I will be the first to tell you that Intel (INTC) is probably an investment that I make my business to know inside and out. This is a big company with many moving parts, and it is very easy for investors to become confused about certain parts of the strategy, particularly the technically heavy parts. The purpose of my articles on this company has generally been to inform, rather than to make specific buy/sell calls (although my readers know I am an uber bull), and in that capacity I believe that my column has been a rather solid success. Today, I would like to bust yet another myth that I have seen perpetuated regarding Intel's process technology as applied to the low power space.
Intel Now Develops Two Flavors Of Its Process
In a recent article published here, the following observation was made regarding the company's 22nm FinFET process:
It's highly probable that Intel's 22nm FinFETs aren't optimized for mobile SoCs and Intel must have saved some of the tweaks for 14nm, because the real battle for mobility will start at 14nm.
This is completely, utterly, and unequivocally incorrect. Intel develops two flavors of its process at each generation, and 22nm is no different; one for high power/high performance products such as the high end laptop processors, desktop processors, and server parts, and the other for low power, highly integrated SoC products such as smartphone and tablet SoCs:
So, the notion that the "battle for mobile really begins at 14nm" and that the "22nm process isn't optimized for mobile" is actually completely wrong. Furthermore, the choice of the FinFET structure - which helps to seriously reduce leakage power (which is power used when the transistors aren't doing anything) - is actually an incredibly mobile friendly transistor structure since having significantly lower leakage (particularly as this becomes a problem at smaller geometries) translates directly into battery life gains.
It seems that TSMC (TSM), the foundry at which the leading ARM (ARMH) competitors will be building their products, will be using a 20nm planar (non-FinFET) process, which means that while I expect TSMC could have some density advantages at this node, Intel's performance characteristics, as well as leakage characteristics, should be significantly better.
The battle for mobile begins at 22nm, and I think Intel takes a decisive lead at the 14nm node, given that it will be on its fourth generation high-K metal gate (TSMC's 20nm will be its 2nd generation), second generation FinFET (TSMC's 20nm will be planar), all with the scaling and performance benefits that come with the next generation of transistors.
Make no mistake - Intel is going to be in a very good competitive position with the 22nm SoCs for smartphones and tablets, and the battle begins in 2013 and should rage rather violently through 2014. I expect the landscape exiting 2014 should be very different than the one that most investors envision today, and as a result, Intel's share price should be significantly higher than where it is today.
Additional disclosure: I may initiate a short position in ARMH at any time.