AMD hasn’t been making a whole lot of Hammers. AMD isn’t going to be making a whole lot of Hammers any time soon.
Those are the only conclusions one can draw from their financial results and the discussion with financial analysts AMD had thereafter.
First, the financial results. AMD’s CPU revenues didn’t increase by anywhere near the amount they would have had they sold a large number of Hammers. They only went up about $75 million. Given that the average price AMD got for XPs also went up for the quarter, and that total unit sales were also up, it doesn’t look like they could have sold more than roughly 200,000 desktop Hammers.
AMD spoke very little about Hammers during the analyst phone conference. They said as much if not more about XPs.
The financial analysts asked again and again in differing ways how many AMD sold, and AMD refused to answer them. They wouldn’t even say if they had made their targets or not.
Finally, one analyst basically asked, “Could you make over a million Hammers next quarter?” and they said that they were capable of doing that and more if they had to, but they were “throttling” (that’s the word they used) production according to “plan” and listening carefully to their “customers” about production.
Oh. You throttle back the production of chips that make you a lot of money, but go full throttle on the cheap stuff. Does that make any sense to you?
It doesn’t unless you don’t think you can sell the increased production.
Another indirect sign that Hammer production won’t be ramping up any time shortly was AMD’s statement that they expected CPU revenues to be down somewhat the next fiscal quarter.
Under normal circumstances, that is a very reasonable thing to say. CPU sales do drop about 10% the first quarter of the year from the prior quarter.
However, here, if Hammer were ramping up, the much greater revenue per CPU sold would more than make up for a mere 10% reduction in unit sales. If AMD planned just going from 200,000 to 500,000 desktop Hammers (which is still a good deal less than 10% of production), CPU revenue would jump up quite a bit, not go down.
The Reasons Why
AMD’s comments made it clear that they couldn’t rapidly ramp up Hammer production even if they wanted to. Maybe they could make 1.5 million, which would only be around 20% of production. Even if you factored in die size, that’s not a whole lot (and besides, what’s better, making twenty CPUs and sell them for $80, or making ten processors and selling them for $250-$300?
However, it’s just as clear that production capacity isn’t the bottleneck; they don’t have much, but they still have more than they need.
If you listened to all the comments, the AMD execs didn’t seem to think Hammer was going to seriously rampup until they got to 90nm, and amazingly, they hardly sounded very enthused about even that.
Some of that lack of enthusiasm seemed due to going to 90nm itself. It didn’t seem like those working on it were exactly having a wonderful time, the task was described as “difficult.” The execs talked about “progress,” and said they were “bullish” and “optimistic.” You don’t use terms like that on something you’ve basically got done; you use them for something you hope to get done.
It’s still early, so don’t make much of that yet, but don’t be surprised if the schedule slips a few months from now.
There was another reason that was sort of, somewhat given for lack of big sales: software. Specifically, lack of 64-bit software. They said they expected sales to pick up once that became available the second half of the year.
Put it all together, and you get: Low production because of low sales from OEMs waiting for software.
AMD’s Financial Prospects
Take the last two quarters and AMD basically broke even. They made a little this quarter, they lost a little the quarter before. Even if they don’t make and sell more than a few hundred thousand Hammers the next few quarters, XP sales look to be holding up well enough so that the combo should keep AMD at roughly breakeven the next few quarters.
The revival in PC sales helps, but AMD’s CPU sales are still pretty bad compared to a few years ago. It’s been their cost-cutting and losing less money in flash memory that have been the principle reason for their financial improvement, not any great progress in CPU sales.
The lack of big Hammer sales is basically a lost opportunity to make serious money rather than something that will imperil the company’s existence.
Your Financial Prospects
After seeing the financial numbers and especially after hearing the conference call, I think what I’ve said about making a relative handful of Hammers and selling them for high prices is going to hold up. They really didn’t want to talk about them in the sense of them being the big and upcoming thing anytime soon.
And honestly, the sudden drop in enthusiasm about Hammer’s immediate and not-so-immediate prospects left me wondering more than a bit what else is happening behind the green curtain.