Intel (INTC), the world's largest chipmaker, is attempting a serious assault on the ultra-mobile computing space with its "Atom" lineup. It is important for investors to understand exactly what to watch for as the strategy unfolds. While it's nice to see the company tout a design win or two, it is much more important for investors to be cognizant of the very real headwinds that the company will face in attempting its mobile push, so that the investment can be tracked more knowledgeably.
Intel Needs A Better Microarchitecture
Intel is the newcomer in the mobile space and it is up against very strong competition from the top ARM Holdings (ARMH) licensees including Qualcomm (QCOM), Nvidia (NVDA), Texas Instruments (TXN), and Marvell (MRVL). It will not be enough to simply offer a "similarly performing" product to the current solutions, nor would it bode well for the company's gross margins in the sector for it to need to significantly undercut existing offerings.
Why? Well, the OEMs need a legitimate reason to pick this alternative chip that runs an entirely different instruction set. In fact, even as a long-time Intel bull, I would say that while Medfield is a great "proof of concept" design, it is simply not enough. On the CPU side, a single core Medfield (a 2008 core design) is slightly faster than most dual core Cortex A9 (a core design released in 2007) implementations.
However, Qualcomm's latest dual core Snapdragon S4 (which is said to be similar to ARM Holdings' own "Cortex A15" design) handily beats Medfield in single and multi-threaded tasks. Why should an OEM choose an Intel chip when faster performing ARM-based alternatives exist?
Luckily, Intel's "tick-tock" strategy (that fellow author Skyler Greene wrote an excellent piece on) should be in full effect. This is a strategy in which Intel will, in each year, alternately release a new microarchitecture or a new process technology node. ARM Holdings is on a fairly long release cycle for its microarchitectures (Cortex A9 was released in 2007, and Cortex A15 in 2010) and is dependent on the major foundries such as Taiwan Semiconductor (TSM) for process node development. Intel has a significant advantage in semiconductor manufacturing over the rest of the industry, and coupled with a leading micro-architecture, Intel could very well take meaningful share.
The caution here is that Intel does not execute as well as I expect that they will, or that the aforementioned ARM-based competitors provide even more competitive parts on a performance/watt basis at a more accelerated rate (Marvell and Qualcomm simply license the ARM instruction set and from there design their own CPUs).
Total Addressable Market Is Constrained
By unit shipments, Samsung held 32.6% of market share, Apple (AAPL) held 16.9%, Nokia (NOK) took 6.6% share, HTC Corporation took 5.7%, and ZTE Corporation held 5.2%. All other vendors comprised 32.9%.
While Intel has inked deals with U.K. based Orange, India's Xolo, China's Lenovo, and with Google's (GOOG) Motorola Mobility, it's clear that the lion's share of the smartphone market is locked up by Apple and Samsung. Unfortunately for Intel, Apple designs its own mobile SoCs based on cores licensed from ARM Holdings, as does Samsung, so it is incredibly unlikely that Intel will end up in the next iPhone or any flagship Samsung phone. In addition, while Intel's fabs are the most advanced available, Samsung's are sufficiently sophisticated that I doubt Apple would want to risk using Intel as a fab when Samsung is able to serve Apple's needs.
Another issue is that Intel will only be available in Android phones. While I expect Android to be the dominant smartphone operating system, I also do not want to discount Microsoft's (MSFT) Windows Phone platform quite yet, especially as I expect the firm's strong history of developer-friendly tools and the surprisingly strong reception to Nokia's Lumia could lead to real market share gains for Microsoft at the expense of iOS and Android. Unfortunately, again, for Intel, Microsoft's Windows Phone will not run on Intel hardware, nor does it seem that such a partnership is likely in the near-to-medium term.
TAM is a very real concern, and it will be interesting to see just how much of the Android market Intel's chips can steal.
So, That Was Pessimistic! Is There Hope?
The short answer to this is "yes." I believe that Intel has the chip design muscle along with the fabrication prowess to develop very compelling solutions. The company will, however, need to utilize both its hardware and software prowess to work with smaller phone vendors to develop novel phones with interesting hardware and software features. I would suspect that such a push would allow phones with "Intel Inside" to become more recognizable and desirable by the public. Interestingly enough, Intel's brand strength built in the PC space could translate nicely to the phone space if done right.
It won't be enough for Intel to produce "good enough" solutions. It can't rely on "good enough" CPU performance/watt, nor can it continue to skid by with "good enough" graphics performance in its mobile systems on chip. The company needs to hold a distinct technological lead if its products in this space are to gain meaningful share, as OEMs will need a very compelling reason to shift to the Intel architecture.
I believe that Intel will prevail, but I certainly do not recommend buying the company solely on speculation that they will make significant inroads here; Intel is a strong company in all of its core businesses, and a strong mobile presence would simply be yet another successful division of an already highly successful company.