Replies to post #43806 on Advanced Micro Devices Inc (AMD)

[BUGGI1000]

@Chipguy
"
a substantial cost disadvantage
"

I don't know, what you want to say here ... seems to me,
that you want to play the old Jerry style speak like
DIE size DIE size ... Thats an important fact and we have
to take care about the development, but with the fact that
AMD is much much more demand constraint then factory
(capacity) constrained, it seems to me not a big issue.
Again, I would like to see a smaller DIE for the A64, which
will come now with 90nm NC step (85mm^2), which is way
smaller than the prescott!
So, lets make a comparison with equivalent process tech-
nolgy. Do you still want to say, that AMD's 90nm design
is much more expensive than that what Intel is now producing
and will produce, when they switch to 2MB cores?

BUGGI

[j3pflynn]

chipguy - re:"Anyway you are sidestepping the critical point that
with Athlon vs Willamette AMD had a substantial cost
advantage while with A64 vs Prescott it is probably
at a substantial cost disadvantage. The dreamer's
talk of inflicting material "damage" on Intel is
laughable."

Except this time AMD has a far better performance lead foundation, has won a lot of mindshare and interest in their product, and is about to bring out it's 90nm process descendants out in force. Last time around, they had blown the 130nm transition with TbredA, and were at a power disadvantage. Not to mention that last time, their only real competitive slice of the pie was the desktop. Quite different landscape now.

With 90nm, their potential production increase is a little higher, and they already have a power advantage to go along with their performance advantage, as well as a more effective and easier to produce supporting infrastructure.

[pgerassi]

Chipguy:

Intel has that they will only get to 3.8GHz by Q1 2005. 6 months to just a 5.5% increase. A 3.7GHz at 1066MHz QDR (14 mult) will not get the big boost over 3.8GHz at 800MHZ QDR (19 mult). If we see 5%, I will be surprised. So 10% in 6 months and 15% total in 12 months going by the schedule, is not like the Athlon Williamette era. Its more like the Athlon P3 era where P3 reached the end of its increases for a year and where the next version (Banias) was a few years off. Except now they had to cancel their Tejas follow on, so its like Athlon vs P3 without the P4 available at all.

Besides the inflection point is where Intel confirmed it must follow AMD. The first time was with DDR. Next was that it had to go dual core with Itanium (it hit the roadblock sooner with a 6.6% speed up taking 15 months). Now the third strike was with AMD64. The fourth one is that due to the roadblock encountered, everything has to go dual core. AMD still has more to go with single core speed ups (17% speed up in less than a year), but it has dual core to turbo boost its marketshare. Intel has to do it to stay in the running.

The inflection also occurs in the highest GM market, server CPUs. The GMs here are in the triple digits and produce enough profits that Intel is willing to spend billions of dollars on Itnanium to stay in the game. Of course some of that had to be put back into Xeon to attempt to stave off the Opteron stampede (Intel's own words). This is a place where market share gains or losses drop readily to the bottom line. This in conjunction with workstations is where Intel makes most of its profits.

Although it can afford to lose the entire market and still be a profitable concern, its GM, earnings, stock price and reputation depend on maintaining its monopoly here. Any movement here to AMD can cement its future. Getting 30% of this market would push its CPU GMs towards the 60% Intel used to enjoy. Getting 50% of this market would put overall GMs into this range even with substantial Spansion revenues and allow AMD to earn more than $1 billion per quarter. YOu know that would shoot the stock price into outer space.

Pete

[yourbankruptcy]

Anyway you are sidestepping the critical point that
with Athlon vs Willamette AMD had a substantial cost
advantage while with A64 vs Prescott it is probably
at a substantial cost disadvantage.

Quite right. This is why the sub-$150 market is shot close for A64, they just can't sell there at profit. And this is why AMD is forced to sell some of A64's into sup-$500 market as Opterons. In normal conditions they wouldn't have to do that.

The dreamer's talk of inflicting material "damage" on Intel is laughable.

I think we just discussed this week that Intel officially announced that price war is over. Please scroll down if you missed that. As AMD has absolutely no power to sell at volume in low-cost market, Intel's market share is protected firmly and there is no matherial damage.

In opposite, AMD is probably very happy to see that Intel is starting to sell some P4's into the $600-$700 area. That must give them breath, too.