Anyway, I think you raise an interesting analogy: working on 20% projects at Google, then using the company's resources to launch the product, does have a number of parallels to VC funding for startups (note: I'm advisor for Google Ventures, so I have some experience with both worlds). In a sense- and nobody everybody will agree with me- Google as an employer is a low-risk, low-capital way to launch my products. Larry and Sergey already took the risks (launching a company with no clear monetization strategy), they figured out a monetization strategy, and now they invest their capital in speculative projects.
Anyway, in my case, after it seemed like my project was in good hands and ready to be a product, I looked for something else interesting to work on. I think the main problem I have working here is that there are too many cool projects I could work on, learning from experienced SWEs and SREs, but I have to stick to one.