Why AI Data Center Cooling Needs to Get Smarter
Data Center Knowledge, Wednesday, October 14th, 2026
Smarter sensing, motor control, efficient power devices and edge AI let liquid cooling respond in real time to AI rack loads.
Texas Instruments' Patrick Zeng argues that rising AI densities require cooling systems that sense and adapt rather than run at fixed settings.
Precise flow sensing, such as ultrasonic sensors, can guide linear valves at each tray, while better leak detection is replacing resistive ropes.
GaN power devices can raise pump-system efficiency by up to 2%, fans are moving from 12V to 48V, and pumps will need to support 800V HVDC distribution.
Edge AI "sensor fusion" across temperature, humidity, pressure and flow data enables faster leak detection and predictive maintenance across any cooling architecture.