now have to factor in something that wasn't a major concern even five years ago. They need to understand not just whether a data center has enough power and cooling, but whether the cables connecting it to the rest of the world are resilient to geopolitical disruption. It's a new kind of risk calculation.

The timeline is also important. The FIG cable and other alternatives won't be operational until late 2027 or beyond. That leaves a multi-year window where Gulf AI infrastructure remains dependent on the current vulnerable cable architecture, even as regional tensions show little sign of easing. For businesses evaluating a long-term commitment to Gulf-based AI services, that window matters.

Looking at the bigger picture, the Gulf situation illustrates a trend that will shape AI infrastructure globally. As AI workloads move around the world instead of staying concentrated in a few U.S. or European data centers, infrastructure planners have to weigh more factors than they used to: not just energy costs and regulatory environment, but geopolitical routing, cable resilience, and backup options. The Gulf's experience is an early preview of that more complex world.

The region's AI buildout continues despite these cable vulnerabilities. Whether it can become a truly reliable global AI hub—the kind that enterprises bet their operations on—may ultimately depend on whether alternative routes become operational before geopolitical tensions escalate further.