What I mean is that Google always has to do things the Google way whether that's an improvement or not. Whether this turns out to be a valuable contribution to the field will play out over the next few years. We'll see. But a company that makes their own DB, their own phone, their own programming languages, etc. is pretty much a prototypical "not invented here" company.
Google has some unique problems, they also have a unique position in that they probably have pooled the largest number of IT talent world wide.
That gives them clout enough to do these things, just like Erlang came out of a phone company (why, couldn't they have used some existing language ?) and Apple made their own phone (I'm sure they could have afforded a Nokia or two) google can and does have the budget to branch out in other fields.
That's not 'NIH', NIH is doing something even though better alternatives exist, for instance, writing your own bookkeeping software when you're a mid-sized company and you should be spending your time on serving your customers instead of writing bookkeeping software.
Google even makes their own servers, at their levels of scale a lot of things that do not make sense for smaller companies actually improve the bottom line.
Google suffers from many defects, bad customer support would be the top of my list, but NIH isn't one of them, at least not in this case.
That's almost like accusing Xerox in their heyday of having NIH, and I'd hate to think what the world of software would look like today if it hadn't been for that.