Google buying Motorola is a strong defensive move against both Microsoft and Apple on two fronts: patents and user experience. Apple is vertically integrated and Microsoft controls Nokia (without having had to buy it). Between that and having filed or acquired a lot of patents, these two pose formidable threats to Android. Buying Motorola lets Google fight both of these threats in one go. As Larry Page’s blog post makes clear, Motorola has a huge trove of mobile IP, having invented the first portable handset and also the then smallest feature phone. Having hardware design capabilities inhouse will let Google more easily develop some reference phones that provide as well rounded a user experience as the iPhone and Windows Mobile 7 (although this will require as much if not more software UX discipline).
Nonetheless, there are many negatives to being in the hardware business, such as margin compression (caused ironically by Android) and real COGS (other than servers and bandwidth this is an unfamiliar concept to Google). Most of all, for Android to succeed it must be widely adopted by many handset makers and the last thing Google should want is for this acquisition to scare off the likes of Samsung and HTC. I therefore predict that Google will only retain parts of Motorola - the patents and the hardware design capabilities, but close down or sell off much if not all actual manufacturing and handset distribution (to the extent that Motorola still has any of these in-house to begin with). That is of course assuming that Google is not tempted to go overboard once they have this done and get into the carrier business as well as Jean-Louis Gassée suggested for Apple. Being vertically integrated would be a terrible idea.