VMware in the enterprise was not about just running VMs, but about having shared and distributed copies, snapshots, backups, hot standbys, ability to live migrate a VM to another host, clustered storage, etc.
It allowed you to give all the modern niceties of systems management to apps that simply cant run in a modern tech stack.
For example, about a decade ago I helped build out a 911 dispatch that was split between multiple sites across the county, but where each was a backup to one another. There was actual hardware necessary that be presented via PCIe and USB pass through into the VMs to let CAD, GIS, AVL, Phones, and Radios work. But, that design also sucks for reliability, as simply needing to reboot a system means coordinating with other dispatches about failing the whole call center. By abstracting that physical hardware through the hypervisor, you could swap over instantly between those VMs.
If the respective vendors had a better tech stack, none of that would be necessary⦠and if we were to rebuild today, thereâ(TM)d be a proper server cluster to mange it. But⦠keeping these pre-existing systems operational is important as replacing all the things every few years in support of newer and better tech stacks has historically been more expensive than simply paying the licensing fees to Broadcom.
Now? I would expect wed pull the trigger and do some upgrades because price is now roughly equal.