Not really, what disappears is the Facebook "potential" to make money, and the valuation is done thinking at this possibility. Now if you look at it in this way, billions of people using Facebook can be monetized probably much better in the long run compared to selling a new HD from time to time to millions.
Actually, it's not purely opt-in. Remember when our tax money "rescued" the banks because of their reckless lending/investments less than two years ago?
That potential will be translated in quarter results sooner or later. I think the gap expressed in potential and real stock value between customizers who use your product and non paying users who consume your product is too big nowadays and will be narrowed abruptly.
"""billions of people using Facebook can be monetized probably much better in the long run compared to selling a new HD from time to time to millions."""
As if this has worked well in the past for other hyper-valued web crap...
(Which reminds me, anybody remembers that ole VA Linux company? Where are they now?)
VA Linux is now Geeknet and is basically ThinkGeek and a few other minor properties (/., sourceforge, freecode). They've shed all of their other business lines, hardware, consulting, Animation Factory, etc.
Virtually all of their business news for years has been ThinkGeek related (which is doing pretty awesome), and various corporate level failings such that all of their other properties have always been a drag on the company, frequently throwing it into the red despite having an extremely profitable e-commerce division.
If Facebook disappeared tomorrow, nothing of value would be lost and another web site would take its place.