Best Practice for Cooling and Powering Large Scale AI Clusters
There are those that believe AI has reached a pivotal moment, like the smartphone in the previous decade, making AI accessible to everyone. This will create a large increased demand for data center facilities, increased expectations on power consumption, and new methods for cooling ITE. The difficulty is that currently there is very limited data center white space available for expansion, power grids are not ready for the challenge, and alarmingly current data center deployment techniques can leave more than 30% of the available whitespace stranded. Innovations to meet the needs of AI are imperative covering data center design, cooling environment engineering, power generation for data centers and methods for optimizing deployments of ITE. This panel will look at future technologies, guidance and best practices needed to deliver on the data center facilities to host AI and HPC Clusters.