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

Digitimes reported today that Advanced Micro Devices' (AMD) "Kabini" system-on-chip for desktops was delayed from a 2H 2013 launch to a 1H 2014 one. While this is widely believed to be the cause of the most recent drop in the share price, I believe that had AMD NOT chosen to delay its desktop-oriented version of its Kabini (that is, its "small," low power core), it would have likely led to a take-or-pay charge as a result of the wafer supply agreement ("WSA") that the firm has with Global Foundries. While there is plenty to be bearish about with respect to AMD, this particular delay is NOT one of them.

Kabini Is A Mobile Focused Chip Built At TSMC

AMD's Kabini APU is a notebook-first chip designed on TSMC's (TSM) 28nm HPM node. While AMD's competitive positioning in the PC space is weak across the board on a performance/watt basis, Kabini was designed to be an inexpensive to be a much more cost effective solution for the $400-$700 notebook price band. Since AMD's performance/watt significantly trails Intel's (INTC) in this space, and since AMD's die sizes on its "big" cores are fairly large, gross margins were likely under serious pressure. This meant that AMD had to double-time getting Kabini out to market before Intel could roll out its Bay Trail-M largely targeting the same segment.

Now, while AMD likely wanted to stop selling the more expensive-to-build Piledriver based APUs into the low cost space and instead sell a single-chip Kabini built on a smaller process geometry (i.e., lower cost), it seems that the company really had no choice. While Kabini for the desktop likely would have brought about some cost savings, it would likely be nowhere near enough to offset a giant take-or-pay charge, similar to the one that AMD suffered through last year. So, for the remainder of the year, AMD will likely be pushing its "Richland" based APUs while it ramps up production of its next generation "big" core part known as "Kaveri."

So, Is It Time To Freak Out?

I don't think so. AMD is a speculative play, and quite frankly, I have trouble understanding why bulls believe that AMD will succeed while Intel will simultaneously fail. It should be painfully obvious to anybody following this space that when Intel succeeds, it eases off trying to eat AMD's lunch at the very low end of the PC space, and that when Intel is in trouble, it looks out for numero uno and leaves AMD and its shareholders to bleed.

My guess is that the PC market will bottom over the next few quarters as Microsoft (MSFT) stops supporting Windows XP. This will probably drive a refresh cycle and from there AMD will be able to ride Intel's coattails to better performance in its PC space. AMD's upcoming GPU refresh could serve to try to gain back some market share from Nvidia (NVDA), although I remain skeptical that Nvidia will - given its much greater resources - let a new product launch from AMD go unanswered. Next, I expect AMD to drum up a lot of hype for its micro-server chips in a similar fashion to what Applied Micro (AMCC) has been able to do with its own X-Gene chip. Finally, I expect the game console news cycle/hype to help keep many of the more speculative investors buying, despite the razor-thin margins on these parts (~20% gross margin).

All told, the story is still the same one that we've known about for quite some time. AMD is in deep trouble, but if management can find itself a few profitable niches (think semi-custom on a broader basis), then there could still be upside to the shares. But be careful before going all-in on a company that has repeatedly destroyed shareholder wealth year-after-year. My position is very small, and I intend to keep it that way. Bet the farm at your own peril.

Source: AMD's Chip Delay Explained

Additional disclosure: I may go short AMCC at any time.