One point that I think might be interesting wrt saturation in consumer computing demands is the possibility of different interaction paradigms driving a dramatic increase in demand.

The example I have in mind for this is something like Oculus Rift, or more generally, VR. While consumer demand for computing power may be nearing saturation under the desktop paradigm, other modes of interaction require DRAMATICALLY more computing power to perfect the experience.

So if VR actually takes off there may be continued, or even greatly increased, demand by consumers for increased computing power. This would help ensure continued investment in the cutting edge, because even small increases could impact the experience in significant ways.