Having huge sprawling swarms of VMs is, for some teams, a problem to be solved, not a fact of life to be designed around.
If I understand it right: VMS were not there because people needed VMS they were there because people needed compute.
We moved everything to k8s and we were able to do this because k8s can