I made good money from Advanced Micro Devices, Inc. (NYSE:AMD) last year. The console win was probably the biggest catalyst for AMD's amazing 2013 run. The $110 dollar royalty income per one unit sold of the Xbox One and PlayStation 4 is a great bonanza for AMD.
Despite the rise of smartphone and tablet gaming, people still can't get enough of the new consoles. There's an iPhone-like demand for the PlayStation 4 that Sony is having a hard time supplying market demand for. Sony sold almost 6 million units of the PS4 already last month and Microsoft might have sold around 4 million of its Xbox One.
AMD is already the biggest winner in the console wars, with more than a billion dollars from its superb semi-custom job for the Xbox One and PS4. Since there's still a strong demand for these new consoles and AMD is still the current leader in cryptocurrency mining, I'm gradually increasing my AMD position.
However, those two catalysts are not the only grounds why I'm going polygamous - getting cozy again with the very sexy AMD while still being married to Intel. AMD's upcoming Seattle Opteron 64-bit ARM-based server CPU is a very enticing preposition.
Unlike my esteemed fellow contributor, Ashraf, I have much faith that AMD can also finish well not only in consoles or GPUs, but also in server SoCs. AMD has the know how and experience to become a factor again in the server market.
SeaMicro Gives AMD A Big Stick On Micro Servers
AMD's purchase of cloud computing-centric SeaMicro in 2012 is going to pay off with the release of the low-power Seattle server SoC. SeaMicro's Freedom Fabric technology will get maximum optimizations from having a 64-bit energy-efficient server SoCs- which allegedly offers lower operating costs than traditional Intel Xeon or Atom C2000 server racks provides now.
I don't want to dive into technical specifics. Despite being a network administrator for almost 11 years, I do not have a computer engineering diploma (or geek badge credibility) to delve into why the low-power 64-bit ARM architecture might become a cost-efficient alternative to the traditional x86 based server solutions.
To my understanding, SeaMicro's Freedom Fabric IP is so good that it reduces full server computers into card-size dimensions using its proprietary ASIC technology. SeaMicro then stacks these (10u) micro server PCs (up to 768 SoC) together to make a high density server machine which offers the same computing power of a more expensive 82u size conventional server rack.
SeaMicro microservers are not only 8 times smaller than traditional server racks, they also uses 3x or 4x lower energy. SeaMicro still uses Atom SoC products for some of its microservers but I'm betting that once the 64-bit ARM-based Opteron 1100 hits commercial phase, AMD will ship out most of SeaMicro products with the said SoC.
The intense competition and need for profit margins, compels even the biggest companies to operate their cloud computing requirements with minimal overhead. Microservers like what AMD/SeaMicro offers now, will become more attractive to corporate clients when they come with energy-efficient ARM SoCs.
Cloud computing firms not only save money on cheaper AMD hardware but they also save money on their energy bills. An electricity cost is the biggest overhead in running server farms. Despite the availability of Intel's Atom C2000 series, Google is still pursuing its need for a very energy-efficient customized ARM-based server processor.
The cutting-edge benefits of Freedom Fabric impressed Verizon last year. Verizon Cloud Compute and Verizon Cloud Storage are powered by SeaMicro SM 15000 microservers.
AMD Has the Credibility And Experience
Intel enjoys a monopoly-like hold of the server CPU market today and AMD's Opteron chips only command 4.5% market share. It wasn't always like this, AMD used to have a 24% share of the server market in 2006.
Unfortunately, Intel's bigger R&D budget and (very intimidating) influence among conventional x86 server players like HP and Dell, eroded AMD's server chip sales over the last 8 years. SeaMicro's Fiber Fabric technology plus good performance/price of Seattle Opteron, may help AMD regain its former appeal among corporate buyers.
The argument that multinationals already hooked up with Intel's x86 servers will resist migrating to ARM-based servers because it will be expensive and difficult, is just a fallacy. It will cost some money to port x86 server codes to ARM instruction set. However, savings from lower energy bill that microservers and 64-bit ARM-based SoCs promises to deliver will offset the cost of migrating x86 codes to ARM instruction.
The promise of very energy-efficient ARM-based server processors prompted Facebook engineers to make its HipHop Virtual Machine JIT compilation engine (which runs all of Facebook's server farms) to be compatible with 64-bit ARM architecture. Reading the blog of those engineers convinced me that it wasn't really hard to make their data servers ready for ARM-based processors like the Seattle Opteron A1100.
All it needs for AMD to again regain its former market share in servers is to have a good marketing team so it wins wider support for Seattle Opteron. The Mullins tablet chip, despite its outstanding features, failed to win support from big OEM tablet makers.
AMD is supposed to ship out demo units of the Opteron A1100 this month. It is my best catalyst for AMD this year. I hope it doesn't disappoint. I'd like to see AMD increase its server sales by 100% by 2015. It will all be in vain if this product gets ignored again due to lack of promotion.
AMD is not the only one trying to cash in early on 64-bit ARMv8 server architecture. Applied Micro is just one of several firms that are coming out with ARM-based SoCs for that big $57.5 microserver pie. AMD's long experience in server processors should be enough to convince enterprise clients to pick the Seattle Opteron over competing ARM server processors.
My faith in AMD's technological prowess is also based on the fact that it was the one who pioneered 64-bit computing. It's ironic, but my AMD winnings in 2013 help fund my December/January Intel (NASDAQ:INTC) purchases. However, the rumor that Microsoft is releasing full native Office for the iPad this March 27 will compel me to sell at least half of my INTC holdings to buy either more AAPL or AMD, or MSFT.
AMD is a strong buy for me. I want to make some money with this ticker again this year. How I wish Google or Qualcomm buys AMD soon.
Disclosure: I am long AMD, INTC. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.