As AI drives up rack densities and cooling demands, Jon Healy, Regional Strategic Operations Officer, EMEA, at Salute, argues that water management must become a central part of data centre strategy.
Conventional cooling practices in traditional data centres fall short of today’s AI demands. The sheer speed of growth, and the widespread adoption that few could have anticipated, has left operators under significant pressure.
As racks run hotter than ever, many operators are now turning to liquid cooling systems because they can support greater efficiency and help maintain performance without overheating. But it is not a straightforward switch-and-replace. Some of these systems require substantial changes, as new infrastructure brings different maintenance needs, so teams also need to be upskilled to ensure successful implementation.
Without appropriate preparation, projects can face delays, inefficiencies and additional operational risk, rather than realising the full benefit of these innovations.
Where complexity comes in
Water, a primary and finite resource, is a major part of the cooling equation and is becoming an increasingly prominent issue. While regulation in this area is still developing, investors and customers are already asking difficult questions: how will large data centres remain sustainable if water use continues to increase? From an operator’s perspective, this means respecting community resources, understanding usage levels and trying to recycle wherever possible. If these stakeholder concerns are not addressed, data centres may face significant reputational and operational risk.
With increased power densities, grid constraints and planning delays, these challenges are becoming more complex. UK regulators, for example, are looking at the resilience and sustainability aspects of newer data centres, and similar pressures exist across Europe. The lesson here is that operators need to think holistically and beyond their immediate capacity needs. In practice, this means integrating cooling strategies with water management from the outset, and designing facilities with future workloads and resilience in mind, rather than focusing solely on speed to market.
Keeping sustainability commitments at the fore
Closed-loop cooling systems are gaining traction, and with good reason. They can support water reuse, reduce overall energy demand and help limit environmental impact. Currently, mid-sized operators are trying to stay ahead in a highly competitive market while facing pressure to adopt modern cooling systems at speed and maintain long-term reliability. Striking the right balance between speed, cost and sustainability will depend on careful planning, the right expertise and long-term operational readiness.
Modular construction is also helping operators keep pace with demand. Pre-built modules can incorporate liquid cooling from the start, reducing on-site construction time and limiting disruption to local communities. In areas such as the Thames Valley or the Midlands, where planning approvals can be slow, modular approaches may allow operators to deploy capacity more quickly without compromising sustainability. The key point is that modular designs can embed sustainability into the build rather than treating it as an afterthought. To meet investor expectations and evolving regulatory requirements while staying on schedule, operators need to combine modular deployment with careful water and cooling strategies.
Scaling to high standards
Innovation in water and cooling will be central to the success of today’s AI data centre infrastructure. The goal is sustainable growth, and that requires more than a compliance-led approach. Data centres need to build confidence among communities, customers and investors, who are increasingly focused on how digital infrastructure affects the world around them.
Operators that take a proactive approach and embed water management and efficient cooling into their designs from day one will be better placed to balance speed, resilience and environmental responsibility. This means thinking beyond immediate capacity needs and planning for how facilities will perform over the long term, under increasing scrutiny and rising demand. The combination of careful planning, transparent resource management and robust operations will help distinguish those who simply scale from those who scale sustainably.
