Sure, there is going to be a lot of new silicon content in a car(but less, or much less cars ), but most of that silicon content would be in new areas - AI and machine learning, LIDAR, Radar ? , Vision - areas where nxp doesn't have any advantage, on the opposite - where startups and other companies like Google and Nvidia beat it.

To get those capabilities, wouldn't it be better for Quallcom to invest in R&D or buy startups ?

Another skill NXP has(via Freescale), is knowing how to build failure proof chips(via redundancy, etc). But my guess is that isn't that terribly hard skill to ackuire, and if not ackuire , there are already probably 3 companies with that skill(TI, NXP/Freescle, Renesas), and at least on startup - so enough people to partner or hire. And in any case , that knowledge is just the start , applying it to the complexities of AI is the real challenge.

And the third skill is of course - access to the automotive sales channel, possibilities of cross selling , etc. But will that be a big issue for the new tech ?