This article will discuss, elaborate, and speculate about statements made by Advanced Micro Devices' management during the Q3 conference call, relating to their pipeline of unannounced semi-custom and cloud design wins.
"Our game console wins are generating a lot of customer interest" - CEO Rory Read
CEO Read continues on in his prepared remarks:
"As we demonstrate our ability to design and reliably ramp production on two of the most complex SOCs ever built for high-volume consumer devices. We have several strong semi-custom design opportunities moving through the pipeline as customers look to tap into AMD's IP, design and integration expertise to create differentiated winning solutions."
Later in the Q & A session CEO Read adds this:
"One of the things you've really got to think about in this business, these are just the first two wins in the semicustom space. We have a pipeline of additional products and it's our intention to win and mix in whole set semicustom offerings as we build out this exciting and important new business."
Game Consoles will morph into ubiquitous TV consoles
One of the things that most Wall Street analysts are not thinking about is the massive transformation of the living room/TV room from a broadcast reception model, to an interactive internet model. Simply consider the multi hundred $ billion market cap Apple has attained, mostly from dominating the initial transformation of mobile phones from audio/text message devices to internet access devices.
Apple has publicly stated their intention on many occasions to compete in the living room, as has Google, and Samsung. Privately many other PC makers are targeting the TV console space, as are other TV manufacturers.
The difference between a cheap internet streaming device and the Xbox One is the user interface.
The user experience, which is based upon the elegance of the user interface, is everything. The huge leaps in the computer/consumer electronics industry have been built upon user interface inventions and innovations.
The Xbox One is just the first of many new TV devices that have a completely new user interface, based upon gestural and voice commands. While voice and gestural commands can be processed in the cloud, building them directly into a device reduces the response time, which increases customer satisfaction.
The Xbox One utilizes the superior parallel processing power of AMD's APUs for analyzing verbal and gestural commands. Without AMD's APU abilities, it is doubtful that Microsoft (MSFT) would have been able to create and sell the Xbox One for less than $500. The Xbox One will be the first TV console that does not require a handheld device.
The TV user interface bar has been set.
Now, it is up to Apple, Google, Samsung, etc. to bring out their own versions of a TV console that does not require a handheld device. What all these next generation TV consoles will require to effectively process gestural and verbal commands, without using the cloud, is a powerful GPU. If a TV console does not have a powerful GPU, and uses the cloud for processing non handheld commands, it will not be able to play high end video games effectively.
While Nvidia (NVDA) is pushing cloud gaming, the inherent time delay is very noticeable, by anyone who is semiconscious.
Just as with the PC, the most important part of the new TV consoles will be the APU, and the embedded operating system. Most of the massive profits derived from the PC revolution went to Intel and Microsoft, while none went to the actual creator, International Business Machines (IBM), and relatively little went to the alternative, Apple (AAPL). The revolution in the TV industry will be the same, except Microsoft will probably not be licensing their operating system to others.
AMD is positioned to derive vast profits from the next generation TV console industry
Because AMD is in all 3 of the major game consoles, new game console software is being optimized for AMD's chips. This is similar to most of the PC software being optimized to run on Intel chips. The key difference is that AMD has not licensed their GPU inventions to Intel.
In my opinion, it does not make sense for Apple, Google, Samsung, etc., to bring out a verbal/gestural command TV console (priced at $500), that cannot effectively play the top video games. Maybe the handheld game controller will be optional, but I believe that the ability to play top video games will become a key capability of all the new high end TV consoles. The difference in cost between a cheap yet ineffective Intel Atom chip for streaming video and the massive graphic processing power of the AMD APU in the Xbox One is only about $50. If a top user interface company like Google or Apple is going to build a TV console to compete with the Xbox One, I believe AMD has the inside track.
Therefore, Apple, Google, Samsung, etc., have a strong incentive to use AMD's APUs, rather than those from Intel or Nvidia in their next generation TV consoles.
Additional key technical reasons for others to follow the lead of Microsoft and Sony in choosing AMD APUs for their TV consoles.
AMD is years ahead of Intel in the GPU area, as measured by performance relative to price. AMD has integrated their superior GPU capabilities with an adequate CPU, into the APUs that are being used by Sony and Microsoft.
Intel has adequate GPU capabilities for PCs that are not optimized to run video games. Intel recently created a PC chip capable of competing with the graphic abilities of AMD's APUs, the Haswell Iris Pro 5200, but it is priced above $400. It is estimated that the AMD APUs being sold to Sony and Microsoft for their TV consoles are priced around $75.
While Intel was concentrating on reducing the power usage for their new chips, AMD was creating an APU for next generation TV consoles, in which power usage is not a consumer concern. What is a key concern for consumers is the price of the TV console. Even if cable providers end up subsidizing the upfront cost of next generation TV consoles, through monthly rental charges, price is still very important.
The only other company with competitive graphic capabilities to AMD is Nvidia . But since almost all the new top video game titles are being optimized for AMD's APU chips, Nvidia is in a difficult position. Nvidia does not have a license to use the X-86 architecture upon which AMD's TV console chips are built. Nvidia chose to create their own game console, which essentially competes with any new TV consoles that are video game capable. Most companies do not like to buy components from a competitor, if they can avoid doing so.
Bottom line, for all the above reasons it makes the most sense for Apple, Google, Samsung, etc., to choose AMD APUs for their next generation TV consoles.
The next generation TV console market is massive. Even people that do not like to watch TV programs, generally like to watch movies, documentaries, and other video that can be found on the internet.
One of the reasons that people like tablets is because they can easily watch video from the internet.
In my opinion, if people can easily access video from the internet, with a consumer friendly interface, most of them will choose their largest screen, and the comfort of their couch, over a PC or tablet.
"Verizon was the first major data center win that we could talk about publicly" - Sr. VP Lisa Su
Clearly, the Verizon cloud win for AMD's SeaMicro will be followed by other cloud win announcements, when those customers are ready.
Some people tried to suggest that only Intel's chips were in all the Verizon deployments, but as noted in this Seeking Alpha article, and by AMD's Lisa Su during the conference call:
"... we do have significant deployments with AMD Opteron as well." "We do see the percentage of Opteron processors increasing because that's what we'd like to do."
Unlike AMD's APUs for TV consoles, cloud chips will increasingly compete based upon energy efficiency
"In the server market, the industry is at the initial stages of a multiyear transition that will fundamentally change the competitive dynamic. Cloud providers are placing a growing importance on how they get better performance from their datacenters while also reducing the physical footprint and power consumption of their server solution. This will become the defining metric of this industry and will be a key growth driver for the market and the new AMD. AMD is leading this emerging trend in the server market and we are committed to defining a leadership position."
"We remain on track to introduce new, low-power X86 and 64-bit ARM processors next year and we believe we will offer the industry leading ARM-based servers"
-CEO Rory Read
AMD is the only other company with vast server chip experience and success other than Intel. Intel currently has a 95% share of the server chip market, relative to AMD's 4.5%.
To some this is bad news for AMD, but to me this current disparity means that AMD has incredible upside potential, to gain server chip market share back from Intel.
It is helpful to remember almost all of Intel's current line of server chips uses patented technology that is licensed from AMD.
AMD's IP and experience in the server chip area, forms the basis of my belief that CEO Read's statement about AMD being the leader in ARM-based servers, is accurate.
ARM architecture inherently more energy efficient than Intel's
Here is a quote from an article in the INQURIER by Lawrence Latif, with the subtitle: "Analysis High volume customers want to roll their own chips"
"ARM based processors are considerably quicker and easier to develop and do not require leading edge process nodes to have the energy efficiency that Intel gains with its advanced manufacturing capabilities"
Since the above statement emanates from the views of AMD server head Andrew Feldman, the founder of SeaMicro, here a few direct quotes from him:
"that the cost of developing ARM based processors is so low that chip vendors' biggest customers will be able to subsidize chip development cost and buy the products on a cost-plus basis."
"Ask Amazon, Google or Facebook whether it would spend $50m co-developing a processor that its competitors couldn't get hold of, and even offer a one or two percent increase in performance on its current software stack, and it is highly likely they will bite your - forgive the pun - arm off."
Mr. Latif goes on to write:
"Feldman's claims actually become pretty obvious when one looks beyond the silicon and considers what high volume server customers run on their equipment. Firms such as Amazon, Facebook, Google, Twitter and anyone else that heavily customizes both the operating system kernel and the userland to meet their specific business demands can gain a significant performance advantage by not having to use a generic x86 or ARM core that's available to anyone with a chequebook."
I wrote in a previous Seeking Alpha article about Verizon-AMD deal: Verizon's CTO John Considine has the vast resources of a Microsoft or Sony to create proprietary services, that cannot be easily copied, which explains the use of customized components, as stated in this Gigaom article:
"Last week, before Verizon CTO John Considine unveiled the company's new Verizon Cloud, he was vague about the hardware underpinnings but did say Verizon built its infrastructure from the ground up."VMware and EMC and Cisco and HP remain great partners but this is our own design. We did our own work in software-defined networking, in storage, in hypervisors and in orchestration," he said in an interview.
The Verizon server deal with AMD, is just another example of the paradigm shift in power to cloud players like Facebook, Google, Amazon, Apple, etc., from old line IT system manufacturers such as IBM, Hewlett-Packard (HPQ), Dell (DELL).
Paradigm Shift in IT
Instead of thousands of companies buying IT hardware from a few powerful system manufacturers, now a few cloud companies are creating their own systems by dealing with component suppliers directly. The paradigm shift in the IT industry will greatly benefit innovative component suppliers such as AMD.
AMD is using the innovative IP of SeaMicro, combined with their vast server chip experience to bypass the stranglehold that Intel held over a handful of old line system manufacturers.
AMD will combine their vast server IP, with the power efficiency of ARM's 64 bit architecture, to take back at least 20% of the server market from Intel, in my opinion.
A thinner client and fatter cloud will increase GPU demands
The vast gains in connectivity are allowing mobile devices to become lighter and smaller, yet still incredibly powerful by using the cloud for all major processing demands.
Google Glass is an example of a device that uses the cloud to process video/pictures and blinks into usable info. Cloud servers will increasingly become more efficient using the parallel processing capabilities of the GPU to analyze videos and pictures, as more software is written for these purposes.
Instead of manually inputting info into a device, your camera will relay the pictures of what you are looking at to the cloud for processing, along with your verbal commands. If you see a hot new product while strolling around, you will be able to simply aim your camera/glasses at it, and ask about it.
The most efficient way to process real world info is by harnessing the inherent parallel processing power capabilities of the GPU.
Therefore, in my opinion the GPU will increasingly become a much more important component of the cloud than the CPU dominated situation of today. There are more quotes and discussion in this Seeking Alpha article, entitled: Next Era of Computing favors AMD.
AMD's IP and experience uniquely positions them to grow and profit from 2 large paradigm shifts in information technology. AMD has already shown a dramatic operational turnaround from the TV console paradigm shift. The Verizon cloud announcement is an example of the potential for AMD to greatly benefit from the paradigm shift in demand for IT hardware.
AMD does not need to compete in the hyper competitive mobile market for thin client chips to greatly benefit from the paradigm shift to cloud services.
Wall Street investors seem to be obsessively focused on AMD's wallowing PC business, rather than the dramatic turnaround and potential in its new growth markets.
I believe that in a few years, almost everyone will have some sort of next generation "smart" TV console, possibly built directly into the TV, and AMD will be the dominant provider of the APUs. I also firmly believe the evidence supports AMD taking back at least 20%, and possibly a lot more, of the server chip market from Intel, over the next few years.
The global TV console market is at least couple hundred million units per year. Even if the average APU price drops down to $50 from $75, that is a $10 billion a year addressable market, of which AMD is uniquely positioned to dominate.
Intel's last quarterly server division sales were approximately $3 billion, which equates to a $12 billion a year. Therefore, if AMD only takes back the 20% market share they held in 2005 that works out to about $2.5 billion a year.
Because of the need for more GPU horsepower in the cloud, I believe that within 5 years AMD could replace Intel as the server chip leader. The cloud is a huge growth market, and therefore the server chip market should also expand, even if the average selling price declines.
Based upon the trends I have noted, it is not difficult for me to see AMD's server business doing $6 billion a year within 5 years.
AMD's present PC business is at a run rate of about $3 billion.
Bottom line, over the next 5 years, AMD is uniquely positioned to see revenues of its new growth businesses expand, such that they are at least 2 to 3 times bigger than the current PC division. More importantly, the new semi-custom and embedded businesses quickly become consistent free cash flow machines, because most of the costs are upfront and partially covered by their partners/customers.