David - I think of it this way.
Intel started about 1.5 years late, and the Nokia fiasco resulted in a 6 month adder to that.
Moorestown was their Gen-1 design, and Medfield was their Gen-2. If you want to categorize Clovertrail, it's more of a Gen-2.5.
nVidia's Gen1 Tegra was a disaster, but Tegra 2 had some good initial design wins. Tegra 3 was more like their Gen-2.5 design, and it happened to be the right part at the right time, and scored some strategic design wins when Android tablets were first starting to ramp. It did rather well.
However, nVidia's problem with their Gen-3 is that they went too aggressively and slipped the design. It's Q3, and still no Tegra 4, which was promised in Q4'12, then Q1'13, then (for sure this time) Q2'13. Well, now Q2 has passed - and there's certainly no Tegra 4i, either, which at this point seems like it can be positioned no better than an entry smartphone part to compete with Mediatek.
Unfortunately, Qualcomm more than made up for nVidia's execution, and clean-swept the majority of the Android tablet market with their Gen-3 (or Gen-3.5, depending on how you want to look at it) Snapdragon 400/600/800 lineup. Bravo to them.
On to upcoming products - Intel will have their Gen-3 Silvermont based cores, on top of their next-generation SOC fabric, and have a chance to compete again. And I think the timing could not have been better.
Qualcomm is going to be at the whim of TSMC's 20nm process, which will have the usual yield issues through most of 2014. Even if Qualcomm executes well on 20nm, I expect to see supply issues plague them through the next design cycle, and Intel will be targeting many OEMs to look at them as a reliable second supplier.
With nVidia shipping an n-1 generation part, and Samsung serving their own interests, Intel is a credible choice. I think that translates into some very achievable wins throughout the next year - including their Merrifield part, which I have high hopes to see in U.S. bound phones.... :-)
Market Data 
Markets 
AI
