The song The Edge of Glor wsa playing in my head as I listened to the AMD briefing on the SeaMicro acquisition. This is a big play for AMD, which could redefine the server space and position AMD’s technology more strongly against Intel in advance of the next technology wave. Generally, when a major technology change occurs the dominant vendor, in this case Intel, is the most at risk because it is unwilling to cannibalize its existing revenue streams prematurely to move to the new opportunity.
It is on changes like this that markets can move. Once done the SeaMicro technology will give AMD the capability to move aggressively on the new architecture and on a price vector. If AMD is successful and Intel lags in an attempt to protect Xeon profit margins, which may then be unsustainable, AMD could emerge as the leader on the next server wave. That’s the bet and it is a big one.
Next Generation: Massive Multi-Core
Ironically, it was Intel that first started talking about massively multi-core systems years ago. Microsoft took the concept one step further and towards the idea of a Random Array of Inexpensive Processors, the resulting acronym is unfortunate and so has never become widely used. Microsoft’s effort initially used Intel Atom processors and the result showcases Intel’s exposure. Intel did not concur with using Atom processors in servers at that time and Microsoft shifted to ARM processors. The resulting expertise its developed in the Windows group likely contributed significantly to their moving Windows itself to ARM for the next version and the expected eventual ARM support for Windows based Micro-servers.
The goal was to anticipate where the cloud market was going and create a ground-up redesign of servers and much like storage was shifted from relatively large non-redundant drives to RAID configurations that had cost, reliability (eventually), and serviceability advantages this idea of lots of lower-cost cores started to catch on. These servers would be massively electronically configurable, able to provision one to hundreds of cores to a particular task on demand. The result would be a class of server that was both more flexible and more cost effective for most cloud loads than existing designs.
The first major company to get behind it was AMD’s strongest large scale partner, HP, who announced its own ARM-based offering Moonshot. This project initially represented more of a risk to AMD than it did to Intel because ARM might have replaced AMD as the alternative processor platform for future servers. And, should the effort be successful, it is at least possible that ARM would replace Intel eventually for this emerging.
Calxeda vs. SeaMicro
There were two small companies that had emerged early as leaders in this new class of server. Of the two, SeaMicro appeared to be the more powerful but they were also wedded to Intel technology. Calxeda is behind HP’s project Moonshot and got a massive boost when HP announced it would be delivering volume product in 2013 using Calxada’s technology. Calxeda would have likely been less expensive to purchase than SeaMicro and would have assured the HP relationship but AMD is at the front end of its ARM efforts and Intel, with SeaMicro, is best able to position against them. By taking the more expensive SeaMicro it can more quickly get to market with an AMD-based offering, given the similarity between AMD and Intel technology, while denying Intel access to SeaMicro.
In short, Calxeda was the better tactical move to preserve the HP relationship but Intel’s likely strategic response with SeaMicro would have eliminated it quickly. SeaMicro was the more strategic because it both speeds AMD’s time to market and, more important, slows Intel’s response.
Wrapping Up: AMD Is On The Edge of Glory
Companies are defined by the big bets made by its leaders. In this move, AMD’s CEO Rory Read is attempting to move aggressively out from under the shadow of Intel and move to slow Intel’s response. Intel already slipped with Microsoft which opened the door for ARM as an alternative technology on their home PC and server turf and now AMD is moving to become the stronger ARM alternative.
The acquisition itself is mostly focused on the SeaMicro technology and its success isn’t based on retaining SeaMicro’s existing customers making it far easier to execute it successfully. Blended parts will likely take 12 or more months to create and during that time Intel will be moving to create an organic competitor internally. The move slowed but didn’t stop them and they remain the better funded.
This leaves AMD at the Edge of Glory but closer than they have ever been to taking and retaining critical server leadership in an emerging market. AMD gets kudos for taking the first big step out of Intel’s shadow, now it just has to execute.
Rob is president and principal analyst of the Enderle Group. Previously, he was the Senior Research Fellow for Forrester Research and the Giga Information Group. Prior to that he worked for IBM and held positions in Internal Audit, Competitive Analysis, Marketing, Finance, and Security. Currently, Rob writes on emerging technology, security, and Linux for a wide variety of publications and appears on national news TV shows that include CNBC, FOX, Bloomberg and NPR.