Event Summary

AI infrastructure is pushing server power densities beyond the limits of conventional thermal management. As accelerators approach multi-kilowatt device power levels and rack densities exceed 100 kW, cooling architectures must evolve to support reliable and scalable deployment. In response, two-phase direct-to-chip cooling is evolving, using distributed and precise liquid delivery to improve control, scalability, and system integration. By enabling phase-change heat transfer directly at the silicon interface, this approach supports high heat flux removal, passive hotspot mitigation, and efficient heat transport—without the complexity and pumping requirements associated with flow-boiling systems. This webinar will explore how two-phase cooling, enabled through controlled liquid introduction, can serve as a scalable, production-ready foundation for next-generation AI infrastructure -supporting high-density GPU deployments while maintaining operational reliability and cost efficiency. The session combines strategic market context from IDTechEx with production deployment insights and technical guidance from ZutaCore, bridging industry trends with practical implementation pathways. Real-world implementation considerations, including system architecture, facility integration, and deployment models, will also be discussed. This webinar is designed for data centre infrastructure leaders, architects, and engineers responsible for deploying and operating high-density AI and HPC environments. IDTechEx Principal Research Analyst Dr Xiaoxi He is presenting alongside Shahar Belkin, Chief Evangelist at Zutacore and Cliff Grossner, Ph.D., Chief Innovation Officer, Open Compute Project Foundation (OCP). You can register using the link above.