My readers know that I was probably beginning to start to sound a little nutty when Intel (INTC) was in the $19-$21 range, promising that the second half of 2013 would be a significant upside catalyst for the stock. After all, the folks on Wall Street want to sell you the drivel that Intel missed mobile, will see its PC chip sales evaporate, and will also lose tons of market share in the server space because a few startups giddy that they can play CPU designer think that they are going to make billions from micro-server chips. Well, I've got news for you -- they're all completely and utterly full of hot air.
Tablets Killing PCs? They're The Same Darn Thing
An argument that I've never understood is how the "tablet" will replace the PC. Fundamentally, a tablet is just a notebook but with the processing guts inside of the screen rather than under the keyboard. Oh, and they pack touch screens. You'll get some of the ARM (ARMH) bulls/Intel bears coming in and saying that the tablet is becoming functional enough to replace the PC for most tasks. Uh huh… so, why doesn't it work the other way? Why will a tablet running Google's (GOOG) Android be much more able to "scale up" to do big-boy tasks than a Microsoft (MSFT) Windows based tablet is to scale down to the tablet side of things?
My point is that fundamentally, even the Intel bears -- who claim that the PC will be supplanted by the tablet -- are simply acknowledging that there is some convergence happening. I certainly agree, and that's why I am happy that Intel not only has a virtual lock-down on the Windows 8 tablet/PC ecosystem (Windows RT is just cut down Windows 8, which is why nobody is buying it), but is quickly expanding onto the instruction set agnostic Android and Chrome platforms.
To put it simply: Intel's chips are right at home on Android, thanks to the fairly instruction set agnostic nature of the majority of Android applications and Intel's army of software engineers working around the clock to work with developers to clean up the remaining compatibility issues with natively compiled programs. Microsoft's Windows, on the other hand, features millions of compiled legacy applications that would require a determined juggernaut to try to get ported over. Microsoft isn't going to sponsor this (it did not even mention Windows RT on it conference call, but did pump up Intel's 22nm "Bay Trail" and "Haswell" processors), Qualcomm (QCOM) and Nvidia (NVDA) are too busy raking it in over in Android land to bother, and Samsung (OTC:SSNLF) has already publicly written off Windows RT.
So, yes, thinner, lighter, touch-screen devices with the processing guts behind the screen will "kill" the PC as we know it, but that doesn't mean that the Intel powered personal computer will stop growing. I think that with the public demonstrations at IDF 2013 of "Bay Trail," slated to power $200 touch enabled PCs, the Street finally "gets" it, which is part of the nice run up in the share price.
ARM Servers: That's Cute…
Intel's position in high performance computing with the Xeon E5 and its rapidly increasing market share in the big iron markets traditionally dominated by IBM's (IBM) Power and Oracle's (ORCL) SPARC thanks to its Xeon E7 will not be touched by the ARM vendors. These processors are highly complex, require large dies, herculean validation efforts and are built using processor cores that were designed to do much more than power a cell phone (and no, 20 of them packed into one die isn't going to cut it as the single threaded performance won't be there). A few startups licensing ARM's Pentium-M level 64 bit core isn't going to cut it, and the R&D required to do something bigger is just not within the scope of these companies.
But then you have the rapidly emerging "Web 2.0" servers that aren't out to do heavy computation, but instead need to service many computationally light tasks. Low power consumption is the key to the total cost of ownership equation (in which actual hardware costs are not the major driver), so having the right chip for the right workload is key to minimizing TCO. Intel will be entering the ring during Q3/Q4 this year with an army of processors for this space, based on both "Haswell" and the upcoming "Silvermont" full redesigned Atom core. Intel's 22nm FinFET transistors help give the company a performance, power, and die size advantage over the ARM startups, which will see their first 64 bit chip based on TSMC's (TSM) 40nm process ship at the end of 2013/early 2014, not to mention that Intel has been doing this for much longer and has more resources to throw on the actual chip design. While the ARM vendors will inevitably win designs, Intel will see significant growth in this space as I believe that market share well north of 50% is not at all unreasonable to expect.
Smartphones: Connectivity Concerns
I don't see this as, "oh no, Intel missed mobile forever!" (a silly notion indeed), but as a chance for Intel to materially grow its revenues. Intel already does ~$54B/year without phones and with still ramping tablets, and any phone wins are essentially "gravy." Right now, Intel's big problem is that Qualcomm is the king of LTE while it -- and every other phone chip vendor -- is still working to ship its first LTE devices. In tablets, this doesn't matter so much, but in phones, whose primary claim to fame is cellular connectivity, having trailing edge connectivity is a deal-breaker.
Intel acquired Infineon Wireless at the end of 2010 precisely so that it could become a real player in the communications space after it backed the wrong horse internally (WiMax, anybody?). Intel currently ships a credible 3G/HSPA+ solution into phones today, and will be shipping its first generation LTE modem, the XMM 7160, by the end of the first half of the year. Intel will likely win designs, but with Qualcomm pouring gobs of money into R&D to maintain its modem lead, and with Broadcom (BRCM) eyeing the #2 spot, it is going to take some real evidence of progress for me to become truly bullish on the high volume smartphone story. I believe Intel will have a superior applications processor, but this is worthless without the right cellular solutions to go with it.
See, I am very bullish on Intel's ability to win compute centric sectors such as tablets, PCs, and servers because Intel's advantages and experience as a computing oriented company are clear as day to anybody with knowledge of the industry. But Infineon Wireless wasn't exactly a Qualcomm, and I have tremendous respect for Broadcom, so while I am encouraged that Intel set up a site right next door to Qualcomm to try to poach its engineers, I really do need to see the same competitive strength in cellular processors as I do in compute processors before I can start adjusting my Intel models to reflect significant growth. XMM 7160 is late and appears to be paired with the upcoming 22nm "Merrifield" processor, when the May 2012 investor meeting presentations suggested that the next generation XMM 7260 would come with "Merrifield" and that the XMM 7160 would be shipping now with "Clover Trail+".
The final concern is that Intel's modems aren't built at Intel -- they're built at TSMC, which means that Intel can't really gain the power/performance advantages of its 22nm process in the modem. I expect that this is the last generation to be built at TSMC (since Infineon Wireless was probably designing this chip before the Intel buyout) and that going forward everything is built at Intel. In fact, Intel has no choice if it wants to build an integrated modem + apps processor part similar to what Qualcomm, Broadcom, and Nvidia all will have by the end of the year.
Margins bottomed in Q1, revenues have bottomed, and we're on the cusp of the assault of new products across the board. There is still plenty of upside left in this stock, and I believe that buying shares on weakness is going to be a winning strategy during the remainder of 2013, barring some terrible, unforeseen macro problems. Intel is now the largest semiconductor company by market capitalization, and now that the market is worried that smartphone chips will suddenly become a lot more like commodities, I believe that the former king -- Qualcomm -- stays in the (respectable) #2 spot going forward.