AMD: With Vista, Time to Re-evaluate Price/Performance

In terms of percentage, Schwarzbach said, more of AMD's customer base is concerned with compatibility, image stability (producing essentially the same system over a longer period of time, in high quantity), acoustics, power consumption, and energy efficiency. "Of course, the enthusiasts are notorious for not being concerned about those things; their cases are very large, [they have] cooled neon fans, several of them blowing full speed, to keep their really high-power rigs running efficiently."

Guilty as charged. But here's where the AMD strategy begins to take a sharp turn from its 2006 course: Suddenly, AMD can't claim a power advantage across the board, and certainly the notion of pairing two processors in a desktop system doesn't help the power envelope much. But hey, perhaps those factors aren't all that important to enthusiasts anyway. And it's enthusiasts who are responsible for helping establish AMD's brand presence; after all, they're the ones who populate sites like Anandtech and Tom's Hardware Guide. So why bother redirecting their attention to matters that don't interest them anyway - or at least, matters that aren't supposed to?

Why not keep their attention focused on the notion of raw power...while making a case for changing the definition of what "raw power" means? The introduction of Windows Vista makes the definition-changing part of that strategy all the simpler.

"The Quad FX platform and the FX 7-series processors are targeted at a unique market segment that values the performance more than the [thermal design power rating]," Schwarzbach said. "So we've optimized that line on the performance attributes, and pushed the TDP in order to do so. These are going in large cases with liquid cooling solutions and lots of thermal airflow designed into the chassis so it can handle the TDP; and folks are buying 1000 watts of power supply to power these rigs."

Isn't the enthusiast segment of the market still pretty slim? "It's pretty slim," Schwarzbach responded, "but the halo effect, the evangelist effect that they have on a person's behavior in the general market, I think have been understated. It's difficult to quantify, but we all know it's real and it's there, which is why you see continued investment from not only the silicon providers, but from the OEMs as well in this really important space. And the margins are great."

If AMD relies upon the high-performance, enthusiast segment of the market to substantiate its brand - the segment that doesn't care so much about low power and efficiency - but the market to whom their message trickles down, the recipient of the halo effect, does care about those things, then doesn't AMD risk creating a kind of disconnect between those market segments, thus strategizing itself into a niche?

"The answer is no," Schwarzbach responed, "because the distinction of the enthusiast market from the mainstream market has been strong...and now there are other applications like multimedia production, extreme video editing, that continue to define what the enthusiast segment is. They have always had different sets of criteria and features that they go after. But what we have found is once these folks have put a product through its paces, they become evangelists for the brand and the company behind that brand. They will have an understanding that most users do not have that same definition for what they're usage requirements are, but they will still give a very favorable impression of AMD, accepting and regardless of what the differences in their feature requirements are."

To that end, AMD CEO Hector Ruiz may have thrown a little monkey wrench into that strategy, when he told analysts during the company's last quarterly earnings conference that over time, customers in general won't care about numbers of cores the way Intel believes they will. "It's not about the cores," Ruiz said, before spending the next several minutes trying to pick up the pieces of what "it" actually is about.

So what is it about? "It's about what you can do with the cores," responded Pat Moorhead. Immediately, he promised, "AMD is going to be first to market with a true, native quad-core processor. Not two dual-cores glued together. So what matters most is the work that the customers are going to actually accomplish with those cores, not the fact that you have four cores. AMD has a very differentiated approach in that we decided, as opposed to rushing to market two dual-core processors and gluing them together, to come up with a native quad-core approach that from the ground up is optimized for four cores.

"It's still the number one important thing, what the customer can actually do with that," Moorhead concluded, "and do with it at the right thermal envelope and the right power envelope as well. That's what matters."

And there was that old AMD pitch again, the 2006 strategy: multiplicity with efficiency. AMD never lost the ability to build an efficient processor; but in a race that many credit AMD for having started, its competitor surged ahead. Now it has to decide how to make the gap seem unimportant, while at the same time closing that gap, and making the closing of that gap matter to all of its customers - the "mega-taskers" and the "mainstreamers."

This while Intel implements groundbreaking changes to its 45 nm production almost immediately, while AMD races to catch up to accomplishing the same feat the following year. For AMD, the year 2007 will definitely be one of redefinition: specifically, one that redefines the phrase "hard sell."