Intel Corp (INTC)

[wbmw]

The focus for them should be first and foremost on GRAPHICS, since this is an area in which I truly believe they have the potential to be the world's best at.

nVidia's GPUs are primarily slave devices, which puts them at the whim of Intel's platform direction, and disadvantaged by the PCI-Express bus. In my view, Intel will always have the opportunity to design a better graphics system design with integration. The fact that they haven't so far is a matter of their architecture and software maturity, as well as the die area and power they dedicate to it. And starting with Haswell, all of that is going to change.

nVidia's best beach-head is their Tegra line, where they can play the same integration game. Only problem is, their first shot at "owning" the architecture is with Tegra 6, where they can abandon ARM's implementation to use their own "Denver" implementation, and when they can align their graphics core to the newest "Maxwell" implementation. Problem is, by 2015, Intel is well into the mobile game (it will be the "tock" after Haswell, on 14nm), and will have had plenty of time to ramp up their graphics capabilities as well (and probably worth another "tock" on graphics as well).

I am hopeful that they are more sensible about this balance in the smartphone/integrated iterations of their chips, which given that they didn't try to stuff a low clocked Cortex A15 into Tegra 4i, seems to be the case.

True, but at the same time, A9 cores are not going to be sufficient in 2014. They'll be competing with 20nm A15s, 2nd gen Swift, Krait 300/400 cores, and not to mention Intel Silvermonts. It will be dead last among the premium contenders.

Ultimately, though, I see Nvidia being acquired.

Maybe, but I don't know why Intel would want them. What about opening up their IP to licensing, or their software team to services?