What is Hybrid air and liquid cooling?
Hybrid cooling runs two heat paths at once. Cold plates remove heat from graphics processing units and central processing units into a rack liquid loop, while fans carry heat from memory, drives, power supplies, and network devices into room air. The liquid passes through a coolant distribution unit (CDU) to facility water; the air passes through a rear-door unit or room air handler to a coil. It fits phased retrofits, colocation halls, and mixed fleets. Its main risk is treating residual air heat as a small fixed number. Liquid capture changes by server model and workload, while neighboring legacy racks still shape airflow. The facility needs both water and air capacity, coordinated controls, leak and dew-point management, and fault plans for each path. Ask the server maker for heat-to-liquid and heat-to-air data at expected workloads, including switches and power shelves. Model first-day and full-build conditions, then test what happens if a CDU, room fan, chilled-water branch, or sensor fails. Procurement should define control ownership across information technology and facilities, meter both paths, and preserve enough air turndown to avoid wasted fan energy. Compare whole-site power and usable rack capacity, not liquid capture alone.