InvestorsHub Logo

The Duke of URL

05/30/06 10:33 AM

#28385 RE: Dan3 #28384

..."Don't know what you think, but my personal opinion is that AMD will have a lot of price adjustments to make in order to become competitive again. Intel theoretically sounds as an unbeatable deal right now. You can expect that "Energy Efficient" marchitecture will probably send the current AMD line-up into oblivion (or just equalise the pricing), since those CPUs right now are even more uncompetitive when it comes to pricing."...

http://www.theinquirer.net/?article=32026

Tenchu

05/30/06 1:50 PM

#28388 RE: Dan3 #28384

Dan, Still - this lets AMD go for 40% to 50% of the market while keeping costs and number of employees about where they are today.

AMD isn't even at 30% now. Don't count your chickens before they hatch.

Tenchu

wbmw

05/30/06 2:51 PM

#28390 RE: Dan3 #28384

Re: Moving to 45K 12" 65nm WSPW from 30K 8" 90nm WSPSW represents growth to about 7 times their current capacity - and they're supplying more than 20% of the market now. Next year's chips will need a lot more transistors than this year's, and the after even more, plus the market is growing (let's hope!). Still - this lets AMD go for 40% to 50% of the market while keeping costs and number of employees about where they are today.

First, let's get past the fact that you've been pushing this line for years: #reply-18472987 #reply-16987841

Also, you seem to be describing transistor capacity. Yeah, moving from 8" to 12" wafers, shrinking a process node, and increasing throughput by 50% nets a 7x increase, but it's absurd to claim that this will also mean 7x capacity.

For one thing, the shift to dual core will demand larger die sizes than we have today, not smaller ones. AMD cannot shrink their products without improving them, and still remain competitive. All signs suggest that they will aim to match Conroe performance, even if they have to brute force it with larger designs with bigger caches and more cores.

Secondly, the market is physically larger by the end of 2008. Even if you assume high single digit growth rates (say, 8%), the TAM at that time will be >25% larger than it was at the end of last year. And when you add in the push for emerging market penetration, that number can be a lot larger.

Based on that, it would be difficult for AMD to get to 40% MSS, even if they could sell all the chips they can produce. And of course, that's the other problem, since Intel now has a competitive micro-architecture. Much better than what you had assumed in the past:

"Intel can talk about how much it spends on its FABs until the cows come home, but it can't seem to ship a process that results in high performance parts that have reasonable power budgets."
#msg-7297399

"You never learn, do you? Next year, when Merom and Woodcrest get "upgraded" in power use, you'll be waxing poetic about whatever snake oil Intel's peddling for the following year."
#msg-7471800

And by the way, let's review the launch frequencies and power levels of Intel's Core micro-architecture:

Merom - 2.33GHz @ 35W
Conroe - 2.66GHz @ 65W
Woodcrest - 3.0GHz @ 80W

All are 64-bit processors with reasonable power levels and performance that leaves the K8 core in the dust (and let me reiterate that these are just the launch frequencies!). Compared to AMD's 90nm parts that you used to think was better than Intel's 65nm:

Turion X2 - 2.0GHz (small cache only) @ 35W
Athlon X2 high efficiency - 2.4GHz @ 65W
Athlon X2 normal - 2.6GHz @ 89W
Athlon FX top bin - 2.8GHz @ 125W

Time to eat crow, Dan.