Intel (INTC) bulls that proclaim that "mobile isn't important" are really missing the "big picture" here. In particular, the "big picture" is that in the future, the vast majority of client/consumer computing will be, in fact, these "mobile" devices. Smartphones already outship PCs by approximately a factor of three, and tablets - which I view as a type of PC - are already the world's most popular PC form factor.
The semiconductors, and in particular, the consumer computing market, have always been - in the words of a friend of mine - a "volume game." Intel got to where it is today and vanquished the big, bad RISC-architectures of old not necessarily by having the "best" chips, but instead by being able to offer the best chips per dollar. Intel, which had tremendous scale and R&D leverage due to its success in the high-growth, high volume PC market, was also able to powerfully work its way up the computing stack.
Today, Intel owns the datacenter business and it owns the PC business, which combined work out to somewhere north of 300 million processor units per year. This is pretty significant scale, and thanks to its manufacturing leadership and the insane amount of R&D that it can afford to design these chips, its products are far-and-away better than whatever its rival AMD (AMD) has been able to put out.
History Repeats, But The Shoe Is On The Other Foot
Intel, at this time, is on the other end of the "RISC-wars" that it won back in the 90s and the 2000s. Today, Intel faces three primary challenges:
- Samsung (OTC:SSNLF) - By far Intel's biggest problem. The world's second most profitable mobile device manufacturer and the world's second largest semiconductor company by revenue is getting serious about chip design. If Intel can't build chips good enough for Samsung to move a significant portion of its volume to Intel, then a large part of the overall computing market will be inaccessible
- Qualcomm (QCOM) - Qualcomm is the "Intel" of the mobile world. With over 50% of the apps processor market and over 60% of the cellular baseband market, it supplies the vast majority of the mobile silicon to mobile device players, large and small. Its products are also good enough to win sockets in Samsung's flagship devices.
- TSMC - Intel's big "competitive advantage" has been its manufacturing technology. While this advantage has worked out nicely in the PC/server space, Intel's design teams have not yet brought to bear the full potential of this technology in mobile. As a result, TSMC keeps getting the business via Qualcomm and others, strengthening its capabilities and bolstering the amount that it can afford to spend on R&D for future technology
If these companies continue to own the high growth mobile market, then Intel's share of the computing market will continue to diminish. From there, eventually its ability to invest in manufacturing leadership as well as in design will diminish. This ultimately leads to layoffs, capex cuts, R&D budget cuts and all of the usual things that come with the "death spiral."
Foundry Won't Save Intel
I can't believe the absolute silliness of the notion that Intel can just "be a fab" as a plan B. Do investors not realize that only 34% of TSMC's revenue is leading edge 28nm (~$6-7B), and that this revenue consists of chips for just about the entire semiconductor industry? If Intel took ALL of TSMC's high end business, it'd be worth about - you guessed it - $6-7B! Qualcomm's mobile chip business is worth $16B-plus per year, and Intel's PC chip business is worth $33B-plus. Foundry is just not the panacea that the newly-minted sell-side and retail bulls seem to think it is. Intel, to GROW from here, needs to sell the applications processors, modems, etc. - NOT collect $8-10 for foundry service for these chips.
Intel needs to become a major player in mobile and it needs to do so soon more for strategic reasons at this point (i.e. keep Qualcomm, Samsung, etc. out of the high end of computing) rather than for pure revenue/profit growth. These guys need to stop talking about Quark, stop talking about foundry, and make sure that their next round of mobile chip launches are on time and on-target, unlike the mess we saw with Bay Trail and what I believe will be a disappointment with Merrifield. Anything else simply won't do. If Intel fails - and we'll know by the launch of the next generation "Broxon" chip if it does - then it will be time to sell the stock.