More Heat, And A Little Light 2586

Some people didn’t like what I had to say about Hammer sales the other day.

It proved to be a good test of AMDroidism. The more loudly and irrationally one squealed, the more likely it is the squealer has contracted the condition.

You can see a representative sample of the kind of comments generated when I do such things here.

For the further educational benefit of those capable of benefiting from it, let’s go over some of the at least semirational points most often raised.

“They’re Selling Very Well” Many knee-jerked this reaction. Unfortunately, that’s all it is. Generally, when a new CPU product line is introduced (whether from Intel or AMD), you usually see that new line get about 20% market share by the second quarter after introduction.

For that matter, AMD seemed to have anticipated that, too. Three months ago, the AMD execs said that they had the capacity at Dresden to make “significantly” more than a million Hammers per quarter, which would be just about right for a normal, regular rampup.

Instead, AMD apparently sold around just one-third of a million Hammers. It’s not clear whether or not the capacity in Dresden just laid idle, or could be used to make XPs instead, but even if it were the latter, it would seem to be rather more profitable for AMD to make Hammers than XPs.

The average selling price of XPs is the big factor in trying to estimate overall sales. Back when AMD released these figures, a $60 ASP was a terrible quarter, guaranteeing big losses no matter what. AMD has indicated that ASPs improved considerably on XPs a few quarters ago, and have stayed steady or slightly nudged up since then. Given that, a $75-80 ASP for XPs should be about right.

Nor does it seem like AMD is bursting at the seams trying to keep up with the demand for XPs. If AMD were making, say, eight million XPs a quarter, that alone would account for more CPU revenue than they showed total for the quarter. We would see a CPU revenue figure more like $700 million rather than the actual $571 million figure.

Of course, it would be better not to rely on estimates and rather hear it from the horse’s mouth. That is just what the stock analysts have been desperately trying to get from AMD during these investment conference calls, just so they can estimate how the company is going to do and advise investors accordingly. AMD refuses to answer them, again and again.

There’s something else AMD does during these calls. If they think the news is good, they talk it up big-time. When the news isn’t good, they get quiet big-time. In 2002, they started talking up Thoroughbred big-time, then suddenly around the time of product launch, “Thoroughbred” practically vanished from their vocabulary, and did so until they fixed it.

In 2003, they talked up Hammer big-time, but in January 2004, it, too, fell off the map. The AMD execs barely mentioned it.

Now use some common sense. Do you think they’d do that to hush up good news? Especially when they have a track record of duck and cover whenever the news isn’t good?

A variant on this theme was “Well, they’re selling very well at Best Buy or my local shop (or wherever).” This is what I call the “iMac fallacy.” Why, yes, if you’re selling one or two models of X, and twenty models of Y and thirty models of Z, it wouldn’t be suprising if the one model of X outsells any particular model of Y or Z. That’s the wrong comparison, though. If you want to know how well X is doing compared to Y or Z, you compare the X sales to the sales of all of models of Y or Z.

There’s a handful of OEM Hammer desktops on sale. There are hundreds if not thousands of PIV or Athlon XP models being sold, too. If two places in your town sell Rocky Road ice cream, and two hundred sell vanilla, Rocky Road sales at those two stores will probably be more than the average store’s vanilla sale. But that doesn’t mean Rocky Road is more popular than vanilla.

“Why Should I Not Buy A Processor That Isn’t Selling?” Many couldn’t see the logic, and I was probably a bit too cryptic about that to try not to go over old ground yet again.

Generically, you have to wonder why AMD seems so lackadaisical about establishing a firm foundation for a new generation of CPUs that is supposed to keep the company alive the next few years.

There are legitimate reasons why AMD would not want to go with a full-blown ramp until the end of the year (90nm, Windows x86-64), but we’re not talking about that. We’re talking about a company only producing a small fraction of the CPUs it says it already has the capability to make.

So why aren’t they? We don’t exactly know why, which doesn’t mean there aren’t any possibilities. In fact, there’s too many possibilities, ranging from the obvious to the speculative.

Pricing is the most obvious problem. AMD is trying to position Hammer as a premium product with premium prices, and people aren’t buying it. The issue isn’t what an FX costs; the issue is what the cheapest Hammer costs. AMD’s pricing strategy has been to try to get maximum dollar from a perceived large group of enthusiasts willing to pay just about anything for one.

Well, they don’t exist in anywhere near the numbers AMD apparently thought they did. Very roughly, they probably thought they could sell two million in the last two quarter, and they probably sold just a half-million instead.

A Dazed and Confused Roadmap is another problem. First, there wasn’t supposed to be any dual-channel desktop Hammers, even well after Intel made it obvious they were going that route, even though the Hammer designs always had dual-channel memory controllers built in.

Then at the last minute, it dawned on them that this wasn’t too bright, so they renamed some Opterons and started working on the dual-channel desktop solution they should have had last September. That won’t be out until late May.

Then they decided that socket 754 had to go from the only desktop solution to the budget basement solution towards the end of the year, while simultaneously trying to sell processors for the soon-to-be budget basement solution for top dollar.

And, by the way, three months after we finally get it right, we’ll be selling 90nm chips, which will sooner rather than later offer more performance at a much lower price.

This is a company that can do no wrong and shouldn’t be criticized? Is it any wonder these chips aren’t selling too well?

And these are just the obvious reasons. Who knows what other ones might be lurking behind the Sunnyvale corporate halls or Dresden walls?

That’s the problem with obsessive secrecy. When you say nothing about a problem, people think it could be anything and everything. People don’t stop thinking there’s a problem; they think it’s more of a problem than what you actually have.

The irony of it all is that this is a good product, and we’ve never been against the core technology. As of the moment, we’d be very surprised if we didn’t find 90nm Hammer systems the best platform to buy down the road, and heartily recommend it.

What we have been against is what the corporate boobs have been doing with it.

For instance, we pointed out very early on that demonstrating 800MHz Hammers was a pretty good sign that AMD was having serious problems developing Hammer and that it would be delayed. AMD said no until they said yes. We said very early on that desktop Hammers had to have a dual-channel option. AMD didn’t think so then, it does now. We said Hammer was really a 90nm product. AMD didn’t think so then, looks like it thinks so now.

We said AMD ought to make clear when they were going to go to DDR2 memory controllers; they’ve started to do that. Every step of the way, the AMDroids have called me an idiot for saying such things, but in the long run, who’s been the idiot?

Of course, when you’re dealing with an AMDroid, none of this matters, which is why the most popular point made was:

He’s An Idiot…

consult the thesaurus for some synonyms, just for a little variety? 🙂

Having said that, I must admit I rather liked this description, “Ed consistently writes as though he has a urinary tract infection and a buttplug at the same time, and is attempting to remedy both by drinking distilled wine vinegar straight from the jug.”

Actually, that’s quite inaccurate, I only drink that when I can’t afford battery acid.

Anyway . . . . 🙂

All joking aside, to an AMDroid, there is only one criteria. “Is he for or against AMD?” Once that binary switch has been flipped, all mental computations stop.

Generally, if you criticize AMD, the switch gets flipped to the “0” position, and critic = idiot.

I mean, really, what have I really been saying? Over the course of time, my criticisms have been aimed at improving the company, the product, and its sales. I’ve said, “Don’t do that, do this instead.” What’s wrong with that?

Obviously plenty. 🙂

For those who just have that single switch, trying to reason with them is like trying to reason with a flatworm. This is why I call them AMDroids; it’s due to the lack of cognitive ability on this subject. There’s no there there on this subject. The concept that criticism of “their” company can be legitimate or even constructive is completely beyond such people.

And they call us an intelligent species.

The problem I have with AMDroids isn’t the AMD part; it’s the droid part. It’s the lack of thought/refusal to think beyond a single rigid point that’s the problem.

As I’ve said many times, most AMD fans are not AMDroids. An AMDroid is one who think AMD can do no wrong. If you don’t think that, you’re not an AMDroid.

Actually, AMDroids aren’t really AMD fans, either. What they really are are fanatics who just happened to choose AMD for their fanaticism.

They’re hardly unique. The briefest glance at the daily paper will tell you about many others, but they all have something in common. They’re all fanatics, and all fanaticism boils down to is a refusal to think past a very limited point.

Be the first to comment

Leave a Reply