Advanced Micro Devices (AMD) hit revenue guidance for the final quarter of 2012, but the numbers were still dismal. Revenue was $1.16 billion, down 9% sequentially from $1.27 billion, and down 32% from $1.69 in the year-earlier quarter. Guidance is for Q1 2013 revenue to be down around another 9%. Net income was $473 million in the red on a GAAP basis, and even on a non-GAAP basis was $102 million short of breakeven.
Despite that, AMD CEO Rory Read was surprisingly upbeat about 2013, predicting the a return to profitability in the second half. Of course, we've heard that kind of optimism from AMD before, only to be let down. Are AMD's claims of a turnaround ahead credible?
AMD is currently known for making CPUs that compete with Intel's (INTC) for personal computers (notebooks and desktops) and servers. In addition AMD makes stand-alone graphics chips (GPUs), competing mainly with NVIDIA (NVDA). AMD has not done well in the server space these last five years, and the PC space has started to shrink in part because tablets and smartphones have become more popular and mostly use ARM-based CPUs, rather than the more capable and power-hungry x86-coded chips made by AMD and Intel.
In 2012, in an effort led by the remarkable vice president of global business units Lisa Su, AMD started to re-target its intellectual property development toward growth sectors. The acquisition of SeaMicro acted as an entry to the dense server space, where AMD's graphics expertise could eventually help with highly parallel computations, and Opteron technology is a better fit than Intel's server chip designs. In addition, AMD has announced it will use ARM technology when appropriate in this field. Although the complete new system will not be available for some time, Read reported that Q4 SeaMicro revenue grew.
A second major line of attack is embedded SoC chips. This is a bit of a vague term; as used by AMD, it seems to amount to the non-PC sector. SoC, or system on chip, typically means that the chip is not a stand-alone CPU. In reality, even what we now call CPUs are not stand-alone CPUs: AMD has been a leader in moving critical components that "glue" the CPU to the rest of the system, like memory controllers, onto a single chip. Embedded SoC in this case means customized for a particular application. Read indicated AMD would be looking only at relatively high-volume applications as margins have been too low in some low-volume systems. Examples of possible embedded AMD chip use would be for advertising displays, casino machines, industrial, and medical use.
AMD does not currently break out embedded revenue, but the goal is to raise it to 20% of revenue by the end of the year. No word on whether that is 20% extra revenue or 20% replacement of eroding PC revenue, and Read made it clear no details would be announced until the OEMs are ready to announce them.
The PC business, of course, is still critical, as is the entry into tablet computing. Read made a point that AMD engineers are executing well, making their deadlines and, in one crucial area, are about six months ahead of Intel. Along those lines, AMD has demonstrated working Temash and Kabini silicon. These APUs will be quad core SoCs for the tablet and mobile markets. They also already introduced the new Richland A series APU, upgrading a sweet spot in their line.
This morning as I write AMD has popped from its pre-conference and results close of $2.45 up 9% to $2.67. Obviously, no one knows if AMD will be able to execute its plan or if, once products are available, they will sell well enough to bring AMD back to profitability. It has the look and feel of a good plan and a big turnaround to me. But I have been wrong about AMD in the past, and the sands of silicon are shifting rapidly and unpredictably. Before getting bullish on AMD, I'd like to see the 2013 products, the revenue, and the profits.
Disclaimer: I have long been long AMD and will not trade the stock for three days after the publication of this article.