Why data centers are going liquid, with Endress+Hauser experts Liza Kelso and Cory Marcon

The podcast discusses the transition from air cooling systems to liquid cooling systems in data centers, driven by the increased power demands of graphically intensive computing, particularly for AI applications like ChatGPT. Historical air cooling via HVAC has become inefficient due to higher thermal outputs from advanced GPUs, which necessitate alternative cooling methods. Liquid cooling methods, such as direct-to-chip cooling and immersion cooling, are highlighted as effective solutions that not only enhance thermal management but also utilize significantly less water through a closed-loop system.

Managing liquid cooling systems comes with challenges, including maintaining temperature stability, ensuring adequate power supply, navigating supply chain constraints, and addressing workforce availability. The discussion emphasizes that large hyperscale data centers need to invest in robust instrumentation and infrastructure to effectively manage these challenges.

Temperature measurement plays a crucial role in the efficiency and reliability of cooling systems. High-fidelity temperature sensors are necessary to monitor the cooling loops effectively and prevent catastrophic failures in the data centers, emphasizing that precise and timely temperature readings can save significant costs related to downtime and equipment failure.

The TM611 temperature sensor from Endress+Hauser is spotlighted as a breakthrough product in surface temperature measurement. It allows for accurate temperature readings from outside the pipe without invasive methods, thus speeding up response times and reducing costs associated with traditional thermal well installations.

The conversation goes into detail about the necessity of keeping the water used in cooling systems clean and pure. Advanced liquid analysis technologies enable continuous monitoring of the water's quality, ensuring that it meets the sustainability requirements set by local communities and the operational needs of data centers.

Digitalization is identified as a vital element in enhancing the reliability of data center cooling systems. By streamlining data collection and providing real-time analysis, operators can track performance metrics, making proactive adjustments to maintain efficiency. This tech-driven approach also supports sustainability goals by providing transparency and accountability regarding resource consumption.

Sustainability in data centers is becoming increasingly important as communities express concerns about water and energy usage. Endress+Hauser explains how they are committing to science-based targets for sustainability, focusing on the measurable impact of their products to assist data centers in achieving their environmental objectives.