What is the utilization of a hyperscaler datacenter?

Posted by banana_zeppelin@reddit | sysadmin | View on Reddit | 9 comments

I was wondering. Does anybody know what the general utilization of a hyperscaler datacenter is? How many headroom does it have to absorb peaks in usage before offloading requests to other datacenters? 50%? 75%? 90%? What are the economics here?