There have been a lot of recent advances in solar and wind. Our group, in particular, has focused on the storage side of the equation, storing the energy more efficiently and reducing waste. Either way, we are confident we can power this thing via only solar in areas without any traditional power infrastructure. However, you're right that stabilizing existing power sources would be a big win. We have a LOT of experience with power issues in rural India, and technologies to mitigate the problems. We'll be using whatever solution best fits the individual deployment situation.
Hey, Kurtis Heimerl here. This work seems to have gone though the telephone game, so I thought I should make some slight corrections to the original (down?) sharable article. Firstly, I'm not a professor, I'm a graduate student (https://www.eecs.berkeley.edu/~kheimerl/). Secondly, there wasn't enough mention of everyone involved. This project is an offshoot of the fantastic OpenBTS work done by David Burgess et al, now working under the Range Networks (http://www.rangenetworks.com/) banner. A lot of work on the open source projects have been done by Alexander Chemeris and Thomas Tsou, and they should be given credit as well. Lastly, the project is still under heavy development. It's worth noting the "prototype" used by Mobile Active had no clever power tricks, and was just a software modification of OpenBTS. We'll have something more substantial soon. Past that though, I'm happy to see people quickly understand what we're going for here. Great to know others think your work is interesting. We'll also be deploying a sample "BTS Application" at burning man this year. Check it out: (http://tier.cs.berkeley.edu/drupal/burningman)