It's a university in the UK, so that's in GBP, not USD, making it even worse. So in 2025, pricing for 425k GBP, I could buy 16 Dell 7725's dual 64 cores (from a HPC perspective, more cores and memory bandwidth becomes an issue, might be different for virtualisation workloads) with disk all up the front, new 100Gbps switches, split them across two sites (we have two data centres), and *still* have change.
That is a redundant 1024 cores for VM workloads, which is frankly a way more than a university is likely to need. For sure, other projects get put on a back burner to do the migration, but at 425k GBP per annum, it is bye-bye VMware because there is literally no budget for that sort of nonsense.
That said, a 425k GBP per annum license fee hike concentrates the mind, and schemes to move off VMware on the same hardware with some judicious modifications all of a sudden become viable. Remember, we are a university, not a bank, so tolerance for risk is different.
For example, for the HPC, I was able to fit a cut-down working service on a single machine while the hardware for the VSphere cluster was reconfigured, Proxmox was installed, and the VMs were migrated back. During those three days, we were exposed to a failure on the single machine. Though for the critical VMs, it was less than 48 hours before they were on the Proxmox cluster.
Would I have ordinarily considered such a scheme, not on your life. The available budget meant it was the *ONLY* viable option.