Wait a minute ...
... this rather assumes the VMs are all on a fairly flat network, and there is a complete and sufficient set of VMs that can understand the egress/ingress network traffic involved in such a migration, and still function with everything they used to talk to from the original site. Think BGP updates, traffic volumes, DNS, AD DCs ... the list is substantial.
The security profile will probably look rather different when running in Azure. Why didn't the company just host their VMs there in the first place? This reeks of "tick the DR box for the regulator" with fingers and toes firmly crossed that it never gets tested with a real event.
Not convinced this is viable for anyone but SMBs with a few 10's of VMs, or organisations that pump out "the next app" barely alpha tested, who care more about ability to pump out "the next app" than any real service level for infrastructure/compute/storage services.