By Andy Ng
Intel (INTC) has seen tremendous growth over the past couple of decades, thanks to the proliferation of the personal computer and the firm's dominant position in x86 microprocessors. As the world's largest semiconductor company, Intel is an undisputed tech titan, but not all has been smooth sailing recently. The emergence of smartphones and tablets has raised questions about the future of the PC market, but more important, it has given rise to a new competitive threat from ARM Holdings (ARMH), whose processor designs populate most of these mobile devices.
Despite some of the negative headlines surrounding Intel, we believe the firm's wide economic moat is intact. While the rise of smartphones and tablets may bring some challenges to Intel's bread-and-butter PC microprocessor business, it also presents the firm with new growth opportunities.
A Look at Intel's Moat
Intel is one of only a handful of wide-moat technology companies that we cover. The source of its moat can be attributed to the immense scale and resources that provide Intel with a significant research and development budget. Intel can invest heavily in semiconductor fabrication technologies to stay at the forefront of Moore's Law (in which Intel co-founder Gordon Moore said the number of transistors on a chip will double approximately every two years). By having the most advanced manufacturing processes in the industry, the firm is able to create faster processors at lower per-unit costs than its closest competitor, Advanced Micro Devices (AMD), as well as the rest of the semiconductor industry.
Together, these form the basis of Intel's tick-tock strategy, where the firm makes a new iteration down Moore's Law every two years (the tick) and launches with a new microprocessor architecture in the years between (the tock). This drive toward ever-advanced chips is clear evidence that Intel has the resources to push the envelope on new processors and technologies. Intel began manufacturing chips on the 32 nm (circuit size) process in late 2009 and started to ramp up volume on 22 nm production in the second half of 2011, which translates to roughly a one-year lead over other logic chip manufacturers in the semiconductor industry. For comparison, Taiwan Semiconductor Manufacturing (TSM) began to produce 28 nm devices in late 2011 and is ramping up volume production this year, while GlobalFoundries' most cutting-edge process is at 32 nm. Both foundries are suppliers to AMD, which currently has chips at the 40 and 32 nm nodes and will have products at the 32 and 28 nm nodes in 2012.
Intel's newly launched processors, code-named Ivy Bridge, are manufactured at the 22 nm node and are the successors to the prior-generation 32 nm Sandy Bridge processors. The Ivy Bridge chips incorporate 3D Tri-Gate transistors, which has generated some buzz and illustrates Intel's superiority in semiconductor fabrication technologies. Other chipmakers remain at least a couple of years away from being able to commercialize the technology. Tri-Gate transistors, which have been in the works at Intel for roughly a decade, offer performance and power-efficiency advantages over traditional 2D planar transistors. The 3D transistor technology will benefit Ivy Bridge, but we believe the bigger implications will be for Intel's foray into smartphone and tablet processors.
Intel Has Benefited From x86 Ecosystem, but Faces Threats From ARM
While we believe scale and technology advantages are key to Intel's wide moat and success in x86 microprocessors, the firm has also benefited from the x86 ecosystem itself. The x86 processor architecture is nearly ubiquitous in PCs and has become the dominant architecture of servers and workstations over time. Over time, most proprietary computer software has been written specifically for the x86 architecture, which has created substantial switching costs, thereby fortifying the x86 ecosystem. This provided a virtuous cycle for Intel, as the immense size of the PC processor market allowed the firm to invest substantially more in R&D to further the competitiveness of its x86 chips over the other architectures.
The ARM architecture, on the other hand, has had a strong presence in embedded and mobile electronics (such as PDAs, MP3 players, cameras, and cell phones), as those devices' lower performance requirements combined with ARM's low cost and power efficiency made the architecture ideal for such applications. ARM's long-dominant presence in mobile phones has extended to the burgeoning smartphone segment. The firm's customers in smartphone chips, which include Qualcomm (QCOM), Nvidia (NVDA), Apple (AAPL), and Samsung (SSNLF.PK), license processor designs from ARM to create their own customized mobile processors, which are then typically manufactured at foundries. In Samsung's case, the firm uses its fabs for both in-house production of its own chips and its foundry business, which is a key outsource semiconductor manufacturer for Apple. As ARM-based processors become more powerful, ARM and its customers are aiming to move upstream toward the realm of the x86 and its stronghold in PCs.
We believe the rapid emergence of the tablet has provided a battleground for the ARM and x86 architectures, as that device sits between the smartphone and the PC on the computing device spectrum. The rise of the tablet is part of the larger move toward cloud computing, where computing tasks are offloaded onto clouds of servers and users access the cloud via client devices, including tablets, smartphones, and PCs. ARM has been much more successful in tablets than Intel so far, thanks to the advantage its processors have in power efficiency, and this has opened the door for ARM to break into the x86's traditional dominance in mainstream computing devices, a category that we believe has extended from PCs to now also encompass tablets. Although Apple's Mac PCs run on Intel's x86 microprocessors, its iPad tablets use a custom-tailored chip based on an ARM design. Google's (GOOG) Android software platform, the popular operating system for tablets as well as smartphones, has predominantly supported ARM-based processors; only recently did Google and Intel announce that upcoming versions of Android will also support Intel's x86 Atom processors, which are aimed at mobile gadgets.
Perhaps the biggest headline came in early 2011, when Microsoft (MSFT) decided to break the so-called Wintel alliance in an attempt to spur adoption of the Windows operating system in tablets. Historically, Windows operating systems only supported the x86 architecture, which worked out tremendously well for Microsoft and Intel in PCs. However, the dominance of ARM-based processors in tablets convinced Microsoft that the upcoming version of Windows 8 would need to support both the x86 and ARM architectures if the firm wanted to extend the footprint of its Windows franchise from PCs to tablets. We expect the trend to continue toward operating systems that support multiple processor architectures on client devices (PCs, tablets, and smartphones), as it enables OS providers to cast a wider net and not rely on the success of any specific microprocessor architecture. It also allows hardware makers to select the best microprocessor available to meet specific design criteria for a new device and not be tied down to any ecosystems.
X86 Ecosystem Should Hold Up in the Cloud and Offer Opportunities in Server CPUs
Despite such pressures, we believe the x86 ecosystem will remain dominant, though it will probably be weakened. While Microsoft's Windows 8 will run on x86 and ARM, it will have two distinct modes--one that runs like today's Windows 7 OS, where existing x86 applications can run as they do on Windows 7, and another that provides an immersive tablet experience with new applications developed specifically for this user interface; it will also work on PCs. The backward compatibility of Windows 8 will be an advantage to Intel, as it means that anyone who wants to run existing or legacy software will need a PC or tablet with an x86 processor because ARM won't be compatible with the legacy mode. We believe this will be an important consideration for many users, particularly those in the enterprise environment where the ability to support older software is critical.
More important, we think the x86 ecosystem will remain strong in servers and the cloud. The emergence of cloud computing essentially shifts the bulk of computing tasks from desktops, notebooks, and so on to the cloud, where servers do the bulk of the heavy lifting and a user interface, such as a tablet, is used to access the cloud. As a result, the performance of server microprocessors is critical. We expect the x86 architecture to continue to dominate in performance and maintain its lead in servers because of Intel's ability to drive microprocessor performance through its large R&D budget and ability to stay at the leading edge of chip fabrication technologies. Much of the software outside of mainframes and proprietary server systems has been written for x86, and it would be costly and extremely time-consuming to shift to another chip architecture. Even though processor architectures used in proprietary server systems still make up a significant portion of total server CPUs, there has been a secular migration over the past decade toward the open-standard server systems based on x86. Thanks to this trend, x86-based server revenue rose from 23% of the total server market in 1997 to 62% in 2010, according to IDC.
ARM has announced plans to eventually enter the server processor market, primarily via microservers, which are energy-efficient servers that contain a high density of processors per unit, thereby reducing the footprint of each server and lowering the total cost of ownership for customers compared with traditional servers. While this is an emerging and fast-growing portion of the server market, microservers are only appropriate for certain computing applications and workloads. More important, Intel has its own microserver solutions with its Atom processors, which are already deployed in microserver environments today, and we believe the x86 ecosystem and Intel's ability to drive processor performance will heavily favor Intel in the microserver segment against ARM.
We think ARM has received support on the client device side because it offered a desirable characteristic for tablets that Intel couldn't--power efficiency. As computing tasks are moved to the cloud, the bulk of value in terms of computer processing is growing on the server side, while the user interface becomes "dumber," which has allowed battery life to become a more important consideration than performance in tablets. Therefore, as long as Intel takes advantage of its moat to maintain its processor performance lead, which will become increasingly critical in the cloud, it will remain positioned in the sweet spot of the server processor segment. As a result, the x86 ecosystem should thrive in the cloud, and Intel stands to capture a significant portion of the value when it comes to total computer processing demand.
This theme also has implications for Intel's long-term growth opportunities. One of the perceived threats to Intel is the possibility that the PC market will be cannibalized by tablets or even smartphones, something that we do not expect to happen. Instead, we think PCs and tablets will coexist, each appropriate for different applications. While tablets make sense from a mobility perspective, there will always be computing tasks more suitable for a PC, such as gaming and photo and video editing, and not appropriate for the cloud. While the emergence of tablets has had an effect on PC unit sales, it has generally been in the instances of individuals choosing a tablet over a PC as a second or third computing device. Although the PC market has matured in developed countries in recent years, growth has been healthy in emerging economies. PC penetration by population in geographies like Asia and Latin America remains quite low, and we expect continued adoption in emerging countries to support healthy PC unit growth for the foreseeable future. Even in a worst-case scenario in which the cannibalization of PCs does occur, the shift of computing workloads to the cloud would mean that Intel will still be able to capture the bulk of the value in terms of computer processing power via sales of server processors. Therefore, Intel is still positioned for growth as long as global demand for computing power continues to rise, whether in the form of PCs or cloud computing.
Intel Can Use Its Moat to Fend off ARM's Threat in PCs
We think Intel can use its competitive advantage to limit the threat of ARM to its main PC processor business. Intel has evolved its x86 processor designs and used its lead in driving Moore's Law to come up with the best-performing PC microprocessors on the market. While the ARM architecture has been successful in tablets, thanks to its power efficiency, it is at a disadvantage when it comes to performance. And performance matters when it comes to PC microprocessors, which is why Intel dominates the x86 CPU market, and we expect this to be true regardless of architecture. As a result, we think ARM-based CPUs will be relegated to the low end of consumer-focused PCs, at least for the time being.
ARM needs to make significant improvements to the performance side of its architecture in order for ARM-based processors to be on par with x86 in terms of running a full-fledged computer, and this will be no easy task. Even if ARM does eventually catch up, Intel's wide moat may prove overwhelming. Given that Intel holds a manufacturing technology lead over the entire ARM-related manufacturing base, it should be able to leverage that advantage to drive production of faster microprocessors at a lower cost per unit than ARM and its partners.
Tri-Gate Could Narrow ARM's Smartphone and Tablet Lead
In mid-2011, Intel announced a new transistor technology, Tri-Gate, which will be used in the firm's 22 nm Ivy Bridge microprocessors. The new three-dimensional Tri-Gate transistors offer significant performance and power efficiency over traditional two-dimensional planar transistors. The firm said Tri-Gate transistors at the 22 nm process node will allow up to a 37% performance boost at low voltages or a 50% reduction in power consumption at constant performance compared with 32 nm 2D transistors.
Other chipmakers, including TSMC, GlobalFoundries, and Samsung (all of which are major manufacturers of ARM chips), remain several years away from commercializing 3D transistors. We believe that this is another testament to Intel's chip fabrication technology lead and should provide Intel with the means to make its Atom processors, which are targeted at smartphones and tablets, more competitive against ARM on the power-efficiency front.
While Atom's x86 architecture has been less power-efficient than ARM so far, Intel is essentially aiming to offset the architectural disadvantage with its major advantage over the rest of the chip industry, which is Intel's lead in driving Moore's Law. Intel also continues to work on architectural design improvements to Atom to make it more competitive for smartphones and tablets, and it will have a couple of new system-on-chip platforms lined up for 2012: Medfield for smartphones and Clover Trail for tablets. Both of these SoCs are based on the new Saltwell processor design, with Medfield scheduled to be launched soon, while Clover Trail (rumored to be a dual-core version of Medfield) should be available in the second half of the year. We believe Intel's two-prong approach, via continued design enhancements and taking advantage of its manufacturing technology prowess, can help make Atom more competitive and possibly narrow the gap with ARM in the next couple of years.
The announcement that Intel and Google will partner to optimize future versions of the Android operating system, the most popular smartphone software platform, for Atom processors is another positive for Intel. The partnership will put the firm on a more even footing with ARM from some performance enhancements that will be possible with the optimizations, a benefit that ARM presently possesses with Android. More important, it will result in better support for Atom from Google. Currently, new Android OS releases support the ARM architecture but aren't available for Intel's chips right away, as the software has to be ported to work with Intel's x86 architecture. Future versions of Android will be released for ARM and x86 at the same time, which should help Intel sway more mobile device manufacturers to use Atom in their smartphone and tablet designs, as there will no longer be a time-to-market disadvantage.
Medfield Has Opened Smartphone Door for Intel, but Uncertainty Remains
In recent months, Intel has been gathering some momentum in smartphones. It has announced a number of partners that will be launching Medfield-based smartphones in the coming months, providing initial evidence that Intel is finally gaining some traction in what had previously been an elusive smartphone opportunity, which we attribute to a couple of reasons.
First, the work with Google to optimize Android for Atom appears to be paying off, and all of the devices from Intel's announced Medfield partners will run on the latest Android 4.0, or Ice Cream Sandwich. Intel supposedly has hundreds of developers working to make Android run well on its chips, and such software support is important to attracting handset makers to Atom. Second, based on our current knowledge of Medfield, we think Intel finally has a smartphone SoC that should help it become at least somewhat competitive against ARM. Initial benchmarks show that the Medfield should do well versus ARM's current generation of Cortex A9 processors on the performance front. While we won't know Medfield's true power consumption until the chip is launched, it appears that it will be in the right ballpark versus ARM, as Intel's smartphone reference design for its original-equipment manufacturer partners claims that it will offer adequate battery life.
However, by the time Medfield-based devices roll out around midyear, the launch of new generation of ARM-based offerings will be around the corner. Chipmakers, including Qualcomm, Texas Instruments, and Samsung, will release new SoCs late in 2012 that are based on ARM's next-generation Cortex A15 processor architecture, which promise a significant jump in performance compared with the Cortex A9-based chips that are in popular smartphones today. As a result, Medfield's competitiveness in the marketplace may be short-lived. Nonetheless, the chip is proving that Intel has the capability to at least get in the door with mobile device makers.
We believe Intel's ultimate success in smartphones is still uncertain, but its wide moat is likely to come into play in 2013-14. Intel may be able to catch up with ARM on both power and performance in smartphone chips as it accelerates its semiconductor fabrication process technologies for Atom to 22 nm in 2013 and to 14 nm in 2014. Although the outcome remains to be seen, making use of its competitive advantage by manufacturing Atom with industry-leading process technologies may be the trump card that Intel needs to close the gap with ARM in the realm of mobile device processors.
Windows 8 Could Give Intel Big Break in Tablets
The ability of x86 processors to run legacy software applications on Windows 8 should provide an inherent advantage for Intel over ARM on the new operating system. Furthermore, details recently provided by Microsoft indicate that Intel should have an even greater advantage in corporate environments, as the ARM version of the OS won't support enterprise manageability features that exist for PCs and tablets running on x86 processors. This means that ARM-based Windows 8 tablets (as well as any potential ARM-based PCs that may hit the market) cannot be joined to corporate domains, which limits the ability of enterprise IT departments to distribute corporate applications or control the use and security of ARM-based devices. Therefore, although ARM-based Apple iPads dominate the tablet market today, corporations will have significant incentive to use x86-based Windows 8 devices, which may help jump-start Intel's foray into the tablet processors.
While future versions of Windows may have more enterprise manageability support for ARM-based devices, Microsoft has extended a helping hand to Intel with Windows 8 to break into tablets. Taking into account that Microsoft typically refreshes its Windows franchise every three years, Intel should have at least that much time to take advantage of the opportunity. The Clover Trail version of Atom, which is aimed specifically at tablets, should launch in conjunction with Windows 8 in late 2012. As in smartphones, the move by Intel to aggressively migrate Atom's manufacturing technologies down Moore's Law in 2013 and 2014 could substantially enhance the firm's competitiveness in tablet SoCs over this period.
Disclosure: Morningstar licenses its indexes to certain ETF and ETN providers, including BlackRock, Invesco, Merrill Lynch, Northern Trust, and Scottrade for use in exchange-traded funds and notes. These ETFs and ETNs are not sponsored, issued, or sold by Morningstar. Morningstar does not make any representation regarding the advisability of investing in ETFs or ETNs that are based on Morningstar indexes.