Intel's "low power IA" project, started in 2006, now into its sixth year is on its last legs. Based on current performance of its Medfield offering compared to what's available from competitors, it's only been able to garner 0.2% market share for smartphones in the first half of 2012. On the tablet side, the abject failure of Microsoft's Windows 8 platform has not helped of sales of CloverTrail based devices. In other words, in the current generation, Intel's low power processors have failed to make any dent in the market and this after six years of working on them. Now the question remains, will they be able to do something in 2014 when they accelerate their process roadmap for mobile. The biggest challenge they face is the absence of any high market visibility customer who will buy their tablet or smartphone targeted chips. As you know, the market for both these kinds of chips is mature and well fed by ARM licensees. Its been this way for several years now. Motorola, their best hope, is still only a fringe player in the smartphone/tablet market. In such a scenario, even if they are able to produce a market competitive chip, they may not be able to sell it.
In light of these observations, what's the prognosis for Intel?
1. They stop working on low power IA and start fabbing ARM chips. This way they can still leverage their process advantage while recognizing their inability to design microarchitectures in the low power space. They need to let go of their grand vision of a single instruction set architecture (IA) satisfying the needs of a range of power consumption scenarios. The window of opportunity for doing this is rapidly closing as the foundrys with their newly found mobile driven wealth are all ploughing money into new fabrication technology development.
2. They need to realize their limits in their ability to innovate and take steps to correct this. For a high tech company such a scenario clearly spells doom. Since they have not been able to convince Wall Street to move their stock in the last 10 years, its high time they assumed this would not change and used other ways to attract better employees.
Medfield was introduced in Jan 2012 at CES and was avaliable in Q2/2012. It will take more than 2 or 3 months of the first half to determine the ramp and success of Medfield.
There is near zero information on response to Win8. I would not bother buying Win8 unless I had a touch device. I expect that many buyers make the same decision. The cross over to Win8 will be driven more by touch input devices than conversion to Win7 which was the first real 64-bit offering from MS.
Another prognosis for Intel?
1. Using extra fab capacity to contract products is always on the table. Intel has a couple years to make and unmake that decision. The foundry decision and how much is independent of IA. They have capacity right now to do both. What is interesting about the Rosepoint die map is the large size of the WiFi section. The WiFi section is larger than two cores. Think about how adding the WiFi to the SoC will stress the most expensive foundry capacity.
2. There is no problem with Intel and its ability to innovate. The problem that Intel had was its success with netbooks. Yes. Netbooks. They filled the fabs and masked the evolution of the iPod MPG - iPod Touch - iPhone.
Intel innovation will be to integrate the phone to a single chip and then to have the process and capacity to ship product when Samsung will be the only other game in town.
It's nice to see that Intel is still deep Inside your head, even late at night..... and you need to justify to yourself that it is no threat to ARM. Thanks for showing us your cards. Who's your Daddie?