As artificial intelligence (AI) becomes more deeply embedded in how organisations operate, the infrastructure behind these digital tools is undergoing a profound shift. Data centre capacity to power AI processes must increase, and rapidly, by expanding and upgrading existing facilities or building new data centres. The African data centre market is growing quickly: research places current estimated market value at USD 1.94 billion in 2025 and is expected to reach USD 3.85 billion by 2030.
This rapidly increasing demand raises technical challenges in data centre cooling systems. And at the centre of the evolution is liquid cooling, a technology long understood but only now gaining traction in large-scale data centre environments. According to Leon Kleyn, Technical Director: Mechanical, WSP in Africa, and Floris van der Walt, Senior Mechanical Engineer, WSP in Africa, the rise of liquid cooling is both a response to market demand and a strategic opportunity for Africa’s growing digital ecosystem.
Why liquid cooling, and why now?
The sudden relevance of liquid cooling is a direct result of the explosion of AI workloads. Traditional computing relied on CPUs that processed tasks sequentially. AI, however, depends on GPUs that process multiple tasks in parallel, which significantly increases energy use and, in turn, heat generation.
“We are reaching the limit of what traditional air-cooling systems are capable of, in terms of performance,” explains Van der Walt. “Liquid has three to four times the cooling capacity of air, so it becomes the next viable option.”
Liquid cooling itself is not new, but the industry has historically not used it due to high costs, perceived risks and complexity. Now, rising heat loads have pushed technology providers and data centre operators towards solutions that can handle greater thermal heat rejection safely and efficiently.
Managing risks and technical complexity
The biggest concern is leakage – liquid coming into contact with expensive electronic equipment. “Leaks are one of the biggest risks,” says Van der Walt. “You need strategies to detect and mitigate them.”
