The Inquirer reports that when AMD introduces the FX-53, it will stop making the FX-51.
If this is true, what that would look to mean is that if you want an FX, you’re going to have to lay out $700. You can’t wait until older models move down the food chain to get these CPUs at lower prices.
AMD is essentially saying, “How dare you expect to pay a measly $400 for an FX?”
This is insane.
Recipe for Disaster?
Granted, AMD has had a problem with perceived value. The market has historically not valued AMD processors as much as Intel processors.
It seems that the folks at AMD think the solution to this is to simply jack up the price and not selling anything for less than a skyhigh price.
This is like Chevrolet expecting to increase its profits by not charging less than $50,000 for any Chevy. All you do by doing that is end up with a lot of unsold Chevys, because most current and potential Chevy owners aren’t going to pay $50,000 for a car no matter how good it is.
Mind you, the issue is not the price of the flagship chip. If AMD wants to charge $700 for an FX-53, I couldn’t care less.
What is the problem is not having lower-rated chips selling in the mainstream CPU price range, which even for Intel is $150-$200. Up to now, AMD supporters could say that not enough time has passed for prices to migrate downward, but if this decision is true, that means AMD isn’t going to let prices migrate downward over the course of time naturally.
Yes, Intel charges even more for its Extreme Edition, but Intel also has PressHots on sale for as low as $175, too. How can AMD justify a perpetual $700 price tag only for a type CPU with an extra 512K cache and dual-channel memory when Presshots with the same features (plus a free space heater :)) cost a quarter the price, and even the top end PressHots (or A64s) cost little more than half the price?
Mind you, it’s not like there’s going to be a ton of new FXs coming down the pike anytime soon. Until 90nm, FX-53 is going to be it. Having two horses in your thoroughbred stable is hardly one too many.
Why wouldn’t you want a $400 FX-51 going one-on-one with a 3.4GHz PressHot? Looks like a pretty good matchup to me (and probably even better in socket 939).
Why would you not want to sell a CPU for $400? How many sales are you forfeiting by not doing so?
No, I’m not trying to finagle a way for overclockers to pay less. Precious few overclockers will lay out $400 for a CPU. Those folks that will are probably almost all Intel business power users, and you hardly are going to get those unhappy with PressHot to change over to AMD by telling them to pay about $350 extra for one.
Right now, it does not seem like AMD wants to sell a lot of Hammers. They seem more interested setting a high price than actually getting paid, and come on, they can’t make money from a $400 CPU?
It’s not a matter of capacity, or the chips being hard to make. In the last analyst conference call in January, the AMD execs indicated that they had the capacity to make somewhat over a million Hammers a quarter, but they don’t seem to be making anywhere near that many, and they apparently won’t tell even financial analyste how many they do plan to make.
To the extent that they want to sell them, it seems like they’re trying to push almost everyone into socket 754, apparently just to create this “luxury” brand.
This is like Marie Antoinette saying about the mob saying that they couldn’t afford bread, “Let them eat cake.” Guess what? The mob found an alternative.
AMD has no problem selling Opterons, which are essentially the same thing as FXs when it comes to die size, cache and dual-memory channel capacity, for $278, $218, $178 or $163, so you can’t say they’re so hard and expensive to make.
You have to ask, “Why?” AMD can’t afford the luxury of a luxury brand today. First, they need to get the new Hammer platform adopted into the homes of a lot of people, and even with some price reductions, make a lot more money making Hammers than they do making XPs.
Later, after they’re getting Intel-like prices for the bulk of their CPUs, then they can get luxurious.
Right now, it seems like AMD is forfeiting five or ten $300 sales to get one $700 sale.
“I Don’t Think We’re At Motorola Anymore”
When I see this, I can’t help but remember where some of the AMD execs come from, what they did when they were there, and what happened during and after their tenure.
The CEO is Hector Ruiz, who had been president of Motorola’s Semiconductor Products Sector before joining AMD. The CFO (chief financial officer) of AMD is Robert Rivet, who was Hector Ruiz’s finance officer at Motorola.
When these two were at Semiconductor Products Sector, they basically reorganized the company a lot (three times in three years), and laid off a lot of people, sold fabs, things like that. It didn’t help a whole lot because after these two left, sales dropped about 40% during the slump in the electronics industry, and haven’t recovered since. Motorola recently spun off the division because it was a drag on the whole company.
The one thing they didn’t apparently do throughout the layoffs and reorgs was push development of the G4 processor. It was during the Ruiz era that Motorola got “stuck” trying to ramp up the G4, and Apple fell far behind the PC world. It should also be noted that Motorola always charged Apple at least $200 for a CPU, and usually rather more. Pretty high prices given the relative lack of performance.
Ruiz and Rivet left in 2000, so they can’t be blamed for everything that happened afterwards, but Apple eventually dumped Motorola in favor of IBM, who had been coming up with better Mac processors than Motorola as a second-source and actually seemed to give a damn about it.
Motorola was essentially the Intel of the Mac world, and they blew it to IBM (which was the AMD of the Mac world). Trying to do the same thing when you’re second-fiddle is only likely to work worse than it did the first time.
Given what AMD has been trying to do pricewise the last year or so, one could reasonably suspect that a few people may be trying to recreate the Motorola experience.
Uhhh, maybe the best way to take a near-monopoly away from a competitor isn’t hiring someone who had one and at least began the process of blowing it. Even if it is (in the sense of learning what not to do), at least from the outside, it doesn’t look like it was much of a learning experience.
Right now, the only conclusions you can make is that either there is something terribly wrong going on in Dresden, or there is something terribly wrong going on in the heads of some AMD head honchos.
If it is the latter, how long will it be before one Motorolee realizes that he isn’t at Motorola anymore with the luxury of a near-monopoly, and how much long-term damage could that cause AMD?