Infrastructure leaders tasked with managing sprawling server rooms may be confronted by the realization that the ageing enterprise data center is not always cut out for the new demands of AI,/ML and high performance computing. In the era of supercomputing, some workloads may require 50-60KW per rack, changing how facilities will need to be both powered and cooled.

This session will outline the steps an enterprise operator faced with increasing rack densities should take to assess the viability of liquid cooling whether keeping servers in-house or on-prem or relocating to a liquid-ready colocation facility.