We reiterate our Sell rating and $4 price target on Advanced Micro Devices following the announcement of its purchase of SeaMicro for $334 million.
We understand AMD's (ticker: AMD) desire to expand its presence in the higher-margin server space, where it has to-date lagged Intel (INTC). Intel's Data Center group has been a key source of margin upside and revenue for the company as sales of PCs slowed.
However, AMD's purchase of SeaMicro, a manufacturer of micro servers powered by Intel's Atom and Xeon processors, seems questionable to us -- it puts AMD in a position where it will potentially compete with its own customers while purchasing chips from its main adversary. SeaMicro's fabric technology is certainly impressive, and can already work with a variety of architectures and protocols, including AMD's x86 processors, and management suggested that SeaMicro products will include AMD by the end of the year -- we are skeptical, as our sense is that Intel technology is partially responsible for the appeal of SeaMicro servers. We view the data-center product cycle as one of the key thematic drivers for Intel, but our sense is that AMD's purchase of SeaMicro does not readily enable it to participate in the cloud buildout.
AMD is purchasing SeaMicro for $334 million, $281 million of which will be in cash. The deal will require a quarter of AMD's cash at a time when research-and-development spending is already constrained, reduce net cash per share to nine cents from 48 cents, and reduce tangible book value to $1.58 from $1.97. Management did not share details of SeaMicro's revenue expectations, but reiterated the guidance provided during the February analyst day, implying little incremental revenue in the near future.
Our 12-month price target of $4 is based on one times our estimate for 2012 tangible book value. Our 2012 revenue/pro forma earnings-per-share estimates are $6.5 billion/58 cents (consensus $6.7 billion/65 cents); 2013 are $6.8 billion/78 cents (consensus $7.2 billion/82 cents).
I think he's wrong about excluding the possibility of integrating some form of SeaMicro IP into a CPU or APU die.
The whole point of the SeaMicro arch is to jettison components from each server instance, and have centralized resources that are virtualized for each server instance through the fabric.
SeaMicro's long-term value is in the fabric and their software that manages the system-wide resources, load-balancing, etc.
Integrating the fabric on-die would increase the SeaMicro value proposition: I.e. the performance/watt of each individual server instance. Integration lowers the "watt" side, and moving beyond Atom increases the "performance" side.
Think about this: The SeaMicro fabric has nearly the same topology (if not scalability), and plays the same role as Cray's SeaStar interconnect... Tuned for volume servers.
I think he is wrong as well, but this approach plays much better with ARM's SOC ecosystem than it does with the vaporware SOC ecosystem in the x86 world. Of course the ARM ISA doesn't have much server traction. So both ARM and x86 lack a piece of the puzzle for implementing the SeaMicro approach as a single integrated SOC. Both can correct the problem with time.
I've been claiming for some time that innovative server startups could take ARM's SOC approach and do something really disruptive in that space. TWT.
I'm very bullish on the prospects of Smartphones and see them as a vital standalone device, just as I do with laptops/notebooks.
It is tablets that I see as the poor second cousin and a device that can't replace either a smartphone or a laptop/notebook.
I may be buying a Kindle of some sorts in the near future and whilst that is in tablet form, I don't see it as the kind of supposed laptop replacement that people like yourself have suggested the Ipad and Xoom would become.
But if you lug a DT around to play games, why not a tablet instead of a smartphone just for the bigger screen? The mode of holding and play would be more or less the same for games on either a phone or a tablet, so its just a matter of whether you don't mind something a little bigger for the bigger screen.
I know the phone fits in a pocket, but as I've stated before, fashion will adjust to take care of that problem for tablets too...
I don't lug my DT around for games. When I go to a particular friends place, I take my DT box coz I have various video's and need a computer whilst there to access the net.
And as I mentioned previously, I can't wait till the right kind of laptop comes along, so I can stop lugging my desktop around. I would probably also see this friend more often, as the hassle of taking my DT with me, means I limit how often I visit.
Your previous comments about fashion increasing pocket size were and are ridiculous.
What are the killer games that are best on a tablet? Or only on a tablet?
For me and most people, I need to have a mobile phone on me anyway, so gaming on it is fine. But why go to the hassle of regularly carrying a tablet with you, just for gaming on a train or whatever?
If I were to play Angry Birds or whatever the latest popular "Mobile" game is, it is a lot easier to hold a phone whilst on a bus or train, than to being holding a tablet.
People at my work place access Facebook on their smartphones all the time, and at home they use their laptop or desktop for that, none of them have the need or are using tablets to do that.
A tablet like a Xoom is doing nothing as a laptop or desktop replacement and who other than yourself, is surprised by this?
Book readers will be an area of growth, but they still aren't replacement candidates for laptop/desktop.
You failed to mention that a laptop is worse compared to a phone than a tablet is. The ONLY advantage of a tablet over a phone is larger screen. The ONLY disadvantage of a tablet over a phone is that it is a larger device, so more awkward to hold and carry.
Different people will make different tradeoffs between these two issues.
As I keep saying, I think something on the size of 5" dia screen is what wins. Its as big as one can get, while still working comfortably as a phone. But TWT. Could be 6" becomes popular.
It seems that part of the motivation for this move could have been to bring key talent back into the fold. The SeaMicro CTO and founder, Gary Lauterbach, was one of the five main designers (of the, I guess, more recent incarnations) of the Opteron. After a long carrier at Sun he spent two years, 2005-2007, at AMD as an AMD Fellow. It will be interesting to see which role he will now play at AMD.
Another key talent, one of the designers of the original Athlon 64 and Opteron, and then AMD CTO, is Fred Weber, who sits on the SeaMicro's board of directors. I noted at the time with disappointment when Weber left AMD in 2005. His replacement, Phil Hester, never created an impression in his short stint as CTO; although interestingly he was the co-founder of Newisys, a server startup betting on Opteron, which was remarkable back then.
In some sense, there is a deja vu feeling about the SeaMicro acquisition. It seems Read is trying a rerun of the successful times in the server space.