This may come as no surprise to you, but the reports are in, and they show something you might've already guess - that Nvidia has the most chips in tablets, while Qualcomm has the most chips in smartphones. Nvidia has a third of the non-iPad market, while Qualcomm owns 42% of the market for smartphone processors.
Why does Nvidia have a lead in tablets but not in smartphones? First off, because Nvidia was scheduled for late 2011, and it was a quad core chip, which sounded great to manufacturers, while Qualcomm didn't have a very competitive chip until early 2012, and it was still a dual core, even though it turned out it was at least as good as Nvidia's chip.
Second, Nvidia usually makes tablet chips first, and then optimizes them a few more months for smartphones. So by the time the Tegra 3 was available for smartphones, Qualcomm's S4 was already there and available for manufacturers. And since Qualcomm has many more relationships with smartphone makers, it was inevitable that they would get the most contracts.
Qualcomm's S4 was also more energy efficient than Tegra 3, because it was made at the 28nm node, instead of 40nm, like Tegra 3, and because its modem was also integrated into the SoC. This is where S4 had a clear advantage in smartphones over Tegra 3.
Tegra 3 was also a smaller chip (only 80mm2 in size vs 100-120mm2 for other competitors), and since it was also made at an older processing node, that means it was pretty cheap to buy and put in $200 tablets. This is where Tegra 3 really started seeing some sales this last year.
So where does all that leave Nvidia and Qualcomm in 2013? I think Nvidia will continue last year's strategy, and it should be working just as well for them, while also appearing in other type of $300 or less devices, such as Nvidia's own Project Shield, OUYA, and others. I think we'll see Tegra 4 in more devices where power consumption is not a huge concern.
In the same time, while Qualcomm should still do reasonably well in 2013, I think it won't be as good of a year as 2012 was, at least in terms of momentum and popularity (they might still get higher revenues, profits, due to the smartphone market growth). The reason I think that is because we will see more Cortex A15 and Cortex A7 big.Little chips in the market, which should outperform Qualcomm's own chips, and we'll see them in both smartphones and tablets. Either way, in 2013 we should see a significant boost to performance in mobile chips.