

HPE Taps Supercomputer Past for Its AI Future
Jul 22, 2025
In this engaging discussion, Trish Damkroger, Senior VP of High Performance Computing & AI at HP Enterprise, shares her extensive experience from Intel and Lawrence Livermore National Labs. She highlights HPE’s pioneering role in AI infrastructure, focusing on innovative technologies like fanless liquid cooling. Trish delves into the convergence of HPC and AI, the concept of sovereign AI and its relevance to national security, and the advantages of Slingshot technology for data centers. Her insights reveal how HPE is set to revolutionize the future of AI computing.
AI Snips
Chapters
Transcript
Episode notes
HPC and AI Infrastructure Similarities
- High-performance computing (HPC) and AI infrastructure are fundamentally similar except in precision requirements.
- Both need tightly coupled, dense compute with advanced liquid cooling due to silicon heat dissipation.
HPE's Large AI Deployments
- HPE deployed 348 racks with over 25,000 NVIDIA B200 Blackwell GPUs for a global media company.
- Another financial client uses 80 racks totaling 5,120 GPUs, illustrating HPE's large-scale AI capabilities.
Long-Term HPC System Planning
- Designing exascale HPC systems involves years of collaboration between customers, vendors like NVIDIA, and HPE.
- Early co-design enables software porting and readiness for future ARM architecture-based chips.