It's easy to listen to the media hype about how Intel (INTC) "missed" mobile. First of all, this is a silly notion, because "mobile" isn't a field that's going away (hint: it's growing), and there's certainly no rule written that says if you miss a product design cycle that you're shut out of the next one. Many commentators also like to claim that the mobile industry doesn't want to work with Intel because they all like ARM (ARMH) better. This also makes no sense because ARM doesn't actually sell chips to anyone: it's up to third party SoC vendors to take the ARM IP and build a usable product. People who claim that LG, HTC, or even Samsung (SSNLF.PK) won't take a superior chip from Intel because they "hate Intel" and "love ARM" are, in my view, wrong.
The Semiconductor Space Is A Supplier's Game
The semiconductor suppliers are not like operating system guys, or even device folks. They make chips that meet certain specs. In the mobile world, the main criteria for buying a chip are:
- Does it provide me with long battery life?
- Is it fast?
- Is it priced correctly?
The first criterion is obvious -- these mobile devices live and die by their battery lives. What good is a smartphone that only lasts 30 minutes on a charge? The second criterion is more subtle. There are a lot of claims that "performance doesn't matter". Oh, really? If that's the case, then why do folks like Qualcomm (QCOM) constantly beat their chests about how fast its Snapdragon is? Why do people continually upgrade to new phones if performance of their older phones is "good enough"?
Performance enables fundamentally new functionality, so saying that it doesn't matter in computing is incredibly naive and short sighted. Once again, don't rely on anybody who claims that "performance" is irrelevant for the next generation of mobile devices -- it's the key enabler of future usage models.
And finally, device vendors want their parts cheap so they can maximize margins on the phones. It's a cutthroat business out there, and in order to make money on these things, the parts need to be obtained inexpensively for the level of performance and value the part brings.
So, why do people count Intel out, again?
I will admit that I was very skeptical about 2 years ago when I heard that Intel would try to make a run for the ARM-dominated mobile space. Not because I didn't think that Intel could make a great chip, but because of software compatibility issues. For the same reason that Windows On ARM is destined to fail, I worried that Intel In Mobile would suffer a similar fate.
However, there are some key differences:
- Android applications (and Android is the dominant player) are mostly instruction set agnostic since they are generally not written in native languages such as C/C++
- Windows Phone 8 (which Intel has expressed interest in via this job posting) is similar in that the majority of applications are instruction set agnostic
- Most smartphone/tablet applications are very simple and non-mission critical in nature, so porting and validating any native applications (of which there are relatively few) isn't an insurmountable barrier
- Intel has a strong heritage of software developer collaboration and support from its position in the PC/server market
- Intel is willing to spend lots of money to make sure that compatibility going forward is a non-issue
Android runs on Intel's X86 instruction set architecture today, and it will be a first-class citizen for that platform going forward.
In fact, Intel and Motorola Mobility (owned by Google) seem to be getting quite cozy, with the two having announced a multi-year, multi-device partnership back at CES 2012. The first fruits of their labor was actually released back in September, and it was known as the Motorola Droid Razr i:
More importantly, another big growth area is tablets. At first, Intel's strategy was to aim for Windows 8 tablets and then on the phones go for Android. At CES 2013, Intel surprised the crowd by showing off its next-generation, 22nm Atom platform codenamed "Bay Trail" not only running on Windows 8, but running Android as well (the tablet in Mike Bell's hand is running Android):
So, with that hurdle out of the way, software compatibility is no longer the big issue that I and many others thought it would be. That being said, as I noted above, Intel is a components supplier, and as such, if it wants to have a dominant position in the market, it needs to have top-notch product.
Intel's R&D: It Costs To Be Boss
For tech companies, R&D is critical. I like big R&D spend, and I distrust any company that claims that it can win a space while being thrifty. A quick look at the trend in Intel's R&D gives one a very clear view into just when the company decided to get serious about mobile processors, RF/modem technology, and the whole software stack that comes along for the ride:
From 2010 to the most recent quarter, Intel has increased annual R&D spend from a hair over $5B to $10B/yr. R&D is projected to nudge up slightly during 2013 to $10.7B. While the efforts here include the PC chips, solid state drive effort, server chips, it is clear that around 2010, Intel went into warp-drive in a bid to significantly grow its offerings and markets. R&D efforts starting in 2010 wouldn't really show up until 2013/2014, but rest assured that this near doubling in R&D will start to pay off around that time-frame. Intel keeps touting the 2nd half of 2013 as when the "transformation" will be clear to everybody, and from the R&D trend, I believe them.
What's more interesting though, is how everyone says that Intel needs to compete with the entire ARM ecosystem. Take a look at this:
Qualcomm, the world leader in mobile technologies today, spent $4.14B on R&D in the last 12 months. TSMC (TSM), the world's leading semiconductor foundry, spent $1.25B, and ARM Holdings, the world's leading mobile core licensing house, spent a mere $256M. The three parts of the ARM ecosystem that compete directly with Intel, combined, will spend significantly less than Intel does.
Further, this heavy dose of R&D is sustained by the absolutely gigantic amount of gross profit that it takes in relative to peers:
Finally, despite arguments that Intel's big spending doesn't/won't amount to much, it is easy to get lost in the mass mobile hysteria and forget that Intel makes more in net income than the mobile champions...combined (that's after the hefty R&D, SG&A, and depreciation):
The question isn't whether Intel will get a big slice of the mobile pie, it's when. And that time is very, very soon.
Intel made its bets back in 2010 and has continued to relentlessly keep fighting to get the first fruits of this labor out to market. Investors are impatient and tend to focus on the near term (6 months - 1 year). That's why ARM Holdings trades at $20B market cap despite posting net income that wouldn't even fund Intel's quarterly dividend check. That's also why Qualcomm is able to have a market capitalization greater than Intel's despite Intel's clear cash generating superiority. It's also why the press has had a field day touting the death of Intel, the PC, and whatever else makes for a good story.
In the 2nd half of 2013, Intel's 22nm "Atom" products for both micro-servers as well as for tablets will hit the scene. At MWC 2014, Intel is likely to release its 22nm smartphone system-on-chip, coupled with a complete LTE solution.
Don't sell Intel before these products hit the shelves. Even though the stock's price action won't be sexy for several months, now is an excellent chance to accumulate in anticipation of the sudden sentiment reversal, in addition to the renewal of revenue growth, that is bound to occur in the 2nd half of 2013 leading into 2014.