As I mentioned in my last post, Intel and Microsoft are two companies I am currently looking at it because they appear to be selling at very low prices. This first post in a series will be on Intel, I will get into their competitors and Microsoft as I go through my analysis.
Company Overview
Intel is a semiconductor chip maker, primarily for PC’s and servers. They design and manufacturer chips used all over the world with the majority of sales now coming from the Emerging Markets (57% in 2010). They are one of the last remaining vertically integrated players in the semiconductor industry along with Texas Instruments. This means, they do everything from design their chips, to manufacture them, and to create low-level software to help software makers design code for their chips. Many newer players in the semiconductor industry do only one of these functions, like ARM Holdings who designs chip cores and licenses them to many manufacturers.
While Intel is the dominant player in the PC and Server industries (which represent >90% of their revenues), they are struggling to get into the faster growing mobile device space (think smartphones and tablets). For this reason, the stock price has remained depressed for some time (as of Oct 18, shares were selling at a P/E of ~9 with a dividend yield of 3.6%). Intel has tried to make inroads into the mobile device space with it’s Atom processor but with little success so far. Their latest generation processor and manufacturing process (based on the 22nm, Tri-Gate design) aims to change this.
Intel is the industry leader in technology and manufacturing. They produce the fastest chips (hence their dominance in high performance PC’s and servers) based on a CISC/x86 architecture and have been ahead of the pack in manufacturing for some time now (just recently releasing chips on the 22nm node vs. 32nm for competitors like AMD and other ARM licensees).
Intel is also going through a transformation from a chip maker to broad-based computing company to position themselves for the future where connected devices will likely come in all shapes, sizes, and capabilities. One step towards this transformation is their acquisition of McAfee (the virus scan makers) in order to begin to incorporate computer security into their hardware designs. This would be another unique competitive advantage, in addition to design/process technology, over their competitors.
Industry Dynamics
The semiconductor industry is extremely competitive and is more so today than ever before. Intel’s primary direct competitor in the x86 PC/server market is AMD. But in the broader computing market, they have to compete against all chip makers who license ARM designs such as Qualcomm, Broadcomm, Nvidia, and Mavell Technologies. All of these players license core designs from ARM and then create chips for use in the mobile phone and tablet markets. These are the markets that Intel is struggling to break into as ARM owns a 95% market share.
Additionally, as Intel is desperately trying to encroach in the mobile devices space (as they need to given where the future looks to be headed), ARM is moving upwards into the netbook/laptop and lightweight server space with their latest Cortex A-15 design (as part of Nvidia’s Project Denver). Intel’s Ivy Bridge chips based on their 22nm Tri-gate process are their greatest weapon yet in the fight to gain market share in the mobile device market. While these chips are hitting the market now, it will be at least another year until ARM licensees hit the market with Cortex A-15 (which will be manufactured on a 28/32nm process) based laptops and web servers. From a technology perspective, Intel appears to have a leg up on the competition.
But there is so much more than sheer performance that goes into success in the computing market. Chip makers like Intel need partnerships with O/S makers and software designers to produce complete devices to build. For example the Wintel (Microsoft + Intel) partnership has been hugely successful for both companies over the past 2+ decades. In mobile phones, the dominant O/S is Google’s Android and Intel has recently partnered with them. While Microsoft has announced its latest version of Windows (Windows 8) will be released for both x86 (Intel) and ARM chips. The big question here is whether ARM chips will be powerful enough to allow Windows to work properly. Possibly not until the Cortex A-15 based chips are released next year.
Beyond just the heated race between Intel and ARM, there are technology considerations that may determine a winner.
Technology Wars
Intel’s (and AMD) chips are all based on the same x86 architecture that debuted with their 8086 processor in 1978. Every subsequent generation since then has been based on this architecture. And it’s important to note that these are Complex Instruction Set Computing (CISC) processors. On the other hand, a new instruction set was developed later on called RISC (Reduced Instruction Set Computing) and this is what ARM-style processors use. RISC vs. CISC has been an on-going debate for some time. Each has its best uses. Over 10 years ago, RISC was dominant in the server space but in the end CISC (Intel) won out. The claim was that CISC was more suited to complex computing. But the reality was that low-cost, CISC-based PC’s simply beat out high-cost, RISC-based workstations.
While I am no expert on RISC vs. CISC, the main takeaway is that RISC allows more compact processor designs and therefore lower power consumption. This is why it makes sense that RISC/ARM processors are dominating the mobile and embedded (auto, gaming, etc…) markets. CISC on the other hand appears to be the weaker approach from a performance perspective but because hardware and software technology now permit, it can compete with RISC in terms of speed. And as another barrier to entry for RISC chips, there is a whole ecosystem of software and hardware suppliers built up around Intel and the x86 architecture.
The other important factor is power consumption. This is where ARM chips are far and away the leader right now. ARM has been designing low-power chips over their entire ~25yr history while Intel is not as experienced. In tests of ARM vs. Intel’s Atom processor in netbooks, Intel wins on performance by a slim margin but ARM wins on power consumption by a large margin. Intel is hoping that their latest 22nm/Tri-Gate technology puts them back in the race. It looks like it will. Here, it seems like two factors are important, the first is process technology (how small is the transistor) and second is how complex (how many transistors on a chip). Intel is the process leader (about one node ahead of the rest of the world) but ARM with its RISC instruction set will be leader in low-complexity.
Business Model
While I’m describing this as an Intel vs. ARM war (which is how the whole world is describing it right now), there are many companies that fall under the umbrella of ARM’s business model and only a few remaining that work like Intel. Intel is the traditional, vertically integrated semiconductor manufacturer. They design their chips, design the fabrication process, operate foundries, manufacturer chips for computer makers, and create low-level software to run on these chips. ARM on the other hand does only one thing, and that is design chips. From there, they license the chip designs out to companies like Qualcomm and Broadcomm, who then supplement the core design and then either manufacturer the chips themselves or send the final designs to a foundry for manufacturing. Most of the semiconductor industry is now organized along the lines of ARM Holdings’ model – where each company serves a single purpose rather than being vertically integrated.
My view is the ARM model is superior. The difference reminds me of the IBM PC vs. Apple MacIntosh wars from the 1990′s. The Apple Mac was entirely proprietary – the hardware and the software. The IBM PC was an open architecture. Anyone could try and create “PC-compatible” machines, which they did. This flooded the market with low priced machines while Apple cornered themselves into fighting the world on their own. While Intel is dominant in the high-end computing market, they will need to fight against the world of ARM licensees to break into the mobile market. By spreading the cost of design across many customers, ARM should be able to keep their design costs lowest while the market will, similar to the PC, be flooded with an ecosystem of ARM-compatible products as it is today. This is a steep uphill battle for Intel. To break in, Intel will have to convince device makers that their technology can compete for the device makers to take the risk. ARM will have a similar advantage in trying to break into the PC market as long as they can get plenty of licensees to jump on board.
The disadvantage for the ARM model is that many manufacturers have to spend large amounts of capital every time they upgrade their manufacturing process (some estimate the next node will cost $15billion across the industry) and this will need to be spread across all the chips produced. Intel on the other hand only has to upgrade themselves.
In addition, my takeaway from reading articles about and press releases by both Intel and ARM is that Intel continues to focus more on the technical details while ARM is properly focused on the customer (what features do consumers want, so we can design them). Again, advantage ARM, especially over the long-term.
But Intel is the 800lb gorilla with the resources to boot and are still the world leaders in terms of manufacturing process. They have a huge footprint in the PC and server markets that will be difficult to upend.
So What
Technology and business models aside, what do consumers care about? They care about getting a device that works, works fast, and has long battery life. They don’t care about ARM vs. Intel. They don’t care about Droid vs. iPhone, they care about getting something easy to use that works. Based on this, I think the Intel-Google partnership is a good one. Consumers are already familiar with the Android interface and Android market has 100,000′s of apps. So it will be up to Intel to create a chip that is faster, with better graphics, and lower power consumption than ARM. Ivy Bridge aims to be that competitor. Whether all this competition is good for shareholders or not, we will see. It will no doubt be great for consumers. For now, in the race between ARM and Intel, I think it’s a toss up.
Next: I’ll look at some of Intel’s competitors in a little more depth before we dig into the financials.
