I hear you… On the other hand I really find this to be quite the exception in post-production at least here in the US. The stems are made specifically for the purpose of giving the client access to processed audio according to the “food groups” that they want. I really, really don’t see the point in sending a client a set of processed “source” tracks that all sum to the mix. Like, why on earth would the client need a checkerboarded set of 6-12 dialog tracks rather than one or two dialog stems? Makes zero sense to me.
The reason it takes longer is possibly that Nuendo was never intended to do what you are suggesting, at least not “efficiently”, so it processes things in series rather than in parallel (this is me guessing here). Because if you feed 12 tracks of dialog into a group and you want them all processed “independently on the group” then you either have to create 11 additional virtual groups for the offline render, or stack the 12 track-renders in series.