Hello @Natha Sahu, Kedara
The bandwidth requirements for Azure Stack HCI VMs using the Azure Arc resource bridge are generally specified per VM rather than for a set number of VMs. Each VM requires a reliable high-bandwidth, low-latency network connection to ensure optimal performance2. Specifically, for synchronous replication, the network between servers should have enough bandwidth to handle the IO write workload and maintain an average round-trip latency of 5 ms or lower.
I think this would largely depend on what you are doing with the VM's. The major tasks which would need bandwidth like downloading of Images from the Marketplace or Storage Account, so they are stored locally on the cluster.
When the VM is deployed it will be a small amount of data which sent to/retrieved by the Azure Arc Resource Bridge, the actual deployment of the VM would be from a local copy of the image. Once a VM is deployed and guest management enabled, then it would depend what extensions deployed and role of the VM. For example 100 VM's which did not have any extensions deployed would not need any ingress or egress bandwidth, only if they stopped, started, resized from the portal by this would be very small indeed. However if you had 100 VM's which were part of a AVD host pool, and had the AMA agent and were gathering lots of metric for performance monitoring and security, were getting updates directly from Windows Update via Update Manager and then had 1000's of session running on them, then the bandwidth would be considerably more.
If I have answered your question, please accept this as answer as a token of appreciation.