DATA CENTRES


Importantly, liquid cooling can be deployed incrementally. Many facilities are adopting hybrid setups, where some racks and rows transition to liquid while others remain air-cooled, allowing capacity to increase without wholesale redesign.


Hybrid architectures and room-level heat collection For most operators, the transition to higher densities will be gradual. Intermediate solutions play a critical role in enabling this progression while maintaining stability. Rear-door heat exchangers remove heat directly from server


exhaust air before it enters the room, reducing the burden on central air systems and helping to stabilise inlet temperatures. Row cooling systems provide targeted capacity within the aisle, addressing localised hot spots without extensive changes to the wider facility. At the room level, heat collection and airfl ow defi nition become increasingly important as densities rise. Perimeter- based air-handling units and thermal wall technologies are being used to shape airfl ow paths at the boundaries of the data hall, enabling controlled heat capture across both raised and non-raised fl oor settings. In mixed-density AI halls, these architectural elements help prevent recirculation and maintain predictable thermal conditions as localised and liquid-based cooling strategies are layered in. Together, these approaches allow operators to increase density selectively, extending the useful life of existing white space while preparing for more advanced cooling methods over time.


Rethinking heat rejection at the plant level Eff ective heat capture is only part of the equation. Once heat is collected, it must be transported and reused or rejected effi ciently, and this stage of the thermal chain is also evolving in response to AI workloads. As cooling strategies move towards higher operating


temperatures, heat rejection approaches are beginning to diverge. An emerging pathway for facilities designed to operate at elevated water temperatures is to trim the cooling. By maximising free-cooling hours and minimising compressor operation, trimming the cooling supports lower energy consumption and aligns well with future-ready architectures built around high-temperature setpoints. In this context, cooling becomes an adaptive extension of the IT load rather than a fi xed mechanical system, adjusting its behaviour to match real-world operating conditions. At the same time, centrifugal chiller technologies continue to play a critical role. Where guaranteed cooling capacity is required regardless of ambient conditions, or where lower supply temperatures remain necessary, variable-speed centrifugal chillers provide a stable and scalable backbone for heat rejection. Their ability to modulate capacity smoothly makes them well-suited to the variable loads introduced by AI, particularly in large, centralised plants.


Control as the integrating layer What ultimately enables this diversity of cooling approaches to function as a coherent system is control. Sensors, analytics


Download the ACR News app today


FANS


and automation now form the integration layer between IT load, thermal management and plant operation. At the unit level, controls regulate fl ow rates, fan speeds and local temperatures. At the facility level, supervisory platforms coordinate setpoints across zones, anticipate load changes and optimise plant operation. Advanced control strategies can maximise free cooling, improve part-load effi ciency and reduce unnecessary energy use. For operators, this level of observability also provides valuable insight. Real-world performance data can be used to refi ne designs, validate assumptions and plan future upgrades with greater confi dence.


Cooling strategy as a competitive advantage As AI continues to reshape data centre operations, thermal performance is emerging as a strategic diff erentiator. Facilities that can transfer heat effi ciently, adapt dynamically to workload variability and preserve fl exibility in cooling design will be better positioned to scale AI infrastructure without excessive cost or risk.


Thinking in terms of the thermal chain encourages a more holistic approach to cooling system design. It shifts the focus from individual components to system behaviour, enabling more resilient, effi cient and future-ready infrastructure. For the cooling industry, this represents both a challenge and an opportunity. As AI workloads push thermal limits, cooling systems are no longer just supporting infrastructure. They are becoming a defi ning factor in data centre performance, scalability and long-term value.


"Traditional reliance on bulk air movement struggles to keep pace with high- density AI racks, where airfl ow


requirements rise rapidly and fan energy


consumption becomes a growing overhead."


www.acr-news.com • April 2026 15


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36