About Olds Research

Olds Research is my independent analysis effort focused on HPC, AI systems, and data center technologies.

I examine what I call the Data Center Perfect Storm — the exploding power demands, cooling challenges, and efficiency pressures that are about to hit enterprises, governments, and large organizations as AI features get embedded into real production workloads. These are the same problems that high-end HPC shops have been wrestling with for decades: extreme power density, heat management, and the need to run expensive accelerator-based systems at high utilization.

My goal is to bridge that gap — to help non-HPC-centric data centers understand and take advantage of what the HPC world already knows about building and running high-performance infrastructure efficiently.

Here's the deal: 

I Don’t Spin.

Most analysts play it safe. They write careful reports that never offend potential clients. They say one thing in public and something else at the bar with their buddies.

I’m not that guy.

After 25+ years in HPC and high-end computing, I call things the way I see them — based on real data, real experience, and real research.

If something is BS, I’ll say it’s BS. If something is genuinely good or even great, I’ll call that out too — loudly and without hesitation. If it turns out I’m wrong, I’ll admit it and say what led me astray.

This site exists for people who are tired of vendor-flavored “insights” and corporate-safe analysis. I exist to address the Data Center Perfect Storm head-on.

The goal is simple: understand what actually works in the real world.

That means focusing on real systems, real workloads, and real constraints — power, cooling, utilization, and cost — not just how things are described in presentations or press releases.