C A S E S T U D Y

performance, minimises unplanned maintenance and most importantly, mitigates the risk of outages where the business impact stacks up very quickly.

How significant is that impact when something fails?

Densities have jumped from roughly 20 kW per rack to 160 kW and we’ re heading towards around a megawatt per rack in some designs. If a high-density rack goes down, you could be looking at hardware exposure in the millions, plus the revenue or service impact per hour. As densities rise, the cost of poor thermal management increases alongside it.

Which parameters matter most for coolant health and performance?

Glycol concentration, pH and conductivity tell you about corrosion risk and contaminants. Temperature and differential pressure indicate how effectively heat is being absorbed and rejected, as well as whether fouling is occurring. Turbidity highlights particulates or biofilm risk. Flow rates confirm each cold plate is getting the cooling it needs. Considering those together in real time helps to understand if the loop is behaving as designed or drifting towards trouble.

How does Ecolab’ s platform support here, especially 3D TRASAR?

3D TRASAR Technology for Direct-to-Chip Liquid Cooling is a proven platform we’ ve applied in liquid-cooling environments. It continuously measures the fluid state, interprets trends to flag emerging risks then drives timely corrective actions. The aim is straightforward: keep the loop clean, chemically stable and thermally efficient so compute runs at target performance without surprises. The goal is to close the loop between detection, diagnosis and response.

How does coolant health monitoring improve water and power requirements in data centres?

Coolant health monitoring supports optimisation of the whole chain from site to chip, not just pieces of it. Operators track water usage effectiveness, but if you only look back at monthly numbers, you’ re reacting late. When coolant health and resource use are visible live, you can fine-tune set points, extend asset life and intervene before a slight inefficiency becomes a significant resource drain or an outage. Performance and sustainability move together when the entire system is optimised. If you’ re not monitoring critical water and energy indicators in real time, they can start lagging. It’ s about understanding water use effectiveness( WUE) and power use effectiveness( PUE) in the moment to take rapid action.

What happens when you scale this across a number of sites?

You need asset, site and enterprise-level views of your insights to help operators understand performance across their network. When real-time telemetry from each facility rolls into a standard dashboard, it reduces how often you need manual sampling, helps teams stay compliant with local rules and cuts unnecessary site visits. Just as important, it provides a unified data language for benchmarking sites, planning maintenance based on condition rather than the calendar and promoting best practices across the fleet – you remove redundancy and reduce complexity at the same time.

Looking ahead, what’ s the next challenge in monitoring cooling performance as data centres continue to scale?

Power densities will keep climbing, which means higher thermal loads. Efficient heat removal keeps chips at the clock speeds designers intend. The significant shift is in mindset – from cooling as a utility to a driver of performance.

Operators have long focused on the utility loop on the plant side. Now, together, we must master the technical loop at the chip as well. Coolant health becomes foundational. Measure it continuously, manage it actively and make it part of day-to-day operations. By doing so, you create headroom for the next wave of compute without sacrificing reliability or sustainability. �

16 www. intelligentdatacentres. com