AMD guns for 10% of low-end server market
AMD is chasing double-digit marketshare in the low-end server and workstation market, according to the chip company's VP for workstation and server marketing, Ed Ellett.
AMD's thinking on this one is interesting. According to Ellett's comments, reported by DigiTimes, IBM and Sun will dither over whether they should stick with the x86 architecture or pull out of the market. Their indecision, says Ed, will allow AMD to get in there and - since AMD doesn't make servers, but Sun and IBM do - persuade those companies' rivals to adopt AMD's Athlon MP processor. The reason for this uncertainty on the part of IBM and Sun, says Ellett, is the falling price of x86-based servers.
Of course, exactly the same opportunity - if it is such, and we're not as sure about it as Ellett appears to be - applies to Intel, through its well-establised Pentium III Xeon and the new Xeon parts, based on the Pentium 4 architecture. And surely Intel will be working very hard to promote Xeon, not only to stitch up its arch-rival but to help recover the margins it's lost by slashing the prices of its desktop P4 line soo aggressively.
AMD has still to announce any big-name workstation or server vendors who have decided to use the Athlon MP. Intel, on the other hand, has lined up well-known (ie. big selling) customers for Xeon, including... er... IBM. IBM is also supporting Itanic, which is at least nominally aimed at all sectors of the server market.
Unlike Intel, however, AMD has nothing to lose, since it has almost zero presence in the server market as it is (thanks to NEC, there is an Athlon-based server out there, but it's based on the older Thunderbird core, not the new Palomino-based part). Whatever marketshare it gains is a extra business, and it can easily boast a massive increase in server marketshare even if it signs only a small percentage of server vendors.
To make a real contribution to AMD's bottom line, though, it's got to get some of the top-tier vendors on board. ®
Sponsored: Benefits from the lessons learned in HPC