Well I'm sure the people working on it think one might take off. But if you're taking the pessimistic view it doesn't matter if they never succeed as long as the smart people are kept busy working for them.
It's a clever argument that might convince certain purse-holders, who are optimistic or pessimistic about the success, that it's in the company's best interest either way. I don't think Google's under full control of the sociopaths yet that this sort of "locked up" argument is needed, or is felt internally by very many upper level managers. It's also just not a great strategy if you're worried about potential future competitors given that employees are free citizens, not slaves, and California has no non-compete laws. Industry (let alone tech industry, let alone Californian tech industry) is littered with successful offshoots started by individuals and teams who used to work together at one company, and quit or were fired en masse to start working on something else in the same domain, taking their experience with them and owing their former company nothing. Facebook's strategy of "buy any threat" is a lot more sound, even if it too is vulnerable when the threat refuses to be bought.