From what I understand about the Spectre/Meltdown patch, the impact to CPU usage is workload dependent; so there isn't one catch all value. I have read the impact can vary from anywhere between 5% and 20% increase in CPU times.
So assuming the impact is 20% in the worst case scenario. In VPM we can experiment with this by making Workload templates of each VM and put in a value of 120% for the CPU 'Percent of consumption' value. For 100 VMs this might be too tedious.
Do you have access to Capacity Manager? if so you can create a Data Center model and add the cluster and create 1 workload for the model. Then you could add a Resource Scaling scenario and put in a CPU factor of 1.2 to represent a 20% increase and see the impact.
HTH
>KIP