The nice thing about this type of technology is that you have enough paranoid people in the US to market to initially (assuming you're in the US or that the US is easy for you to sell to, and no, I'm not trying to be condescending of people that are paranoid). Another interesting bit could be to donate one piece of the technology to opposition movements in countries run by dictators, when somebody buys your devices. How you'd smuggle it is one question, but I'm sure there are NGOs that would gladly help you.
Fred Wilson might be interested in it too: http://news.ycombinator.com/item?id=2158529
And that's a cool suggestion, applying the TOMS shoe one-for-one model.
In order for a technology to become mainstream there has to be an obvious, immediate benefit.
Piracy comes to mind, but the ISP's haven't cracked down enough yet for that to be a major enough issue.
Bandwidth caps could also be a prod, but I don't think people would be any more annoyed with those than the inevitably higher latency (between geographically disparate people) of a distributed network.
And then there's the bootstrapping problem.
