Understand this story
Anu Kätkä proposes that liquid-cooled data centers require accurate real-time measurement across both liquid-cooling and air environments. Related reports point out that liquid cooling expands operations from single air-side measurement to a dual domain of liquid cooling and air: direct-to-chip cooling delivers liquid to processors through CDUs and cold plates, and liquid-side parameters such as liquid temperature, pressure, flow rate, liquid level, leak detection, as well as coolant health indicators including pH, turbidity, conductivity and refractive index, all require dedicated instrumentation. Air-side management remains indispensable, and the coolant supply temperature usually needs to be kept at least 2°C above the measured ambient dew point to prevent condensation. The reports also note that AI training workloads are dense and inference workloads intermittent, and a cluster can go from idle to full load within minutes; only accurate real-time measurement can support dynamic CDU control.
AI-generated from public reports
Latest development · Report
Liquid cooling is changing the metrics data centers need to measureDevelopments
1 confirmed developments
Anu Kätkä proposes that liquid-cooled data centers require accurate real-time measurement
Source:Data Center Knowledge
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 2
Data Center Knowledge
Liquid cooling is changing the metrics data centers need to measure
Liquid cooling expands data center operations from single air-side measurement to a dual domain of liquid cooling and air. Direct-to-chip cooling delivers liquid to processors through CDUs and cold plates. On the liquid side, liquid temperature, pressure, flow rate, liquid level, leak detection, as well as coolant health indicators such as pH, turbidity, conductivity, and refractive index all require specialized instrumentation. Air-side management remains indispensable; the coolant supply temperature typically must be kept at least 2°C above the measured ambient dew point to prevent condensation. AI training workloads are intensive and inference workloads are intermittent, and clusters can go from idle to full load within minutes. Only real-time, accurate measurement can support dynamic CDU control.