Equinix’s Kamel Al Tawil discusses the cooling, power, and sustainability challenges shaping the future of AI-ready digital infrastructure in the UAE and wider region.
Artificial Intelligence is rapidly transforming the data centre landscape, driving unprecedented demand for high-performance computing infrastructure capable of supporting GPU-intensive workloads. Across the UAE and the wider Middle East, organisations are investing heavily in AI-ready digital infrastructure to power everything from generative AI applications to advanced analytics and autonomous systems. However, this surge in compute capacity is creating new challenges around power consumption, heat management, sustainability, and operational resilience.
Traditional data centre architectures, designed for far lower rack densities, are increasingly struggling to keep pace with the thermal demands of modern AI environments. This shift is accelerating the adoption of advanced cooling technologies, particularly liquid cooling, which is emerging as a critical enabler of next-generation AI infrastructure. At the same time, operators must balance growing compute requirements with energy efficiency goals, regulatory expectations, and long-term sustainability commitments.
In this interview, Kamel Al Tawil, Managing Director, MENA at Equinix, discusses how AI is reshaping data centre design, why liquid cooling is becoming essential for high-density environments, the key considerations for successful deployment, and how advanced cooling technologies can help support the UAE’s ambition to become a leading regional hub for AI-ready digital infrastructure.
Interview Excerpts:
How is the rapid growth of AI and GPU-intensive workloads reshaping data centre design and cooling strategies in the UAE and wider region?
As the Middle East accelerates its AI infrastructure investment, AI and GPU-intensive workloads are shaping data centre design by driving a rapid increase in power density. For example, data centres in the UAE consumed 3 TWh in 2025 alone, representing approximately 2% of the country’s 173 TWh total electricity demand. This share is expected to rise further as AI adoption expands across the economy. Managing this level of intensity requires a well-equipped infrastructure and advanced cooling systems to ensure effective heat transfer. Direct-to-chip liquid cooling (DLC) has emerged as the dominant approach for supporting high-density AI infrastructure, due to its ability to improve thermal efficiency by delivering coolant directly to high-heat components.
Why is traditional air cooling becoming harder to scale for high-density AI environments, and where does liquid cooling offer a stronger advantage?
Traditional air cooling is becoming less viable as GPU clusters reach significantly higher thermal loads, requiring new approaches that can dissipate heat more efficiently. Modern GPUs can exceed 200 kilowatts per rack and are expected to reach up to 1 megawatt, compared to the 5-10 kilowatts per rack environments that data centers were designed to support a few years ago. At these densities, air cooling is no longer sufficient. Liquid cooling offers a stronger advantage because it uses a liquid, such as water or a refrigerant, to cool the data centre. This enables the cooling system to be positioned closer to the heat source, allowing it to support significantly higher-density environments while using less energy and reducing reliance on fan power. In addition, advances in piping technology have reduced operational risk, with a very low probability of leaks. Liquid cooling also enables the consolidation of more computing power into fewer racks, which means less equipment to buy, operate, and eventually replace.
What are the key operational and infrastructure considerations data centre operators must address before deploying liquid cooling technologies?
Deploying liquid cooling requires integrating cooling into the overall AI and infrastructure strategy. These choices need to be made early, as they directly impact how easily organisations can scale high-density AI workloads later. In most cases, designing for liquid cooling from the outset is more effective than retrofitting existing environments, which can be complex, costly, and higher risk. All new Equinix data centres are, therefore, designed with this option in mind. Another key consideration is close alignment across the full ecosystem, which comprises customers, OEMs, and colocation providers, to ensure specifications such as temperature, pressure, and flow are properly coordinated.
“Systems must be designed with coolant distribution units that manage heat exchange between facility chillers and server racks while maintaining redundancy to mitigate unplanned cooling outages.”
How can data centre operators balance rising AI compute demand with energy efficiency, sustainability, and regulatory expectations?
Increasing regulation of the data centre environment, including requirements around efficiency and emissions, is shaping infrastructure design and reinforcing the need for compliance with evolving sustainability standards. Operators will need to balance these requirements while remaining flexible enough to support evolving customer and workload demands. Liquid cooling is one of the key tools enabling this shift, helping operators address increasing power, cooling, and hardware efficiency requirements associated with high-density AI environments. It also supports dematerialisation, enables wider operating temperature ranges, and improves resource efficiency, while creating heat reuse opportunities that enhance data centre performance. It also improves reliability by reducing thermal stress on hardware, which extends equipment lifespan and minimises downtime in mission-critical environments.
What role can advanced cooling technologies play in supporting the UAE’s ambition to become a regional hub for AI-ready digital infrastructure?
Advanced cooling technologies will be essential to enabling the UAE’s next phase of AI infrastructure growth, particularly as the country scales high-density compute environments to support national AI ambitions. As AI workloads continue to increase rack power densities beyond the limits of traditional air-cooled environments, advanced cooling systems will become critical for ensuring long-term infrastructure scalability, operational reliability, and energy efficiency. Liquid cooling can support this by reducing reliance on energy-intensive air cooling systems, enabling more efficient heat management, and supporting higher power AI deployments.
Source: Tahawul Tech