You're talking like these are the same people ideating and then dropping products over and over.
No: the people who build each product, even at Google, certainly enter a "nurturing and responding thoughtfully to user feedback" phase with that product.
The "problem", if you want to call it that, is that there are always new product people joining the company. And what the heck are they going to do to get recognized and promoted, if they don't build a new fiefdom to call their own?
The actual question I have about Google and its business strategy, is why these new products always seem to displace old products (and their teams), rather than resulting in the internal equivalent of a merger.
I can see good reason to have four teams building four distinct web VoIP clients, if one of them is the "flagship" one and three of them are experiments. I can also imagine one of those "experiments" getting really popular. But why should that result in the "flagship" dying? Why not just merge the teams together, and put the team from the "experiment" in charge of the pooled UX design talent for a new release of the "flagship" client?