Quantum #101
Issue #101 of the weekly HPC newsletter by HMx Labs. A new HPC VM from Google, analysis of CPU based HPC in the cloud, more GPUs from AMD for Meta and GPUs in your storage cluster too by VAST.
Google dropped its new H4D, HPC optimized, VM instance with zero fan fare last week. So now we have two 5th generation (Turin) AMD CPU HPC VMs in the cloud, Released in the space of as many weeks. I thought Google would have made more noise about this but it went GA in a quiet release note and little else.
You’re waiting for me to drop some comparison benchmarks now aren’t you? 😆 Patience… maybe later this week?
Next Platform gave us a pretty nice analysis of CPU based HPC in the cloud by looking at AWS’ VMs, triggered I guess by the release of the HPC8a last week. It’s a shame it didn’t go wider and include the other two major CSPs but is a good read nonetheless
AMD seem to be cementing their position by inking a deal with Meta for GPUs to go along with last week’s CPU announcement. Either that or they’re getting good at playing the press release game and the deal does look remarkably similar to the one drawn up for OpenAI last year.
VAST Data managed to get a huge amount of column inches this week by including a GPU into their storage fabric. I guess that provides another use case for GPUs if the AI demand ever tails off…
I talked a lot about AI coding last week, which wasn’t really intentional. I guess it’s just what was top of mind. (It’s definitely getting harder to find and talk about HPC news that isn’t just AI news rebranded). What’s quite interesting is after having published our piece on vibe coding in HPC I came across several social media posts and videos echoing similar sentiments. This is an interesting shift away from the hype at the start of the year… Perhaps something I’ll talk about more in our next update on HAL.
In The News
Updates from the big three clouds on all things HPC.

Next Platform has a good write up on the AWS HPC VM types

AMD and Meta expand on their CPU deal from last week to include lots of GPUs in a deal that smells very much like that between OpenAI and AMD last year

VAST got a lot of attention, most of related to this
https://www.hpcwire.com/2026/02/25/vast-adds-gpus-into-clusters-with-cnode-x/
From HMx Labs
Lots of talk about vibe coding this week it seems, both from us in the form of an update on our HPC vibe coding experiment and using it to do COBOL migrations


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


