AI Factory Imperative Series: Cooling Imperatives for the AI Factory

Thermal Strategies for Scaling AI Infrastructure.
Heat is the defining operational challenge of the AI factory. As rack densities climb and GPU clusters push thermal output to unprecedented levels, traditional air cooling approaches are reaching their physical limits. The data center industry is at an inflection point — and the path forward calls for a fundamentally new approach to thermal management.

In this session, David Toralez walks through the five thermal imperatives that define AI-ready infrastructure: the transition to liquid cooling, CDU (Coolant Distribution Unit) strategy, fluid chemistry considerations, hybrid air-liquid integration, and the operational collaboration needed between IT, facilities, and engineering teams. This session provides a practical framework for understanding how to design, deploy, and operate the cooling infrastructure that AI demands — at any scale.