Well, I don't really doubt the idea that justifies the department's existence is that Microsoft Research will produce new developments that can be used in money

Well, I don't really doubt the idea that justifies the department's existence is that Microsoft Research will produce new developments that can be used in money-making products, but the research itself is not necessarily linked to revenue in any real way. For example, the Haskell programming language and the experimental Singularity OS are things that really have no chance of ever themselves making Microsoft money, but Microsoft Research drove their development nonetheless. The most obvious underlying motivation is not finding revenue streams, but pushing the state of the art in ways that Microsoft can learn from and use to improve other projects.