VM Thick Or Thin Provisioning For Video Surveillance?

When it comes to storage performance in a virtual environment, thick provisioning is always recommended regardless of the application. It means the space for a drive will be initialized or zeroed out from the beginning, so the hypervisor doesn't need to expand the disk over time. Thin provisioning has a growth penalty each time the hypervisor needs to grow the VMDK. And in the case that the NAS/SAN supports thin provisioning, it may need to expand the storage pool prior to allowing the VMDK to grow.

However, while I'm paying the penalty up front, provisioning a 14TB drive on a NAS, I'm wondering if the performance penalty of thin provisioning is all that bad. It sure would satisfy my desire for instant satisfaction to just go with a lazy zeroed drive and get on with things.