Data centre construction is rapidly transforming, driven by the increasing demand for AI-powered computing. This year, companies are speeding up development timelines, optimizing energy use, and innovating to tackle the ongoing shortage of skilled IT professionals. Data has long been the backbone of analytical insights and tech progress, fueling the relentless expansion of data centres.
Today, significant regional investments in digital infrastructure further propel this growth. Consequently, data centres are expanding as the region heavily invests in IT infrastructure. According to a recent study, Saudi Arabia operates 24 data centres with a total capacity of 123MW and plans to build 37 more. Meanwhile, DC Byte reports that the UAE has 67 data centres with a capacity of 899MW. These expansions have surpassed many expectations.
Adapting Infrastructure to AI Needs
AI is swiftly transforming data centre requirements. Its expansion mirrors the Internet boom of the 1990s, but AI is advancing even more rapidly. Initially a simple automation tool, AI now enhances business processes, customer service, and problem-solving.
Leading tech firms are rapidly constructing data centres to meet the growing need for AI computing. This investment is not just about increasing power but also about evolving AI’s application, management, and scalability. AI as a service is accelerating adoption in areas like customer service, financial planning, and decision-making.
Data centres are also leveraging AI to streamline operations. With a scarcity of skilled IT workers, AI tools enhance efficiency and simplify tasks. Generative AI now provides real-time answers and advice, enabling smaller teams to perform as effectively as larger ones.
Rising Power Demands of AI
AI has significantly increased power demands at data centres. A Goldman Sachs report forecasts a 165% rise in power consumption due to AI. The International Energy Agency states data centres account for around 1% of global electricity consumption, with their annual electricity use being about half that of household IT appliances like computers, phones, and TVs.
Due to limited additional power in traditional hubs, operators are leasing extra grid capacity or partnering with energy suppliers directly. Some are even exploring nuclear power as a reliable option. This shift is expected to continue growing into 2025 and beyond.
Focus on Energy and Sustainability
AI-powered data centres require substantial energy, raising concerns about sustainability. The region’s solar energy presents a valuable resource. Many operators are investing in renewable energy, and Microsoft aims to power its UAE data centres entirely with renewables by 2030. These initiatives will shape the future of AI-centric data centres.
Nuclear Power and Enhanced Cooling
Implementing small modular nuclear reactors provides a steady stream of low-carbon electricity; combining this with liquid cooling maintains demand in check. AI servers generate more heat than older systems, and traditional cooling with large fans and chillers consumes extra megawatts. Circulating water or dielectric fluid over heat sinks efficiently removes heat, cutting cooling electricity use by about a third. This creates a balanced approach: nuclear power cleanly covers base load, while liquid cooling reduces the load itself, allowing the facility to expand its AI capacity without exceeding energy or carbon limits.
Hans-Jürgen Niethammer, head of data centre market development at CommScope.
Enhancing Fibre Connectivity
As AI data centres expand, they require more fibre connections to meet communication demands. AI-driven GPUs need fast, complex links, boosting both cooling needs and fibre usage. New data centres are adopting denser fibre systems with compact connectors, supporting a transition to native 800G networking, which fulfills the bandwidth requirements of heavy AI usage.
Flexibility in Colocation Data Centres
Colocation data centres will play a crucial role in 2025, needing to remain flexible and scalable to meet evolving business needs. Standardized high-density fibre networks can enhance these centres’ efficiency. Simplified fibre setups reduce complexity and reliance on large IT teams. Manufacturers are developing plug-and-play solutions that simplify system management while ensuring reliability.
Future Outlook
The evolution of data centres in the AI era is accelerating. Both large-scale and multitenant facilities must expand fibre networks, employ advanced cooling techniques, and devise new power management strategies. The shortage of skilled IT professionals persists, but AI tools are bridging that gap. As AI becomes more integral to business, data centres must deliver the power needed to support these transformations. Ongoing innovation and adaptation will shape the future of data centres in the AI age.
The author is head of data centre market development at CommScope.
