The whole point of true Type I Hypervisors is to put a bunch of them together in view to build-up a cluster farm. This, in turns, permits to enable virtualization-level High-Availability and automatic guest hot-migrations shuffling, to optimize the load across the farm. In fact, optimizing host’s CPU load also increases guest’s performance, as the hypervisor schedules guest’s jobs.