Panels Archive

Building and Managing Exascale HPC/AI Supercomputing Infrastructure


Moderator: Ben Di Qual (Everpure)
Panelists: Joe Greenseid (NVIDIA Corporation), Caroline Yap (Frontier Technologies Laboratory, Cambridge University), Christopher Zimmer (Oak Ridge National Laboratory (ORNL)), Benedict von St. Vieth (Jülich Supercomputing Centre (JSC)), Christine Luo (Oracle), Andrew Jones (Microsoft Corporation)

Abstract: Designing and operating exascale infrastructure for HPC and AI workloads is no longer a distant challenge — it is accelerating faster than ever, bringing unprecedented complexity in performance, cost, scalability, and security. National labs, hyperscalers, neo-clouds, commercial enterprises, and academic centers all face these problems, often arriving at very different solutions shaped by their workloads, regulatory contexts, supply chains, and economics.

This 90-minute panel brings together operators from across the sector to compare approaches, debate trade-offs, and surface the practical lessons most worth carrying across organizational boundaries. Topics under discussion include cost-effective scaling, exascale data movement, security and multi-tenancy, energy and power management, and operational resilience.

The panel delivers insights that benefit the broader HPC and AI ecosystem — turning isolated operational lessons into a shared, comparative view, aligned with the theme of SC26: HPC Unites.


Back to Panel Archive Listing