Intel Outside: Breaking Into Mobile

On November 19th, Intel CEO Paul Otellini announced that he would be stepping down at the company’s Annual General Meeting. The news was a shock to investors as Otellini, 62, was widely expected to stay on until mandatory retirement at age 65. Intel’s share price had bled 20% over the prior 3 months and the news of Otellini’s departure prompted another 3.5% drop on what was an otherwise flat day for the markets. In the face of weakening sales of PC processors, it appears that a titan of the semiconductor industry, once the world’s largest semiconductor company by market capitalization, has fallen victim to the rise of mobile devices – smartphones and tablets. Processor companies, such as ARM Holdings and Qualcomm, have grown and prospered off of staggering mobile growth. Otellini’s retirement represents an opportunity for Intel to recapture its lost glory.

Intel has struggled to gain market share in mobile technology, due to fundamentally misunderstanding the competitive dynamics that shape the mobile industry. A new CEO brings hope for a change in Intel’s mobile strategy, without which, Intel will struggle to replicate its past excellence and maintain its dominant market position.

Complicating the Value Chain

A simplified value chain in both the PC and mobile processors market is described with the following four steps:

(1) Processor Architecture – A list of commands for the processor to execute is developed;

(2a) Processor Design – A processor is designed to execute the instructions in the architecture;

(2b) System on Chip (SoC) – Additional functional components are integrated into the processor;

(3) Fabrication – Processors are manufactured at semiconductor fabrication plants, known as fabs. Foundries perform this task for ‘fabless’ companies, who outsource this activity;

(4) Device Manufacturing – Processors are integrated into electronic devices;

In the PC processors market, Intel performs steps 1 through 3, and has been able to secure a dominant 80% market share. The mobile processors market however is significantly more fragmented and complex. While Intel’s architecture is the standard for PCs (and Macs), ARM’s architecture commands 95% of the mobile processor space. Unlike Intel, ARM does not manufacture processors, instead choosing to license its architecture to companies further down the value chain. This means that other companies can compete based on their products, services and strategies at later stages of the value chain based upon ARM’s architecture – a complete opposite of the PC market.

ARM also licenses processor designs, known as IP cores, to its customers. While most companies license ARM’s IP cores and enter the value chain at step 2b, Qualcomm and Apple actually license the ARM architecture, designing their own processors at step 2a. Apple just recently started utilizing this strategy with the introduction of the iPhone 5’s A6 processor. In mobile, the SoC step usually involves the integration of Wi-Fi, GPS, and 3G components into the processor design. Qualcomm currently holds a 48% revenue share in the mobile market because of its strong IP holdings in these wireless communication technologies.

In Fabrication, most of the mobile processor companies are fabless and rely on foundries for manufacturing. The largest foundry is Taiwanese Semiconductor Manufacturing Company (TSMC) which has a 49% market share in manufacturing. The production process used by fabs and foundries is constantly changing as new processes are developed to fit more transistors, the building blocks of processors, onto a single chip. In addition to the products’ size benefits, smaller transistor size leads to lower power consumption, making the production process a main point of competition for foundries. Intel’s process has historically been two years ahead of competitors in the foundries space, and thus, their lead in manufacturing is a great advantage.

Intel’s Move into Mobile

The PC processors market is suffering from declines in PC sales and slowly eroding margins. Last year, PC shipments fell for the first time in over a decade, resulting in reduced demand for these processors. Alongside its poor Q3 2012 financials, Intel lowered its Q4 guidance and more actively managed expectations beyond then.

While PC sales languish, sales of mobile devices – and therefore mobile processors – continue to explode. Over 731 million mobile processors are expected to be sold in 2013, 22.3% more than last year. With this growth and related supply shortages comes an opportunity for a new player to successfully enter the mobile processors market. Few would seem better positioned than Intel.

Intel’s first major push into the mobile market is with its new Atom mobile processors, based on the same architecture as its PC processors. Intel is pushing these processors to be integrated into new Windows 8 tablets and convertibles (a laptop-tablet hybrid). However, this will not be a smooth entry for Intel as ARM is wedging into this tablet market as well, with Windows 8 adding partial support for ARM’s architecture.

Intel is also pushing its Atom processors for Android-based smartphones. The first smartphone powered by the Atom processor was launched in India in early 2012, but no phones have launched yet in North America. The widespread adoption of ARM’s architecture is preventing Intel from gaining traction in mobile. Android manufacturers are reluctant to switch, as large switching costs often accompany such changes and the Atom processor is not yet compatible with all Android apps. Intel only has a 0.2% market share of the mobile market.

Can Intel’s Architecture Succeed?

On average, ARM’s architecture is more power efficient than Intel’s. While Intel has generally had the upper hand in performance, processors have reached the minimum performance threshold for most users given constraints on battery life. Incremental performance generally increases battery drain which destroys value for end consumers. ARM also recently announced its new Cortex-A50 processor series with offerings that are either three times more powerful or five times more power efficient than the previous lineup, exceeding Intel’s in processor power efficiency. Add in that some Android apps are not fully compatible with Intel and it’s clear that ARM’s architecture will continue to dominate the smartphone market.

Although Intel’s processors can run older Windows applications on Windows 8 tablets, this may not be a sufficient advantage for Intel as the Windows 8 tablet experience is centered on new apps and its online focused features. Since the tablet market is already filled with ARM based iPads and Android tablets, and with the uncertain performance of the Windows 8 tablet, ARM’s architecture also seems to be well positioned in the tablet market. The Intel architecture is unlikely to make significant progress in the mobile market.

Intel’s Real Advantage

If Intel’s architecture in smartphones and tablets is at such a disadvantage, should it even bother to compete for architecture dominance? Architecture design is not a high value component of the value chain – ARM only makes about 11 cents per processor. Currently, Intel is trying to compete in this step of the mobile value chain, which only makes sense if it can drive volume to other segments of the value chain. Given their current market share and ARM’s competitive position, it seems unlikely this will change.

Intel’s management is seemingly blinded by their goal of trying to recreate their dominance of PC processors in mobile that they are ignoring existing opportunities. Intel’s real advantage in the semiconductor industry is its strength in production process innovation, resulting in smaller processors, improved power efficiency, and a greater capability for SoC integration – all important bases of competition in the mobile processor market. Intel has also announced that due to the slowing PC sales, the company will slow down production. The slowdown in sales has reduced margins because their facilities are now underutilized. Given the need to have the most advanced production processes, expensive investments in fabs must be made to remain competitive. If these assets are underutilized, processor companies are less able to afford the continuous advancement in their manufacturing facilities that forms the basis of their competitive advantage. All these factors indicate that a foundry-focused strategy where Intel manufactures, rather than designs, mobile processors would be a great fit for its strategy and capabilities.

Margins in the foundry space are fairly high compared to Intel’s traditional business. Foundry EBITDA margins are the highest across the value chain at 38% due to lower R&D and selling/administrative expenses compared to Intel’s traditional business at 24%, making a foundry approach more financially attractive.

By entering the mobile market with its current architecture, Intel is trying to force its control structure in the fragmented mobile processor market. This is difficult when fractions of segments of the mobile value chain are worth more than Intel and its step 1-3 business in the PC market (Qualcomm’s market cap recently surpassed Intel’s). Mobile is a different market landscape than Intel is used to. Intel has grown complacent due to its dominance in the PC processor market and needs to be more flexible in its offerings if it wants to compete in mobile, even in the manufacturing space.

Competing as a Foundry

With established incumbent TSMC controlling a large portion of the foundry market, Intel needs to be compelling to attract customers. Intel’s major advantage over TSMC is its excess production capacity. The current foundry space is starved for additional capacity with both Apple and Qualcomm attempting to buy exclusive rights to a certain TSMC production line in order to meet their supply needs. Both parties offered US $1B for this exclusive right which TSMC turned down. Intel not only has a better production process than TSMC, but it also has unutilized capacity that can meet market needs.

This exclusivity attempt by Apple, coupled with strained relations with its current foundry, Samsung, indicates that one of the largest customers in the foundry space is up for grabs. Intel likely has the capacity to meet Apple’s volume requirements and should be vying for the Apple foundry business. While production capacity alone might be enough to attract Apple, Intel has much more to offer to bring in this large volume contract. Intel’s position as the leading production process innovator allows it to produce more efficient, smaller processors which have a greater SoC capacity. With mobile device trends moving towards longer battery life, thinner form factors, and greater integrated capabilities, Intel could frame its offer to Apple as a superior processor that would significantly improve the end product. Intel would offer both a stable supply and a better processor product.

Intel can actually go one step further for Apple by offering some of its IP resources to be included in the SoC stage of Apple’s processor design process. Intel’s recent acquisition of Infineon’s wireless solutions business would be particularly attractive to Apple as Intel could offer its newly acquired cellular modem technologies. Qualcomm has traditionally included cellular and other wireless technologies in its SoC chips but this strategy has been unavailable to Apple as it does not have access to these technologies. By offering these technologies, Intel would be backing up the value chain from foundry into the larger SoC stage of the value chain, improving its attractiveness to potential customers and its market share in mobile.

Changes in the Computer Segment

ARM’s impact on Intel goes beyond the mobile market as ARM is currently also moving into Intel’s traditional stronghold, the PC processors market. ARM has forecast that it will capture 20% of the notebook computers segment by 2015. While much of the PC processor market will remain unchanged in the short term, this market is also moving in a more fragmented direction and Intel needs to adapt its approach to retain these customers.

Apple’s recent foray into the processor design step in its mobile devices has given rise to speculation that it is looking to switch from the Intel-based processors in its Mac lines to its own custom ARM-based processors. Industry analysts believe that Apple has already started looking into this possibility and could make this change by 2017, or even earlier. With processor sales to Apple accounting for 10% of Intel’s revenues, this could result in four to five billion dollars in lost revenue.

The foundry and SoC approach would also be a good fit here as it would allow Intel to maintain a portion of the value chain it could otherwise lose. One key distinction between mobile and PC is that the Intel architecture actually has an advantage in the PC space given that most established software is written to work with this architecture only. Intel should take advantage of this by designing custom IP core solutions with its PC architecture for its foundry customers, essentially offering customized processor design services. This would motivate Apple to stay with Intel, as Apple is seeking more control in processor design, while subsequently increasing what Intel can charge over the base foundry prices.

Moving Forward

Intel has prospered from its market dominance in PC processors. But the technology industry is very dynamic, and Intel must adapt to the structural change that mobile technology represents. Intel does not have the power to change the fragmented nature of the mobile processor market. As such, Intel should be willing to forgo control in segments of the value chain where it cannot compete effectively. By emphasizing stronger customer relations, Intel will be better able to compete in a fragmented mobile market as well as defend its traditional turf. If Intel can execute on these manufacturing and customer-focused strategies, it will be able to experience a renaissance on the back of mobile device growth instead of being strong-armed out of the industry.