When Advanced Micro Devices (AMD 2.60%) announced its next-generation Mullins product for tablets, it showed some impressive performance numbers from the part. Further, the chips are rated at quite an aggressive thermal design power, or TDP. While these TDP ratings generally don't tell the whole story -- and that includes from AMD or any of its competitors, including Intel (INTC -10.62%) -- they do give a pretty decent idea of what kinds of devices a part should be able to fit in. At the upcoming Computex show beginning June 3 in Taipei, Taiwan, be sure to watch closely for tablet design wins on AMD's part.
Finding tablet designs has been a problem for AMD
AMD has found it difficult to win designs in the tablet market, in no small part because of a very poor competitive positioning with its products. Its first tablet attempt, known as Hondo, was a performance and power disappointment. Its next attempt, Temash, was a significant improvement over Hondo but still couldn't quite cut it on a performance and power basis -- and probably from a platform bill of materials level. Mullins, AMD's third-generation tablet product, does look a lot more competitive.
However, despite the impressive performance that the initial previews have suggested, we still don't know a number of critical pieces of data, such as power consumption and battery life numbers and platform bill of materials costs required to support the platform. When the first designs hit the market, the hardcore tech sites will probably get us some power consumption numbers. If those look good, and if the delivered performance is as claimed, then the last major gating factor to design wins and commercial success is chip and platform cost.
Intel's contra-revenue scheme is a headwind, but not in the way you might think
The reason I'm so concerned with the platform bill of materials cost is that Intel had some difficulties here with its own Bay Trail-T platform and as a result needed to provide contra-revenue support to offset that cost to the OEMs. While Intel's Bay Trail-T, the one that required much support, was based on a PC platform but optimized for tablets, the Mullins chip appears to simply be a binned -- i.e., creme de la creme -- variant of its PC-focused Beema platform.
While AMD has indicated that it has done a number of things to lower the bill of materials costs for this platform relative to the prior generation, it is still likely that as a PC-focused platform it has a bill of materials cost more in line with Intel's Bay Trail-T than a platform from Qualcomm or MediaTek. If Intel is providing the appropriate bill of materials offset and AMD cannot, then even if AMD has any real-world performance or power advantages, they could be neutralized as a result of cost.
Foolish bottom line
Each generation, AMD promises a new world-beating tablet platform on some vector or another, but so far those paper victories have led to not much in the way of design win traction. However, keep a very close eye out at Computex for any design wins based on the Mullins product from AMD. If the company can deliver, then it'll be much easier to be more positive on the shares, particularly in a frothy market increasingly devoid of compelling bargains.
