This is topic that was discussed a lot some years ago as a huge potential when the mobile devices would have enough computing power. I returned to it now to see how the vision is developing. Not too well if Google is trusted. A search for "mobile grid computing" does bring back over a million hits, but many of  the links point to research papers or articles from 2003-2005.

I threw a question into the LinkedIn community, and got some responses back. Two respondents said that the problem for using mobiles as hubs for grid computing type of tasks is not the the CPU, but battery life. At first thought this sounds like a too simple answer, but could it really be so?

When asking the question "When do we have enough CPU power for meaningful mobile grid computing?",  I was also trying to see if someone would come back with a calculation of average computing power of a PC and a smartphone, add a growth multiplier, multiply by number of devices and come up with a year when the global computing power from mobile devices would exceed the global computing power of personal computers.

Perhaps it is impossible to estimate, as the definicions of a "mobile device" and a "personal computer" are already vague and keep changing all the time.

***
Post Scriptum: See the wrap-up from LinkedIn answers.