It's starting to look like 2014 really is the year that Intel (INTC) goes from bit-player in the tablet space to a certifiable juggernaut. If the company can hit its stated goal of north of 40 million tablet chip units shipped, then it should exit the year with about 15%-20% of the market. With tablets taken care of, investor attention is likely to turn to smartphones during 2015.
Qualcomm's domain is not so easily breached
The tablet market is very much like the PC market; these devices are usually Wi-Fi only and are typically sold unsubsidized through retail channels. Think of this market as similar to the laptop market, except the price and performance bars are shifted lower (the sweet spot for a notebook is probably $500-$700; it's $125-$250 for a tablet).
Qualcomm (QCOM) currently dominates the smartphone silicon market, a business that is surprisingly pretty different from the tablet/PC chip market. Cellular capability, performance, and power efficiency are the defining features, with application processor performance secondary in all but the very high end. This is where Qualcomm excels, and its leadership in this market will not be so easily breached (although many competitors -- current and former -- have tried).
How about Intel?
Today, Intel's share of the smartphone apps processor market is within a rounding error of zero, hardly improving over the last several years. Additionally, the company's stand-alone cellular modem business has been on the decline as the portfolio had, until recently, consisted solely of 2G/3G solutions. On top of that, the integration of the cellular baseband into the applications processor by competitors further eroded the market for stand-alone modems. However, not all is lost.
The company is shipping its first-generation stand-alone LTE solution, the XMM 7160, in devices such as the Galaxy Note 3 Neo and the Galaxy K Zoom today. This modem isn't top of the line, nor is it even particularly competitive with any recent Qualcomm modem, but it is a good, cheap solution for less demanding markets and applications.
Knowing it can't really make a mobile living from selling bottom-of-the-barrel LTE modems, Intel has been very aggressive in talking up its XMM 7260 LTE-Advanced modem. This part looks attractive, offering category 6 LTE-Advanced support, support for both TDD/FDD LTE modes as well as TD-SCDMA support. While Qualcomm's upcoming MDM9x35 should be a more fully featured solution, Intel is likely to position this as category 6 LTE-Advanced goodness at a lower price than Qualcomm's flagship.
Will anybody buy it?
Samsung (SSNLF) is using the XMM 7160 in the Galaxy Tab 3, Galaxy Note 3 Neo, as well as the Galaxy K Zoom. Further, Intel stated at Mobile World Congress 2014 that multiple design wins based on XMM 7260 would be announced this spring. The design wins have yet to materialize, but Intel still has 20 days to make good on this promise.
Looking to 2015
During 2015, Intel will have a low-end, low-cost integrated applications processor and modem family out known as SoFIA. The LTE variant should integrate a feature-reduced variant of the XMM 7260 and integrate four Silvermont CPU cores. This means Intel could have a very good shot at winning low-end business from companies such as Samsung (whose internal apps processor efforts are higher-end focused) and the many white box vendors in Asia.
Where things get iffier is at the high end. This year's Merrifield and Moorefield have yet to appear in handset design wins, and the high-end follow-on known as Broxton will not be ready until mid-2015. This means the latter needs to fight the Snapdragon 808/810, and while Intel should offer very competitive (my expectation is clear leadership) apps processor performance relative to these parts, the modem story is not as clear.
Leaked roadmaps suggest that the follow-on to XMM 7260, known as the XMM 7360, will still only support category 6 LTE-Advanced. This puts it behind Qualcomm's category 7 capability soon to be found in the company's Snapdragon 808/810 processors. On top of that, the XMM 7360 will still be a 28-nanometer part while Qualcomm's will continue to enjoy a performance/power benefit in moving to 20-nanometer. That said, Intel's apps processor will enjoy 14-nanometer, second-generation FinFET technology, potentially tipping the overall platform power/performance story in Intel's favor.
In 2015, Intel could drive pretty significant smartphone unit volume at the low-end/value portion of the market, and could even do so profitably. However, the high end of the market -- dominated by Qualcomm -- still looks to be quite challenging, particularly if Qualcomm is able to keep a meaningful lead on the modem/integration side of things. Intel will persevere and throw as much money and time into this market as it needs to, but it is important to temper expectations on the timing of such a smartphone incursion.