Pete Hall, MEA Head for Ciena, on the region’s race to build AI-ready infrastructure
As Saudi Arabia and the United Arab Emirates accelerate investments in artificial intelligence and hyperscale data centers, a crucial question is emerging across the industry: can connectivity infrastructure keep pace with the explosive demand created by AI?
According to Pete Hall, MEA Regional Head for Ciena, the region is already moving quickly to ensure that network capacity grows alongside its ambitions.
Hall says that over the past two years the pace of infrastructure investment has increased significantly, particularly in the UAE and Saudi Arabia. “The level of investment and planning we are seeing across the region is very deliberate,” he explains. “AI infrastructure depends on several layers coming together, including GPUs, data centres and the connectivity that links them. From what we are seeing, those plans are being developed in parallel rather than in isolation.”
The Middle East also benefits from structural advantages that position it well for the next wave of AI growth. Hall points to geography as one of the most important. Sitting between Asia, Europe and Africa, the region acts as a natural global interconnection hub, making it strategically valuable for large-scale data movement.
Government policy is another major driver. Across the Gulf, national strategies around sovereign cloud and AI are unlocking large-scale investments. These initiatives are accelerating the deployment of data centres, connectivity infrastructure and digital ecosystems designed to support AI innovation.
In addition, the region has fewer legacy infrastructure constraints compared with many mature markets. “In the Middle East, there is access to land, scalable energy and greenfield network environments,” Hall notes. “That allows countries like the UAE and Saudi Arabia to build infrastructure optimised for the future rather than retrofitting older systems.”
However, AI-native data centres introduce new technical demands on networks. Unlike traditional cloud environments, AI facilities generate far heavier east-west traffic patterns as GPU clusters exchange massive amounts of data during model training and inference.
This shift is dramatically increasing the bandwidth required between data centers and across regions. “When you cluster high-performance GPUs together, the connectivity demand rises very quickly,” Hall explains. “We are already seeing significant increases in backbone capacity requirements and much higher traffic volumes moving between facilities.”
Hyperscale cloud providers are also pushing the boundaries of network technology to support these workloads. Many are adopting next-generation optical systems capable of delivering speeds of 800 gigabits or even 1.6 terabits between data centres.
“They are constantly looking for ways to scale capacity faster,” Hall says. “That includes activating multiple fibre pairs and deploying advanced optical technologies to move data more efficiently.”
Despite the rapid expansion of AI infrastructure, Hall does not believe fiber capacity will become the primary constraint. With proper planning and investment, existing terrestrial and subsea networks can be upgraded to support far greater bandwidth.
Instead, the more pressing challenges may lie elsewhere. Power availability in major metropolitan areas could become a constraint as AI workloads grow, while the global shortage of skilled AI and cloud specialists may also slow deployment.
“The demand for AI infrastructure is growing extremely quickly,” Hall says. “If anything becomes a bottleneck, it may simply be the speed at which demand expands.”
Sustainability is also becoming a key consideration for network providers. As hyperscalers push for greener operations, companies like Ciena are focusing heavily on reducing the energy footprint of network technologies.
Through continued investment in research and development, newer optical systems can deliver significantly more capacity while consuming less power and occupying smaller physical footprints. This not only improves efficiency but also reduces cooling requirements within data centres.
Looking ahead, Hall believes the key to success lies in adopting what he describes as a “fibre-first” approach to AI deployment. Connectivity planning must take place alongside data center development rather than as an afterthought.
“Too often organisations focus on building the data center and deploying the GPUs first,” he says. “Then they realise they need to connect it to the wider ecosystem. Connectivity needs to be designed from the beginning.”
Ultimately, Hall sees the region’s AI ambitions as achievable precisely because governments, technology providers and hyperscalers are moving in the same direction.
“The Middle East is uniquely positioned,” he concludes. “With the right planning around power, fibre and talent, the region can build one of the world’s most advanced AI infrastructure ecosystems.”
