I agree with your point, but looking at the projects,there's a clear difference between the companies: most of MS's projects(maybe except hololens) have large business value risk, on top of the technical risk.

On the other hand, most of Google's projects don't have much of a business value risk. If they'll work well, most likely they'll make good deal of money(unless the competition wins - but i believe Google also thinks hard on making projects hard to copy , or some other substitute challenges the market - but hey what can you do?)