There is one thing humans are good at: pattern analysis on a blink (a-la-Gladwell). Looking at one company after seeing 1,000 gives you some ideas on how that particular one would develop (that is the common VC argument). History repeats itself, after all. It is not rocket science, but it has been empirically proven. If you do proper pattern analysis, you get it right most of the times (works with the stock market too).
One quick look at the following picture and a pattern develops in front of my eyes.
The picture says that Android now sells in the US better than the iPhone. I have a hard time believing it by looking around me, but NPD has been quite reliable in the past, so we have to assume they are right.
Quick look at the mobile market today: there is an Apple operating system which is closed and no device manufacturer can put its hand on (it is only shipped by Apple). There is another operating system that is available to any device manufacturer who wants it (at low cost), and it can be put on a small, medium, large device (in fact, any connected device would do it).
Any pattern recognition? I bet. That's the PC business. One Apple operating system which was closed, and one Microsoft operating system that hardware manufacturer could adopt and ship at "low" cost (for the time). Apple was better and now they have 4% of the PC OS market share.
I feel history might be repeating itself. The major difference is that now the operating system must be open source (or you are out), and be zero in the BOM (sorry Microsoft), and that the amount of devices is actually ample.
Actually, the race to connected devices might be the key here. Apple can chase Android on smartphones or pads, but Android is getting on millions of "other" devices (set-top-boxes, home appliances, cars, and so on). Devices Apple does not care about. Devices which will run in the billions in a few years: they will give Android an enormous installed base, which could lead to a lock-in.
Lock-in... I think I have seen that before too.