re: some of the convergence and future services, those may be drivers of processing power
at some point it remains to be seen if the software will actually get written to enable any of that, or if people will start using it. The engineering problems with developing the software and services are a lot bigger than the need for more CPU power.
Current cell phones don't have the processing power often required by an office workstation.
More importantly, they don't have full-size screens, full-size keyboards, mice, or a general purpose OS that can run Microsoft Office or a full version of Acrobat. Good luck figuring out how to hook up most printers with an Android phone OR iPhone. The present generation has already got plenty of processing power for most office workstation tasks; what it doesn't have is the software infrastructure, or the right sort of I/O or user interface. The first two are solvable; the third is, I think, never going to be.
I mean, you can dock a phone with a PC-like chassis, but at that point, why bother to use its processor at all? You can access its storage (either locally, or wirelessly) to make it appear to be a single OS/environment if you have the software set up, but for that matter, if you go all-cloud as you seem to espouse, you can just tie the environment to the cloud.
It does if, for example, you want to be able to walk into your house, wirelessly connect your phone to a big screen on your wall, to share video with your family or office, and then utilize the power present in your phone to edit that video quickly instead of needing to rely on a second device connected to the TV.
Why use your phone for that? You need negligible computer power within the phone to grab files off of it or push them down to it, and a not-quite-so-power-efficient CPU in the TV (or your laptop!) is much cheaper relative to the amount of power it needs.
For that matter, phones back to 2010 have plenty of power (and HDMI ports) for connecting to TVs, and while Intel's "WiDi" idiocy uses a lot of CPU in order to share any arbitrary screen contents, in principle sharing a video (that is already compressed) over wireless uses almost no CPU.
There's a concept of the right tool for the job; a small handheld device isn't the right tool for every job, and by the limits of the form factor and basic physics probably never will be.