Seeking Alpha
Long/short equity, contrarian, independent research, tech
Profile| Send Message|
( followers)  

There's no question that Intel's (NASDAQ:INTC) investor meeting left a sour taste in many investors' mouths. In a great market, NOBODY wants to hear anything like "flat revenues next year." I don't really blame people for wanting to steer clear of Intel here, and I realize that the stock really is likely to be dead money until such time that the market is convinced that the company will return to growth. However, while bearish arguments are certainly welcome and encouraged, outright deception should not be tolerated by the investment community. There's a big misconception that bears continue to perpetuate about Intel's FinFET processes that needs to be addressed - and it will be in the following paragraphs.

FinFETs Aren't Any Good For Analog/Mixed Signal - LIE

One of my readers left the following comment on a recent article in which I pointed out that Intel was building its cellular basebands at TSMC until it could transition the product pipeline to its own internal processes (again, Intel bought Infineon about 3 years ago, so everything that was in the works then was built at TSMC):

The problem is internal to Intel. Intel's 3D transistors are too limited for all SoC functions. I have not seen any information that Intel can do mixed signal on their 22nm or 14nm processes. All the public information is that Intel farms mixed signal out to foundries.

Unfortunately, while this is an absolutely wonderful scare tactic, it's also 100% FUD. A cursory Google search in which I typed, "Intel 22nm mixed signal" led me to the following Intel slide from a few years ago:

So, right off the bat, there's some initial "evidence" that Intel can, indeed, build mixed signal circuits on its 22nm FinFET process (mixed signal just means that the given circuit has both analog and digital portions on the same chip).

Now, I wouldn't be a responsible analyst if I didn't go much deeper in this article to not only show that Intel's FinFETs CAN do analog circuits, but that the notion that FinFETs are bad for analog circuits is complete nonsense. With that in mind, off to the evidence.

Intel Gave An Entire Presentation On Its 22nm Process...Talked Up Analog

Maybe I was the only one paying attention, but last year, Intel's Mark Bohr (the guy responsible for the process technology at Intel) gave a pretty thorough overview of the company's 22nm FinFET process for system-on-chip solutions (the 22nm FinFET process for high performance processors had already been detailed). The slide deck can be found here (it's titled "Silicon Technology Leadership for the Mobility Era) and the accompanying webcast here.

Mr. Bohr showed the following slide:

(click to enlarge)

So if you look at the chart, the performance of the analog devices built on Intel's processes has been degrading over the last several years from the 65nm node to the 32nm node. However at the 22nm node, Intel was able to pull it together to improve performance to beyond-65nm levels. The notion that FinFETs (or Intel's implementation of FinFETs) aren't any good in analog circuits is simply ludicrous.

Conclusion

Intel is out of favor and of course, it's not "cool" to like Intel. In the minds of many, Intel is a bumbling corporation that wouldn't know what a mobile chip was even if a box of 10,000,000 of them showed up at Brian Krzanich's door. Of course, this is patently ludicrous. Did Intel screw up the products that came on 32nm and older? Absolutely. Did Intel's design teams not fully utilize the 22nm advantage across its SoC products by missing schedules and designing for the wrong market segments? Yep, no question - but at 22nm Intel is much more competitive than it was in the prior generation (leadership power consumption, CPU performance, really good graphics for tablets).

But does anybody truly believe that Intel's engineering and management teams don't know what they're doing and won't be able to eventually be a major player in phones, tablets, and everything else? Seems to me that people confuse lack of focus historically with lack of ability today - something that frustrates the living snot out of those of us who do extensive due diligence on the company.

Source: Intel: Another Lie You Shouldn't Believe