On the other hand, if research for a particular moonshot ends up creating something that the other products can use (and become more profitable/successful as a result) - such as Google Brain - the moonshot itself doesn't need to succeed for the project to be considered worthwhile.