Advanced Micro Devices (AMD) finally gets it! In my previous writings, I have pointed out that AMD's new CEO -- Rory Read -- finally stopped trying to compete for the raw performance crown in PC processors and has instead decided to try to target the competition's weak spot. It is my view that Intel (INTC) has left a significant window of opportunity for AMD at the low end of the PC market as its "Bay Trail-M" part -- based on Intel's upcoming "Silvermont" architecture -- will not see release in systems until Q4 2013, giving AMD about a one quarter head start.
This is what I have wanted to see AMD pursue for a long time, and it's finally clear that the company as a whole "gets it." AMD likely cannot win the high end race against Intel, but the low PC space -- which sits in between the mainstream/high end PCs and tablets -- could be a place where AMD's willingness to accept a lower profit margin (bigger dies for less money) could lead to dramatically outsized returns for its shareholders. It is clear that AMD's presence in the new Samsung (OTC:SSNLF) PCs indicates that it has the right part for the right market at the right time. This, not promises about "HSA" and "Steamroller," is what will bring AMD back to profitability and will drive significant shareholder value going forward.
Remember Brazos? AMD's Best Selling Chips Ever
AMD's best selling chip ever was actually a small, svelte little platform known as "Brazos":
(click to enlarge)
While Intel was busy trying to push woefully underpowered Atom parts for the low end PC space (so that it would not risk cannibalizing its higher end parts), AMD saw some really nice sales and profitability growth thanks to these superior "Brazos" parts. Brazos was released on January 4, 2011 and drove significant FCF growth:
I believe that the situation with the upcoming "Jaguar" based "Kabini" APUs bears striking similarity to the "Brazos" launch (which is the core bullish thesis for AMD), and I do believe that this should validate AMD's claims that the firm should be FCF positive in the second half of 2013.
Jaguar Is Sort Of The Same Thing, With Two Critical Differences
AMD's "Jaguar" based APUs known as "Kabini" and "Temash" are successors to the "Brazos" and "Brazos 2.0" (which was a warmed over "Brazos") parts. These aim for the same markets -- low cost, low power PCs. This is a proven strategy that allows AMD to sell smaller, lower power, and easier to design and produce dies, rather than the bloated, expensive, and poor selling higher end APUs such as "Llano" and "Trinity." The strategy is certainly sound.
However, AMD faces two longer-term headwinds that were not present last time:
- The proliferation of Google's (GOOG) Android as a low cost computing platform of choice, which means that ARM (ARMH) based competition exists, suggesting AMD is not even guaranteed the #2 spot that it would normally have by default.
- Intel isn't putting out complete crap this time. "Bay Trail-M" SoCs should actually be faster/lower power on the CPU side of things, but AMD will still have a graphics advantage. In the prior generation, it was a clear blowout in AMD's favor.
Now, the upshot here is that despite the additional headwinds to a 2011-like boost from Kabini, these parts will undoubtedly drive revenue and earnings growth. Essentially, AMD is at rock-bottom in the PC space at the moment, so anything that can improve the situation will have a positive effect on the stock.
Game Consoles And Semi-Custom APUs: Sustaining Cash Flow
The problem with AMD is that every so often, the company rises from the dead, does something interesting, and then gets smacked down by competition or its own incompetence. This has made AMD more of a swing trader's dream and not so great of an investment. The looming threat of Intel hitting hard with its own low power CPUs (with process node and R&D spend advantages), and with Android taking more of the broad computing market share (enabling ARM competitors... at least until most of them kill each other off and the market consolidates as it always does in the CPU world), AMD needs a way to guarantee some level of positive cash flow. That's where the semi-custom APU strategy comes into play.
With the semi-custom business, the volumes generally aren't all that high (Xbox One and PS4 are exceptions), but the designs have a really long tail, so AMD will have a reason to keep buying wafers from GloFo, and the top and bottom lines should be propped up nicely enough. AMD also is exposed to a number of different areas outside of the traditional PC business, which lowers the overall risk profile of the business.
AMD is finally doing it right and settling into its role as a chip vendor that delivers an interesting value proposition to OEMs. It's better for AMD to attack this market intentionally than by designing large, expensive chips that end up relegated to the bargain bin anyway. I think that "Jaguar" based APUs will help the company return to profitability/free cash flow positive, and it is good that AMD has finally realized that it's all about making money. AMD's low power APUs are cheap to make and can be sold with decent gross margins at a cost competitive price point.