Quantum #99
Issue #99 of the weekly HPC newsletter by HMx Labs. HPC is riding AI’s coat tails, we get some new VMs, an example of how to do observability and I ask a lot of questions.
HPC is riding AI’s coattails. This time it’s not me making that claim (though I have in the past) but some real HPC heavy weights including Jack Dongherra.
In case you missed the memo, or your only exposure to HPC is in in national labs and universities where buying HPC specific hardware is still the answer, the rest of the world moved to using commodity hardware to build supercomputers long before the AI hype train started rolling. In financial services this has always been accepted as the only way to run large compute clusters in fact. In some regards the explosion of AI focused hardware is actually great news because it increases the number of options available. (Yes, I’m aware it’s also cannibalising FP64 GPU hardware).
For those of you that are new to this game though, including using LLMs as part of your HPC software solutions, I think there have been a few things in this week’s news that you’d do well to bear in mind. As the Next Platform also points out this week, with a very optimistic take reconciling the AI bubble with reality, exposure to a single vendor in these conditions could have serious consequences for your commercial health. If my social media feed is to be believed even Microsoft is busy addressing that risk with OpenAI (who recently sunset 4o again). Whilst commodity hardware and models can be a great way forward, don’t cede control of your environment in adopting them. Especially in regulated industries.
For our part, this week I’ve mostly looking for input from your rather than dropping what passed around HMx towers for wisdom 🤣 If you have some ideas on unifying multiple workloads into a single benchmark or how to solve our growing pains and figuring out what role we should be hiring for next then give me a shout.
In The News
Updates from the big three clouds on all things HPC.
This is a nice way to couch AI doomerism in some optimism. I guess that makes it balanced?
Timothy Prickett Morgan
I don’t think it’s a secret the HPC is riding AI’s coattails so it’s nice to see it being acknowledged
https://www.hpcwire.com/2026/02/12/hpc-is-riding-ais-coattails-so-now-what/
From HMx Labs
We’re wondering if we need to be hiring a chief of staff. Or maybe a founder’s associate. Or operation manager. Maybe an SDR? Yea we’re confused. Any tips welcome!
And since we’re asking for input this week, why don’t we just carry on and ask for your thoughts on what should go into a benchmark? The corresponding LinkedIn post has some discussion already
Know someone else who might like to read this newsletter? Forward this on to them or even better, ask them to sign up here: https://cloudhpc.news
