"Google is now adding about 500 workers each month"
Hmm...500 people. By my calculus, 500 people should be able to implement about 100 business concepts every 1-3 years. That seems off somehow, but...well, they can't all be system admins, can they?
Take a look at google labs: http://labs.google.com/
How many of those things make money? Some of them seem to be a part of a long-term Orwellian strategy:
http://webaccelerator.google.com/
...but I don't really see the revenue stream. What are they doing with all of those people except mocking up start-up environments and hoping for the occasional big hit? Wouldn't it be cheaper to just let the free market work, and buy other people's companies as needed?