Intel (NASDAQ:INTC) thrived in the transition from mainframe and minicomputers to personal computers. Several other companies once made the CPUs at the heart of PCs (remember Zilog?), but only AMD (NASDAQ:AMD) has survived, barely, to continue to make chips using the traditional x86 instruction set. Innumerable technology companies failed during that technology transition.
For years now some technologists have argued that the emergence of a "better" architecture, ARM, would spell the doom of Intel's "energy hogging" x86 chips. So far they have been wrong. PCs, whether servers or desktops or notebooks, are still dominated by x86 chips, almost all made by Intel.
Mobile: Intel still at competitive disadvantage
At the same time Intel has failed to make any significant headway in the mobile market. By mobile I mean smartphones and tablets, not notebook computers. Intel bulls point out that Intel does sell into these markets, has certain competitive advantages, and intends to be a major player in them. In addition Intel makes most of the x86 chips used in datacenters (the "cloud"), without which the mobile devices would not be able to run their rich app ecosystems.
You can check the current price of Intel stock from your smartphone, but likely you will be using some other company's silicon to do that. The main argument that Intel could gain a significant and profitable share of the smartphone market is based on its lead in process technology, the ability to put larger numbers of transistors in the same area of silicon. In theory since Intel typically has a one to two cycle lead in process technology over competitors, it should be able to produce a smaller, more-energy efficient, yet computationally more powerful chip to run smartphones.
But Intel has had that process technology lead going back to before Apple (NASDAQ:AAPL) introduced the iPhone in 2007. The iPhone itself was an incremental step evolving from Asian feature phones and BlackBerry phones that first appeared in 2003. There are good reasons to believe history will simply repeat itself. Intel sold its Xscale phone chip division to Marvell Technology (NASDAQ:MRVL) in 2006. While Marvell struggled with its own smartphone chip sales after Apple and Android-based crushed BlackBerry (BBRY), it went on to become a major player in the Chinese smartphone market and has been expanding into other markets, with Q3 mobile and wireless semiconductor chip revenue of about $251 million.
In other words, Intel dumped its ARM-based phone division to concentrate on x86 based chips. No big deal if they were not repeating the same mistake eight years later, but they are repeating the same mistake. The x86 architecture is great, but it co-evolved with PCs mostly running Windows. Modern ARM chips co-evolved with smartphones. Smartphones are not just little tiny PCs with a cellular modem attached. The domination of the smartphone chip market by Qualcomm (NASDAQ:QCOM), which began making chips for old-fashioned cell phones back in 1992, is no accident.
Intel can hire engineers and buy intellectual property to make smartphone chips, and in theory it could build those chips on a smaller process node than fabless companies like Marvell and Qualcomm, but it can't make the x86 architecture efficient at running software stacks that co-evolved with ARM.
In tablets the playing field is more level, but the advantages Intel has (other than the ability to waste large sums of stockholders' cash before they revolt) are minimal. The hardware costs of adding cell phone connectivity to tablets is minimal. Most people use Wi-Fi rather than cell service to connect their tablets mainly because of the high recurring costs of cellular service contracts, not because of the incremental cost of silicon. Some tablets run Windows, where x86 architecture has some advantages, but still has disadvantages in power consumption. But the hard, perhaps insurmountable reality is that the industry has standardized around ARM and Android.
This year Intel is paying OEMs to use Intel chips in tablets. The resulting Q3 operating loss in its Mobile and Communications group was $1 billion. Despite that Intel generated $5.7 billion in cash from overall operations in the quarter. Perhaps Intel can buy its way into this market. But why? Even without the subsidies, margins are likely to be near zero. Qualcomm is not going to give up market share without a fight, and also generates a lot of cash. As the market leader, Qualcomm can do to Intel (in mobile) what Intel used to do to AMD: cut margins temporarily to maintain market share.
PC and Server chip advantage is sturdy
Perhaps Intel would do better in the long run to concentrate where they have a clear current advantage, server chips. Intel also has an advantage in PC chips, with no competition from AMD at the high end. With no competition Intel has enjoyed fat margins for server chips and high-end PC CPUs. The PC market has stabilized because power users need local processing power - the cloud is no substitute. All this is reflected in Intel's Q3 Data Center group revenue, which was up 16% y/y, with units shipped up 6% and prices up 9%.
Cloud and other datacenters have been expanding mainly to serve the proliferation of mobile devices. How much should investors bet on cloud datacenters? Currently cloud growth is being driven by video, which requires higher bandwidth, vastly more storage, and considerably more computation power than non-video services. At some point, presumably, datacenter growth will peak and level off. But that is likely years in the future.
Meanwhile, those negative about Intel will tend to see a datacenter market share challenge coming from ARM-based server chips, or perhaps from AMD's announced but still to be commercialized chips combining ARM and x86 architectures. I am not sure ARM in the datacenter is ready for prime time, or will ever replace more than a small amount of x86 based servers. ARM could make some sense in situations where the cost of electricity is a major factor. But for the foreseeable future the ability to process large amounts of data quickly will be the trump, and that favors high-end 86x server chips from Intel.
Long term: don't count on rapid growth
The takeaway is that I don't see Intel's revenue growth either crashing or greatly accelerating in the next two to five years. That could change if there is an unforeseen breakthrough by Intel or one of its many rivals.
In the long run, I expect Chinese chip designers to undermine all but the very best of the American and European chip designers. The trend has already started, supported by a Chinese government that is just a lot more business friendly, and nationalistic, than the U.S. government.
Dividend and Yield
So I am treating Intel as a stodgy old company. All I really need to know about it is the dividend and how that compares to other safe dividend and interest-bearing investments. Here's how Intel's dividend yield compares with some other dividend paying technology stocks (based on recent stock prices):
|Dividend yields on December 3, 2014|
|Applied Materials (NASDAQ:AMAT)||1.68%|
Apparently Intel is hearing this from other investors. On November 20 Intel announced a dividend increase from the prior $0.90 per year. At the close on Monday, December 1, Intel's dividend was $0.96 annually, or $0.24 per quarter. At the closing price of $37.17 that is about 2.6%. Not too shabby compared to the 10-year U.S. bond at 2.34%. Note Intel could increase the dividend even more by shifting cash flow currently used for stock repurchases.
Disclosure: The author is long AMD, MRVL, MCHP, AMAT, XLNX.
The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article.
Additional disclosure: This article reflects the author's research and analysis for his personal investing; it is journalism, not financial advice.