That is interesting experiment. I would think there are two basic limitations to ultimate speed, heat dissipation and basic gate propagation delay. The first is being addressed here, but I wonder what the intrinsic propagation delay limit would be?
Lefty