From the INTC thread
For a while there I was of the opinion that AMD is giving Intel a serious run for its money. And they were, at least for a while. When I bought a few hundred additional shares of INTC at $35 1/2 a week or two ago I did it after having to struggle with myself whether this was the right decision. My valuation of the company based on fairly modest future projections was telling me I was doing the right thing, while my brain was screaming NOOOOOO....
But enough BS, on to the facts. Today we learn that AMD has been unable to crack the server market despite their supposedly "award winning" product:
public.wsj.com
However, most of the substantive news came yesterday. At first sight the headline seemed to promote AMD's newfound success:
AMD Prepares Chip Onslaught On Intel
zdii.com
First we begin with the mandatory hype of a product that does not yet exist except as a "spec":
The Sunnyvale Calif., chip maker on Monday revealed plans to launch a number of chips including ClawHammer, a new processor that will yield clock speeds of 2GHz and higher. ClawHammer will mark an important milestone for AMD, as it will be the first of the chip maker's forthcoming Hammer-family of processors, due in 2002.
Yada yada yada, the Hammer is the best thing since sliced bread. But keep reading my fellow Intel longs, the real treasures are hidden toward the end of the article:
Despite AMD's high expectations for it, the Mustang core has seen a few setbacks. Most recently, AMD scrapped plans to offer large cache-size server chips based on the new core, which can offer up to 1MB of integrated cache.
What happened to the notoriously easy to manufacture Athlon design (at least according to TMF Oak)? What about the vaunted SMIFs in Dresden? Those were supposed to forever solve AMD's cache yield problems. Yeah, I know that the official line is that the box makers did not want the new product, but that is like saying that the Republicans or Democrats give a rats ass about the "will of the people" in Flordia. But wait, the news gets even better (for INTC longs):
Palomino is also the next big thing for desktop PC buyers. The Mustang derivative is due in the first quarter of next year. The chip is slated to ship at 1.2GHz and 1.33GHz in the first quarter of 2001. AMD chairman and CEO Jerry Sanders has predicted that the chip would sample at 1.5GHz in January. However, AMD has now backed off that claim. Company officials said Monday that the projection was simply too aggressive. AMD now plans for Palomino to hit 1.33GHz in the first quarter and ship 1.5GHz in the second quarter of next year. A 1.7GHz Palomino is scheduled for the second half of 2001.
No, Jerry can't be up to his old tricks again, can he? Over-promise and under-deliver? Naaah, this is the old Jerry. The new Jerry delivers only Pentium crushers. Well, maybe next time around. Apparently we will have to wait for Hammer if Jerry's prediction is to come true. But wait, THE most important news is yet to come:
AMD plans to transition Palomino and Morgan from its current 0.18-micron process to a new 0.13-micron process in the first half of 2002. The resulting chips are code-named Thoroughbred and Appaloosa, respectively.
Appaloosa, Lallappalloosa (sp?). May I remind those of you not up to date on their Intel news that Intel will begin shipping 0.13um product in volume by mid 20001, and perhaps as early as the end of Q1. From what I hear the initial tests of their 0.13um process yielded and performed well.
If there is anything to worry about now is Intel's execution on those networking and communication fronts. Has anyone heard anything new on this front? |
