A new Ciena survey of 200 service providers across the UAE and Saudi Arabia finds that the networks connecting distributed AI compute are becoming a revenue stream in their own right. 96 percent of UAE respondents and 90 percent in Saudi Arabia expect high-capacity AI connectivity to drive net-new growth over the next three to five years, though most concede their optical networks are not yet ready to carry it.

[For more news, click here]

The Gulf’s artificial intelligence story has been told mostly in gigawatts and chips. Abu Dhabi’s Stargate UAE, a five-gigawatt campus its backers say will cost more than $30 billion, and Saudi Arabia’s HUMAIN, a roughly $100 billion program spread across 11 data centers, have dominated coverage with their scale and their rosters of American partners. A new survey from Ciena, the US optical-networking company, sought to find out what connect the compute once it is built and who pays to run the connections.

The answer, according to service providers in both countries, is increasingly the network itself. In the UAE, 96 percent of respondents expect high-capacity AI-driven network services to be a primary driver of net-new revenue over the next three to five years, against 90 percent in Saudi Arabia. Confidence in managed optical fiber networks, used to stitch together distributed AI compute clusters, runs at 97 percent in the UAE and 96 percent in Saudi Arabia. The survey, run with Censuswide between 13 and 23 July 2026, polled 100 providers in each market as part of a study of more than 1,200 respondents across 12 countries.

Why Distributed Compute Changes the Math

A single AI training rack can draw more than 100 kilowatts, roughly ten times a conventional enterprise rack. The constraint in growth in this case is not land or capital, but power availability. This is pushing hyperscalers to spread compute across multiple sites, which turns the links between those sites into a product in their own right. 99 percent of UAE providers and 91 percent in Saudi Arabia see this “scale-across” hyperscaler demand as a meaningful contributor to wholesale revenue. To put it simply, the more AI capacity fragments across locations, the more the interconnect matters. This reframes the network from into a monetizable rather than cost center. It also explains how the survey on fiber and optics is actually a survey about where the next wave of telecom revenue will come from.

The Same Optimism, a Different Clock

There is a difference in urgency in both markets. In the UAE, 97 percent of respondents call optical network upgrades urgently needed to support premium enterprise AI service-level agreements, with 56 percent describing them as critically urgent. In Saudi Arabia, 86 percent see the upgrades as urgent, with just over half saying they are needed immediately.

“Globally, service providers are balancing optimism with urgency about AI-driven revenue opportunities,” said Pete Hall, Regional Managing Director for the Middle East and Africa at Ciena. “The survey findings for UAE and Saudi Arabia showed the same optimism, but urgency sits at the center of the conversation. In the UAE, service providers are treating network upgrades and automation as something to act on now, not plan for later. In Saudi Arabia, service providers are also focused on AI inference, service reliability, and sovereign infrastructure requirements.”

Two Builds, Two Strategies

Those different emphases track the different shapes of the two national programs. Saudi respondents were more likely to tie future revenue to AI inference with 58 percent expecting growth in inference data centers to lift demand for data center interconnect, above the 49 percent global average and well ahead of the UAE’s 37 percent. Saudi providers also lean toward multi-cloud connectivity management as a revenue model, at 64 percent, while UAE providers still put traditional fixed-capacity services first. The UAE is further along on quantum-safe encryption, with 54 percent saying they have launched or will launch such services within a year, against 38 percent in Saudi Arabia.

Analysts have described the two builds in comparable terms, which is, the UAE’s effort, led by G42, has been framed as a more open international compute play anchored to US frontier labs, while HUMAIN carries explicit national-champion features, including Aramco-linked energy and a sovereign Arabic model. The survey’s split on inference and sovereignty reads like the network-layer echo of that divide.

Why US Investors Should Care

For American investors, the through-line is that US money and hardware sit on both sides of the trade. Stargate UAE’s partners include OpenAI, Oracle, Nvidia, and Cisco. Ciena itself trades on the New York Stock Exchange and stands to gain as the recurring business of moving AI traffic between campuses scales up. The compute announcements capture the capital expenditure. The network is where much of the operating revenue and the vendor competition will land.

While regional research has repeatedly flagged the gap between announced gigawatts and capacity operators can use, it should be taken with caution as new supply arriving in 100- to 200-megawatt increments is gated by grid connections. IDC has projected AI spending across the Middle East, Türkiye, and Africa rising from $4.5 billion in 2024 to $14.6 billion by 2028. However, with substations setting the pace, the Ciena survey indictaes the tension. While ambition and optimism exist en masse, the networks underneath may just not be ready.

Related Articles

How the Middle East Became the World’s Most Exciting Digital Infrastructure Market

Riyadh’s Global AI Show 2026 Turned Sovereign AI From a Buzzword Into Binding Contracts

Saudi Arabia’s AI Adoption Has Crossed a Tipping Point, and Gen Z Is Leading the Pushback