In this episode of IT SPARC Cast, Lou brings you a packed 8-minute walkthrough of the biggest themes and technologies from SuperComputing 25—the largest and busiest HPC show he’s ever attended.

In this video, Lou covers:

🔥 Cooling Wars: immersion cooling, PG25 liquid loops, cavitation risks, phase-change fluids, and long-term hardware reliability.

🧠 CXL & Memory Expansion: shared GPU pools, multi-host memory fabrics, and how CXL can extend server life.

☁️ Hybrid Cloud AI Platforms: two research-born vendors (including one FedRAMP-compliant) redefining HPC + cloud orchestration.

⚡ Infrastructure Giants: the mind-blowing cooling and power equipment that will shape future enterprise data centers.

And John reads out Listener Feedback regarding AlmaLinux as the successor to CentOS.

If you want a concise, expert-level briefing from the SC25 show floor—this is the one to watch.

What it on Youtube Here - https://youtu.be/Ve57fs7efFY

00:00 – Intro

01:08 – Greeting from Super Computing 25

Lou sets the stage after returning from SuperComputing 25, describing the massive scale of the show, packed floors, and how SC25 has effectively replaced events like Interop and SuperComm.

NEWS & TECH BREAKDOWN

02:22 – The Major Theme: Cooling, Cooling, Cooling

Lou explains that cooling dominated the show, with two primary approaches emerging:

1. Immersion Cooling

•Full-system submersion in mineral oil or engineered fluids

2. Active Liquid Cooling (PG25 Mix)

•Issues explored: erosion, cavitation, biological growth, thermal cycling, solder fatigue

3. Phase-Change Approaches

•Solutions that vaporize at fixed temperatures (e.g., 55°C boiling point phase-change fluids)

Why It Matters: Enterprise hardware longevity, reduced thermal stress, and predictable cooling efficiency.

05:41 – CXL & Memory Expansion: The Future of Server Life Extension

Lou discusses a major standout category: CXL (Compute Express Link) technologies allowing:

•Shared memory pools & GPUs across multiple hosts

•Extending server life by adding external memory instead of replacing hardware

•Switching architectures enabling dynamic assignment of terabytes of memory to GPUs

Enterprise takeaway: “Do more with less” becomes practical—critical during recessionary or budget-tight periods.

⸻

07:55 – Hybrid Cloud AI Platforms

Lou meets with two research-born companies offering advanced hybrid cloud and orchestration stacks:

•One FedRAMP-compliant, built for U.S. federal and defense workloads

•One European research derivative, designed for container-heavy hybrid environments without VMware reliance

These solutions focus on orchestration, HPC-to-cloud overflow, container scheduling, and distributed compute for AI.

09:19 – Wrap Up

John closes by encouraging viewers to watch the upcoming shorts and emphasizing how SC25 showcased the next generation of enterprise-class tech. He also covers Listener Feedback on our first short from SC25 regarding AlmaLinux

Social Links

IT SPARC Cast

@ITSPARCCast on X

https://www.linkedin.com/company/sparc-sales/ on LinkedIn

John Barger

@john_Video on X

https://www.linkedin.com/in/johnbarger/ on LinkedIn

Lou Schmidt

@loudoggeek on X

https://www.linkedin.com/in/louis-schmidt-b102446/ on LinkedIn

Hosted on Acast. See acast.com/privacy for more information.