Even as an ardent Intel (INTC 1.77%) bull, it is not difficult to understand why some investors may be running out of patience with the company's mobile strategy. Each year, Intel's product offerings become faster and more feature-rich. But it also seems that each year, the company closes the gap but doesn't quite get there. In mid-2015, the company rolls out its next-generation system-on-chip called Broxton, and this chip will be the one that finally changes the game for Intel.
Why has Intel failed so far?
The ironic thing about mobile system-on-chip products is that despite their low cost, these little chips integrate a ton of IP. It's not good enough to just get one or two blocks right -- a successful system-on-chip vendor integrates all of the right IP and gets that solution to market as quickly as possible. Intel's problem is that when it first began its system-on-chip efforts back in 2008, it was missing a lot of critical IP. Five years and many acquisitions later, Intel now has the IP mostly right, and the company's recently launched Bay Trail system-on-chip is proof of that.
However, having the right IP in place isn't enough. While the company's Bay Trail chip is an excellent solution for tablets from a performance and power perspective, it's not as highly integrated as chips from its competitors and, unfortunately, the platform itself sports a rather rich bill of materials. While Intel is helping its OEM partners out with subsidies, which should help it reach its aggressive goals of "more than 40 million" tablets for next year, this doesn't solve the smartphone problem.
Smartphones are just plain tough
At Intel's analyst day, the company partitioned the smartphone market into five different buckets, as illustrated here.
The vast majority of the smartphone market -- more than 50% -- falls into the "ultra-low cost," "entry," and "value" buckets, which means rather inexpensive phones. The rest of the market, performance and mainstream, is still a sizable chunk -- and the silicon ASPs in higher-end devices is naturally higher. But this market is also not growing as quickly.
The performance and mainstream markets are the markets in which Intel can -- and should -- differentiate based on performance. But the irony here is that the company's chips have repeatedly missed the mark time and again compared to competitive chipsets from Qualcomm (QCOM -0.20%). They're getting better, but they're still not there yet.
In the mainstream segment, the market is absolutely brutal. Margins are cutthroat and there is a distinct focus on entire platforms and highly integrated solutions. Intel is well behind competitors like Qualcomm, Broadcom, MediaTek, and Marvell here. But the company hopes that its newly announced SoFIA parts for this space will finally help it be competitive -- although the first iteration will include a 3G modem, making it uncompetitive with the LTE-capable parts coming to market soon.
Broxton: It's the game changer
At the risk of sounding naive, it's clear that the company's upcoming Broxton chip will finally be the part that gets Intel some real design wins in the high end of the market. The company pointed out at its analyst day meeting that this part would drive leadership performance and would be suitable for the highest-end smartphones. One Intel engineer that I spoke with flatly indicated that if Broxton didn't dominate the competition that he would retire. Bold words, indeed!
While the competition certainly isn't asleep -- again, Qualcomm is very good at its job -- Intel will have a fundamental density and transistor performance advantage with its 14-nanometer processors. Intel will also need to nail the design. But if the company delivers what it has promised at its analyst day, Intel's competitors will run into economic and physics-related barriers that should allow the company to have a meaningful lead in performance per watt and integration of compelling IP at the high end.
Foolish bottom line
If Intel can deliver on Broxton, then it will have proven its abilities to really play in the high end of the market. The lower end of the market will still probably remain elusive until its highly integrated part gets moved to the company's 14-nanometer process, which will bring cost, performance, and power advantages. But this isn't as important as gaining traction in the higher-margin, high end of the market.
Investors with a long time horizon could be rewarded handsomely if Intel finally comes through, and with a roughly 3.7% dividend yield and a modest valuation -- roughly 13 times forward earnings -- the risk here really seems to be opportunity cost rather than permanent capital impairment.
