Unlike deep-learning tasks, datacenter-grade (Tesla) GPUs don't have an advantage in mining over much cheaper consumer GPUs. As an example, a cloud-based server with dual Tesla P100 would get you 120..125 MH/s on ETH, and would cost $2,500 per month. Or, I can build a rig with consumer GPUs (1060/1070) for under $2,000 (one-time).
Interestingly, you cannot run consumer GPUs in a datacenter, as Nvidia driver terms prohibit it.
