As for returns, the average value of YC companies that are more than two years old is over $100M. I'm not aware of any other model that even comes close to that.
If anything, I'd like to find more ways to fund even more extreme moonshots (e.g. http://techcrunch.com/2014/08/14/y-combinator-and-mithril-in...). Funding an actual, literal moonshot would be wonderful.
I see this effect here in Australia where founders and startups are trying to follow the YC model to success despite the ecosystem here being really different. The thing I really love about YC (apart from HN) is that you guys are trying to do things differently - the last thing the world needs is a 1000 YC clones all following your lead and nobody trying different approaches.
I too would like to see a literal moonshot. I have thought a lot about this topic over the years and I think everyone is stuck because they have been concentrating on the wrong area. The cheapest component in the whole process is the humans - the way to get a real moonshot off the ground (sorry for the pun) is to put risk back. There are plenty of people willing to be heroes so why not optimise everything around that - the engineering costs go way down if you are willing to tolerate a high failure rate.