As AI workloads drive rack densities higher, direct liquid cooling is becoming essential. Yet attention often focuses on chips, servers, and cooling equipment, while the piping infrastructure connecting these systems receives far less scrutiny. Its design and material selection can influence not only cooling performance, but also project cost, deployment speed, and long-term operational reliability.
 
In this session, Mark Bulmer explores the piping system through a total-cost-of-ownership lens. He examines how material choices affect upfront investment, prefabrication and on-site installation, flushing and commissioning requirements, hydraulic efficiency, coolant integrity, and ongoing maintenance. Drawing on current industry guidance and practical project experience, the session highlights how decisions regarding coolant compatibility, air management, corrosion, passivation, and system cleanliness can create hidden costs and schedule risks.
Attendees will learn how a system-level approach—and the appropriate use of polymer piping—can support faster installation, more predictable commissioning, and lower operational risks throughout the lifecycle of high-performance AI data centers.