Every "sovereign AI" pitch in the Gulf assumes the region controls its own data pipes. It doesn't, not yet. Over 90% of Europe-Asia data traffic still runs through the Suez Canal and Red Sea corridor, according to reporting by Fortune, a single geographic chokepoint carrying the traffic behind a region betting hundreds of billions of dollars on AI compute.
That is now changing, on paper. Four subsea cable projects worth more than $2 billion combined are under construction or newly announced across the GCC. None of them will be ready this year, and one of the largest cannot currently even get its cable-laying ships into position.
What's actually being built
Qatar's Ooredoo Group is leading the highest-profile project. Fibre in the Gulf (FIG) is a $500 million, roughly 2,000-kilometre system with 720 terabits per second of capacity, designed to link all six GCC states plus Iraq while bypassing the Suez Canal and the Bab-el-Mandeb strait entirely, according to Fortune's reporting. Ooredoo set up a dedicated entity, Ooredoo Fibre Networks, in February 2026 to run it, and completion is scheduled for late 2027.
It isn't alone:
- SilkLink — an $800 million, 4,500-kilometre fibre network backed by Saudi Arabia's STC, running from Syria to the Mediterranean
- WorldLink — a $700 million hybrid subsea and terrestrial network, an Iraqi-UAE consortium, connecting the UAE through Iraq to Turkey
Collectively, the projects announced across the region target 8 to 10 gigawatts of AI compute capacity, per Fortune's analysis, all of it dependent on data infrastructure that mostly does not exist yet.
Why the rush
The trigger is geopolitical, not commercial. Houthi attacks in the Red Sea have halted new cable construction in the region since 2024. Meta and Google have both delayed subsea projects through the same corridor over security concerns. Ooredoo Group CEO Aziz Aluthman Fakhroo told Fortune that the recent Iran conflict "reinforced the proposition" for FIG's necessity — and then admitted the same conflict is currently blocking it: cable-laying ships cannot get inside the Strait of Hormuz.
Fortune's reporting also notes that no new subsea capacity has been built in the Red Sea region in five to seven years, against annual data flow growth of roughly 30%. The infrastructure gap was opening before AI accelerated it. AI has just made the gap expensive to ignore.
What this means for programme delivery
None of this is a reason to delay GCC AI investment. It is a reason to stop treating "the region has sovereign AI infrastructure" as a settled fact when you scope a programme.
- Check what's actually routing your data today, not the roadmap. A vendor's marketing deck showing a 2027 cable does not change where your traffic runs in 2026. If a client's compliance case rests on regional data sovereignty, confirm the physical route, not the announced one.
- Build redundancy assumptions around a five-to-seven-year infrastructure gap, not a clean slate. Firms planning latency-sensitive AI workloads (real-time voice AI, agentic decisioning) in the Gulf should assume current-generation connectivity risk persists through at least 2027, and treat FIG, SilkLink, and WorldLink completion dates as targets, not commitments, given the geopolitical dependencies already delaying at least one of them.
- Separate data centre capacity from data transport capacity in due diligence. The region's $30 billion-plus data centre build-out and its subsea cable build-out are two different infrastructure races, on two different timelines, run largely by different players. A programme can have committed compute and uncommitted connectivity at the same time.
- Track this alongside data residency law, not instead of it. Saudi Arabia's data embassy legislation and similar frameworks elsewhere in the GCC answer the legal jurisdiction question. They don't answer the physical routing question. Both need checking, separately, before a programme's infrastructure assumptions go into a client proposal.
The bottom line
The Gulf's bet on independent data infrastructure is real, well-funded, and strategically sound. It is also, right now, mostly future tense. Programmes that assume 2027's cable map describes 2026's actual routing are building on infrastructure that doesn't exist yet, in a region where the last stretch of cable-laying was stopped by a war. Scope accordingly.