Proposal-writing talent is unevenly distributed among researchers, and so is grant money. Furthermore, there are strong positive feedback loops where grant money leads to more publications, bigger group size, and increased reputation, all of which make it easier to get subsequent awards.
It struck me, then, while skimming the list of Google Research Award winners from 2010, that just about everyone on this list (certainly the people I know on the list) is already highly effective at getting grant money. In general (I claim) the research output of good grant-writers is not limited by money, but rather by the fact that there are no more hours in a day and by other personal and institutional factors. $6M of Google’s money spread over 112 awards is about $50,000 each: a relatively small amount that will not make a qualitative difference in the kind of research performed by a well-established, highly-funded scholar. It seems likely that Google could find a way to give $6M to the research community that would lead to a much greater impact — for example, giving $50K to each of 112 postdocs or first-year assistant professors. For these people, the money may well create a qualitative difference in their work, providing part of a year of funding for the postdoc or a year of graduate research assistant support for the prof.
Related: Michael Nielsen, in a great blog post, observes that that best results might be obtained by giving out grant money randomly.