Well, more CPU power is only needed for running more CPU-hungry algorithms.
Axe II has got more CPU than Axe I because Fractal needed to write more complex algorithms for improving the sound quality. Unless Kemper discover a more CPU-demanding algorithm that noticeably improves profiling or playback performance, there's no need for more processing power.
Again, it's nothing to do with comparing the Axe and the Kemper (it's a non-sense comparison just like Apples and Oranges).
As said, it all comes down to the only one question: How close and accurate do you want it and can hear it?
(not between the Axe and the Kemper, but between a real Amp and the Kemper)
From some demo I heard, there was still some distinction from the original amp and the Kemper (of course, don't get me wrong, it's very tiny... and actually it looks to be the tiniest one I heard so far... but it might still be there. To really confirm it, I'm waiting to test it in real situation). Chasing the very last percentage of accuracy is always requesting a lot of efforts (from very complex algorithms and then DSP power). That's for this unique reason we might see in a future an Axe III, an Eleven Rack II, a Pod HD II, a Torpedo II, and even a KPA II, etc... several years from now.
However, even if there's still a place for improvements right now, it will pretty soon reach maturity because technologies are catching up with human beings; By 2020, a regular consumer computer will got more processing power than the human brain...