I have been fairly adamant in my stance that, as great as it would be for Intel (INTC) bulls such as myself for Apple (AAPL) to adopt Intel's "Atom" products for its iPhone, it is not likely to happen within the next couple generations of the iPhone product. Apple has assembled an incredibly talented team of CPU designers that, with an ARM (ARMH) instruction set license in hand, managed to produce one of the best mobile phone processors available on the market. In terms of raw performance, only Intel's single-core Atom Z2460 is able to outperform Apple's dual-core A6 processor in a variety of benchmarks.
That being said, the main reason that Apple is not likely to move to an Intel chip - even as Intel continues to widen its process technology and micro-architectural lead over the rest of the ARM players - is that Apple has already spent gobs of money staffing up a team to build its "A" series of systems-on-chip. Is Apple supposed to simply lay off this team that it is still in the process of building, even if Intel's products have an advantage? Nope. For a device such as the iPhone, custom tailoring its system-on-chip to meet the device's exact needs is a useful asset.
So, how can Intel get into the iPhone? Well, two ways: first, Intel could simply act as a foundry for Apple. I do not believe that this is likely barring some sort of grander strategic relationship, since Intel has explicitly stated that it would only consider a foundry deal if it were not enabling its competitors. Building Apple's SoCs for them would be enabling a competitor to its own "Atom" product.
It's All About The Modem, Baby
The next way would be to reclaim the baseband/modem spot that it formerly held before the honor was stolen away by Qualcomm (QCOM). As I have noted previously, Intel is not behind Qualcomm (or any other SoC vendor) in the applications processor side of things - the Atom Z2460 is quite competitive with any of the ARM-based products. It's the modem/baseband that gives Qualcomm such a staggering lead over its competitors in the mobile space. While Qualcomm has been shipping LTE modems for a while now (first mover advantage), Intel will begin shipping its own LTE products in early 2013 to coincide with the dual core Atom Z2580 launch.
So, if Intel were to want to get back into the iPhone, it would need to do so by securing the connectivity portion of the device. According to iFixit, the modem/baseband from Qualcomm is actually the most expensive part of the BOM at $34 a pop. Now if we assume 73 million iPhone sales annually, this could potentially represent a $2.5B business for Intel. A relative drop-in-the-bucket compared to Intel's annual revenues of ~$54B, but any upside here counts.
Unfortunately for Intel, and fortunately for Qualcomm, Intel's modem/baseband technology is not quite up to par technologically as of today. However, as part of my due diligence, I ran across this at the Intel Job Opening page:
"MCG is expanding its research and development activities in the field of mobile communication platforms to fulfill our strategic target to become the leader in mobile platforms. We are planning to set up a complete modem team focused on HSDPA and LTE in the San Diego market. Sound amazing? Join us!"
Besides lamenting that I would likely not be qualified for the jobs there, I realized that Intel is dead-serious about becoming a significant competitor to Qualcomm in the baseband/modem space. Should Intel truly succeed here, then it has a shot at winning back the spot in the iPhone. Further, with a leading modem solution, it would be able to sell both an apps processor in addition to a world-class modem to phone vendors. This will have a significantly positive effect on revenues and is a great way to leverage the firm's enormous fabrication capacity.
Do I think that Intel will win the apps processor spot in the iPhone? No, not anytime soon. Apple's got a good team, even if it is not quite at the cutting edge of process technology as Intel's products will be. Apple's stuff is more than about raw CPU performance.
But could Intel win the modem/baseband spot? Yeah. It could. Qualcomm is a formidable opponent, and such a "win" would not happen overnight, but Intel has a habit of absolutely dominating anything it unleashes its engineers onto, even if it takes a couple of generations to get it right.
Intel isn't serious about mobile? Hardly. They're playing to win.
Additional disclosure: I am short ARMH.