I have a podcast template with 6 voice tracks, a dialog bus, SFX / music tracks and a final mix/check bus before master. I have 1-2 VST effects on each of these. Mixing down a 2 hour podcast episode will easily take an hour on a pretty decent machine - with 15% CPU utilization.
I realize this is complex, but couldn't the mixdown process be modularized to take advantage of multiple CPUs in the system, and thus probably reduce the mixdown time greatly?