My colleague, Hernan Bruno did a great job (see http://blog.insead.edu/2012/02/what-does-a-100bn-valuation-for-facebook-mean/) estimating Facebook's value from data broadly available on the company as well as sensible assumptions about growth, retention, discount rates and other relevant parameters. The calculation is simple and considers wide and reasonable intervals. The conclusion of Hernan's blog is, to say the least, depressing. His estimates are far from the broadly mentioned $75-100 billion dollars, somewhere below $50 billion.
So where are the other $50 billion? I think they are in the yet undiscovered media technology that social networking reserves to us. The pattern is not new and over 100 years of (modern) media history has taught us that it takes decades before a new medium finds its killer application that allows a media format to extract value from the market. If you are skeptical, then check the early ads on radio in the 1920s or TV ads at the end of 1940s. They are a far cry from today's snappy ads. It took a long time for broadcast media to understand which advertising format would work for which category. With social networks we face the same situation: we all feel that there is huge potential in connecting a billion consumers and allowing them to share literally everything in their lives. The $50 billion question is: how can this social graph be turned useful for third parties who want to sell products to this population.