Intel (INTC) reported second quarter 2013 earnings yesterday and provided guidance for the balance of 2013.
The $12.8 billion in revenue was well within the guidance range, and the $.39 GAAP earnings per share was spot on Street expectations. I'm always amazed at how a company the size of Intel can project this scale of revenue and earnings to an accuracy of less than 1%.
The disappointment, if you can call it that, was that the full year guidance was taken to flat vs. up "low single digits." What does that mean? Well, two quarters are now in the rearview mirror; first quarter at $12.6 billion and second quarter at $12.8 billion. So far that is $25.4 billion. Third quarter guidance is $13.5 billion, which it will make, so we are at $38.9 billion for nine months. Expecting $53.4 billion total for the year means that the fourth quarter will have to be about $14.5 billion.
The interesting thing about Intel's fourth quarters is that they are a very good predictor of the following year's performance. Fourth quarters don't have anything special in them; no build up to the holiday season, no "back-to-school" blip. Fourth quarters have about 10% fewer work days in them for Thanksgiving and Christmas. Far more times than not, the fourth quarter numbers times 4.something end up being the following year's revenue. I'm going to use a conservative 4.1 times the probable fourth quarter revenue of $14.6 billion. That would put 2014 Intel revenue crowding $60 billion.
I was checking Intel's sales history in Value Line (it's good for that); Intel's growth is nothing if not lumpy. The company came through the Great Recession in pretty good shape, only dropping 8.3% in revenue. 2009 saw the company come roaring back with a 34% one year growth, followed in 2010 with another 23.8% increase. Intel grew $18.9 billion in two years! I challenge anyone to find another hardware company that has done that kind of growth. That amounts to a current day Qualcomm (QCOM) or TSMC (TSM), from scratch, in two years. If compounded since 2009, Intel would have an 11% compound growth rate, even including the three flat years.
OK, enough of history; what is Intel going to do for me now?
Haswell isn't a bad start. A little side note: I have been looking for a Haswell based Ultrabook so I can finally put my Dell boat anchor on a railroad track somewhere. There are none. Of course it dawned on me that the phased introduction of Haswell had desktop parts first. Any low power mobile Haswells went to Apple (AAPL) (another favor for Apple?). With that insight, I find it amazing that Intel did as well as it did in the second quarter. The third and fourth quarters will see a wave of mobile Haswells jazz up the PC business.
Haswell will give us fanless, 3 pound notebooks with 12 hour battery life with the performance of supercomputers of a decade ago.
The new Atom core will come to life as Baytrail for tablets, Merrifield for smartphones, and Avonton for the low end server market. The Silvermont has a dynamic (as required) ability to be either one-fifth the power of the previous Atom, or 3 times the performance, and it can change on fly as required.
But get this: The Baytrail is the only tablet processor with the capability of running either Android or full Windows and switch between the two without re-booting!
Let me repeat that -- Baytrail can run either Android or Windows.
That little factoid, which the analysts seem to yawn over, renders all other tablet processors obsolete. Period.
Presumably the Merrifield will have the same capability of Android/Windows OS compatibility for smartphones next year.
We learned (I learned) for the first time that the Clovertrail+ design win on the Samsung Galaxy Tab 3 also included the Intel LTE chip. The real drag on the mobile business for Intel, in my opinion, has been the delay by Infineon in delivering a respectable 4G LTE chip. That problem appears to also be in the rearview mirror. Qualcomm still holds the high ground on the latest LTE chips, but that advantage will dissolve over the next year.
Examining the reported financials, I found a tidbit that I thought couldn't be true. Intel client (PC) units was up 5% in the second quarter over the first quarter and down only 6% year to date over year to date 2012 -- look for yourself. You have to scroll all the way down the last page. Where is this "death of the PC" that everyone is focused on?
Given the lumpiness of Intel's growth record and the capacity laid down during the past three years, I would expect another Intel growth surge in mobile to the distinct disadvantage of Qualcomm, Nvidia (NVDA), TSMC, and ARM (ARMH).
The mobile business, between tablets and smartphones is or soon will be about 1.3 billion units annually. Just a wild guess at the application processor and communication processor silicon dollar content on an average basis, considering high end to low end, would be about $25 per unit or $32+ billion in business that would be new to Intel.
What I have finally come to appreciate about Intel is that the company is not a racecar; it is a D10 Caterpillar capable of slowly crushing anything in its path.
The process technology is in place at Intel, the manufacturing capacity is in place, and now the Haswell and Silvermont products and the LTE chips are in place for a massive assault on the mobile business from top to bottom, beginning right now.