Except for Intel the entire industry has gone to this model, and Intel profits from it. Chips have become a designer business, a software business. The algorithms implemented inside the chip make all the difference.
The answer is simple. Separate the design of chips from their production, at least internally. Consolidating four mobile businesses into one does not do that.
I believe there are two things Intel must do now that it has its mobile resources in one place:
Turn the process around. Have finished products drive the train. That means going to companies that make phones, that make tablets, that are in the WiFi services space, and and make sure you go to OEMs with finished designs, not just chips and ecosystems of software.
Think software. It's the software inside the chip that counts, not the number of gates, not the space between circuit lines. Put your money on better software in the chip and you'll make better, more customized chips.
The first is something I was advocating for almost a decade ago, when I was asked to look into Intel's business process during my days as a market researcher. It's basically what has made Broadcom (BRCM) so successful. The second is what Intel's newer rivals, like ARM, have been doing for many years now.
A successful company has a “not invented here” syndrome. Intel has become accustomed to dictating to the rest of the electronics market the designs of finished products. It still makes the bulk of its money this way, defining each year's PCs and servers based on what its engineers can do with silicon.
But that's not going to be good enough. What's happening today in mobile markets is going to happen in other markets soon. And if Intel doesn't get its act together – if it doesn't learn from what its rivals are doing right – all the re-organizations in the world will be of little use.
I'm still long Intel but I wonder for how long.