As GPUs and CPUs continue to densify, they are rapidly surpassing the limits of air cooling. This, coupled with the advent of AI and ML which is being rolled out on a commercial scale, is accelerating the need for far more efficient types of cooling which can rapidly move heat away from the rack. So, could direct to chip be the solution?

This episode will take a critical look at the current state of cold plate technology, and explore the reasons behind its adoption, and the fundamental role it will play in cooling the chips of the future. Beyond its functionality for superior heat transfer, direct to chip systems also boast the ability of easier retrofitting, unlike its immersion counterpart. So, could this lead to a more streamlined rollout, a more incremental adoption of technology in an industry that is change-phobic, and a better ROI? Furthermore this session will question if, and how long, it could be until we see cold plate technologies being deployed at scale.