I was in Stratford to talk Shakespeare, not semiconductors, but sometimes there’s just no escape.
There I was two weeks ago, enjoying an expensive meal at Carter’s Restaurant, when I heard tech talk at the table next to me. “The P4’s a piece of crap,” a largish guy in a red shirt told his older buddy. “AMD’s going to be the one. They came out with the Athlon and it had its problems, but they worked those out. It’s going to be the better chip.”
AMD could not have asked for a better testimonial. If I had chosen to butt in on the conversation (which I didn’t — sometimes you just have to shut the meter off) I would have asked him for more details on those problems, since you almost never learn about the bugs except from resellers or users in the field. In the end, it may not matter, because this recent scandal over IBM’s decision to drop AMD chips from its PCs in North America and Europe demonstrates that technical glitches are only half its battle.
As some of today’s media coverage has pointed out, the IBM story is really much ado about nothing. IBM is no longer a major retail player, where AMD has proven strongest. Though Big Blue was an early supporter of the K6 series, it’s unlikely AMD will suffer serious financial repercussions from IBM’s rejection, unless the two companies did much more business together than was previously believed. Of course, tongues start wagging as soon as a customer drops a business partner (especially mine) but there’s a bigger issue to be addressed here: what does AMD have to do to position itself as an enterprise company?
After establishing a beachhead in retail (helped somewhat by the disappearance of Cyrix) AMD no doubt spent a good deal of money hiring a branding company to come up with a name for its enterprise processors. Athlon comes from the root of “biathlon” or “triathlon,” and in a sense this is part of the whole problem. Ever since 1 GHz speeds came within reach, Intel and AMD have been locked in a race that never seems to end. The Athlon name, intended to convey the idea of speed, unintentionally underscores the pressure the company faces to cross the finish line. In this case, the finish line represents a goal AMD has set for itself to capture 30 per cent market share by the end of the year.
I don’t think they’re going to get there, and that’s too bad, because they’ve got a good product. Though it will likely trail the next 2 GHz P4 chip in terms of clock speed, its core is capable of handling more instructions in the same cycle. It’s a competitive chip, but getting customers to notice that may take more time than their attention spans will allow.
In almost every one of the many product roadmap discussions I’ve had with AMD representatives, they’ve spoken at length about their competitive advantages over Intel. With Intel, on the other hand, it is as though AMD had never existed. Its name never comes up, and Intel will only acknowledge there is a race, or a price war, if relentlessly questioned. This is how you act when you’re a market-leading company. AMD has to take a similar approach, ignoring Intel publicly while doing everything it can to mimic its marketing strategies with a smaller budget.
Yesterday, for example, I spent more than an hour watching an Intel Webcast discussing the benefits of Itanium, using a video that likened addressing, prediction and parallelism as various steps in a grocery shopping exercise. AMD needs to do the same thing — show that its processor core is the equivalent of a bag boy who can load two bags at once. The worst thing this company could do right now is offer side-by-side comparisons with Intel that will fail to pack the punch of Intel’s more straightforward message. Courage, AMD: just because you’re the underdog doesn’t mean you have to act like one.