Kyle Geib

Thermal management is a priority as data centers race to add AI capacity. Operators are weighing the pros and cons of single-phase vs. two-phase direct liquid cooling solutions as they seek a quick, effective approach. Single-phase cold plate cooling is the default choice and likely to dominate in the next five to ten years because it's easier to deploy, costs less, and aligns with future silicon roadmaps.

The magnitude of demand for AI has surprised everyone. McKinsey predicts that global demand for AI capacity could triple by 2030. And it's created some challenges for data center operators, who are rushing to modernize power and cooling infrastructure to support the technology. One way to accelerate data center deployments is through AI pods - prefabricated infrastructure blocks that contain all the power and cooling required for AI. Speed to market: What's the rush?

As data centers deploy liquid cooling systems to support high-density AI workloads, operators need to decide on coolant distribution unit (CDU) placement. The main question is whether to deploy in-rack CDUs or floor-mounted CDUs. Making a choice comes down to the specific requirements of the liquid cooling architecture in each implementation. The CDU acts as the hub of the cooling infrastructure, circulating liquid from high-density hardware in a closed loop to dissipate heat.