Michael Dell is still young and kicking
I once criticized AMD for their small vision, 30% of the market by 2008 seemed so insignificant, and at the same time I praised Dell for their attitude -- DELL folks always talk in up beat tones.
Now, we know DELL signed up with AMD when AMD made sure it could supply 10 million CPUs. While I question Michael Dell's PC mentality, I have to say he does have the ability to push volumes. Making millions of $399 PCes is a major supply chain management problem, and DELL knows how to do it.
AMD will exit 2006 with 40% market share (run rate), mark my word.
posted by Sharikou, Ph. D. @ 9/08/2006 01:46:00 PM
19 Comments:
For those of us not in the know, could you please define market share (run rate), since it is not a common metric (like plain, old market share)?
Also, how do you reconcile your prediction with Ruiz' comments to Wall Street analysts that AMD would be at 40% by end 2008? (IIRC) I certainly understand the concept of sandbagging a target, but it would cost Ruiz some credibility on the Street if he sandbagged by 2 years. That would not be viewed as a good thing when the CEO has that little visibility 2 quarters out. No one (well, maybe Intel) wants to see that happen.
A run rate of 40% for the 4th quarter would mean that if sales continued the same way over the next year the market share would be 40%.
I don't see anything wrong with Ruiz being conservative in his marketshare estimates. At the very least it could lull Intel into a false sense of security.
In simple english what this should mean is at the end of qtr 4 AMD's market share on that day should be 40%. Hence they "exit" 2006 with 40% share.
Sharikou - be clear about whether you are talking about share by volume or share by revenue. These are two different things.I see you lining up wiggle room for yourself as usual. Perhaps you learnt your lesson after missing the Q2 Intel GAAP loss prediction.
AMD will be suffering lower & lower margins from the price war in the next 2-3 qtrs and will move to a GAAP loss once they have to start writing off interest on their loan to acquire ATI. They have no choice to keep reducing prices as long as Intel does because they will have to keep their factory filled. Remember - a good qtr for AMD is 90 mln $ profit and a bad qtr for Intel was 800 mln $ profit. It's easier to go to 0 from 90 than from 800.
http://sharikou180.blogspot.com/
A more balanced point of view.
Sharikou - please define explicitly what you mean by market share runrate (in matehmatical terms / formula).
A good example for your readers may be to calculate what AMD's market share (runrate) was exiting Q2'06.
Thanks in advance
A run rate of 40% for the 4th quarter would mean that if sales continued the same way over the next year the market share would be 40%.
In other words, not a 40% market share, but rather an extrapolation of the market share based market share growth rate (which is defined as d(market share)/dtime). Any statistician will tell you- extrapolation is a dangerous business. And I'll tell you that there are 100s of potential external factors (ranging from Intel pricing to global macroeconomic events like war in Iran or a successful terrorist attack) that can trash that extrapolation in no time at all. Past performance is not necessarily a predictor of future performance...
Sharikou: "AMD will exit 2006 with 40% market share (run rate), mark my word"
Can u please explain here once and for all what exactly do you mean by "run rate".
"I once criticized AMD for their small vision..."
translation - I hate Dell because they don't use my beloved AMD.
"...and at the same time I praised Dell for their attitude..."
translation - I love Dell because they are now using my beloved AMD.
41 is not young and kicking!
http://www.msdf.org/images/biophotos/dell_michael.jpg
*barf*
This is not made to offend anyone but Michael:)
To the glass ½ empty crowd. The main point is AMD market share is trending UP and Intel’s market share, employee count, bank account and image is trending down.
Consider a daily diet of bad news for Intel advocates over the past year and you’ll find there’re posts are trending down also.
Intel is already bankrupt when comes to ideals for saving this once proud and capable corporation. The 40% number doesn’t mean much because before long the number will be 50% and then 60% and so on. The AMD world wide snowball is rolling down hill.
“Intel is already bankrupt when comes to ideals”
INTEL press release yesterday;
CSI “Common Systems Interface” to link CPU just like rival AMD. Set for release 2008. This most certainly proves Intel is 2 or more years behind.
9.11=Doomsday for AMD. THG will release Kentsfield benchmarks:)
AMD has shown nothing but closed demos of 4x4.
http://www.xtremesystems.org/forums
They've been roaming around in overclockers's hands for a while, now official results.
Just some perspective on performance..
"If so, thats some scary performances. Quite literally twice as fast. Makes sense i guess"
"conroe@ 4ghz has worse cinebench than kentsfield at stock[2.4ghz]...someone at amd just passed out lol"
On a TMPGEnc test, a 2.4GHZ Kentsfield outperforms a 3GHZ Conroe.
http://www.tgdaily.com/2006/09/09/preview_kentsfield_ processor/
"The quad-core was sent through the entire test parcours and showed impressive performance."
"Kentsfield easily shattered previous benchmarks records and highlighted its horsepower especially in threaded applications such as audio and video processing. The chip was able to set new record levels in traditional benchmarks such as 3DMark 05 and 3DMark 06 as well. Overall, Kentsfield turned out to be about twice as fast as the Pentium EE 965, while consuming about the same power."
"Tom's Hardware will publish the full article with technical details and the complete benchmark result on Monday, 11 September."
"This most certainly proves Intel is 2 or more years behind."
You can say the same about AMD in the notebook platform/chipset area. Core 2 is already beating K8 without MC/HT, the lead will extend when they do implement them
CSI “Common Systems Interface” to link CPU just like rival AMD. Set for release 2008. This most certainly proves Intel is 2 or more years behind.
I have long proved that Intel is 5 years behind AMD.
"INTEL press release yesterday;
CSI “Common Systems Interface” to link CPU just like rival AMD. Set for release 2008. This most certainly proves Intel is 2 or more years behind. "
What ..did you forget the Integrated Memory Controller ..that Intel is planning for 2008.
Intel's position reminds me of Abhimanyu caught in the Chakravyuh of the Indian legend of Mahabharat.
Chakravyuh is a battle formation that is impregnable.. Young Abhimanyu was taught only the secret to enter the Chakravyuh
but not to come out of it, and is eventually killed by the enemies who lay the Chakravyuh.
"INTEL press release yesterday;
CSI “Common Systems Interface” to link CPU just like rival AMD. Set for release 2008. This most certainly proves Intel is 2 or more years behind."
I told a friend of mine who works at Intel a few months ago that Intel must've been working day and night to get CSI out as early as possible. At that time everyone was saying CSI would not be out in 2009.
He was surprised at where I got this info. I guess Intel wanted to guard it carefully before it's really ready. Well, it's just common sense, that there is no where for Core 2 to go forward have it not a scalable interconnect like the hypertransport.
Putting CSI onto Core 2, however, would be yet another punch into Itanium's face. Not that many people still care about it, though. ;-)
"9.11=Doomsday for AMD. THG will release Kentsfield benchmarks:)"
Kentsfield is just two Conroe in one package, sharing one FSB.
If any benchmark run twice as fast on Kentsfield as on Conroe, that benchmark must've been able to fit completely into Conroe's L2 cache. In other words, that benchmark do not reflect real-world application performance.
Intel fanboys used to call 4x4 being just a dual opteron. Then how much different is Kentsfield from dual Conroe? The fact is, 4x4 is targeted at gaming with its special HT links among CPUs, memory modules, and graphic cards. OTOH, Kentsfield is indeed nothing but two Conroes packed into one MCM.
Anyone touting Kentsfield as technological advance, other than that it's an MCM with huge die area, is (paid or unpaid) a marketing pumper. (Yes, THG).
"CSI “Common Systems Interface” to link CPU just like rival AMD. Set for release 2008. This most certainly proves Intel is 2 or more years behind."
Yet somehow for 2P or less systems latest Intel chips are equiv/better performance. Of course that will change when AMD next gen chips come out, but to say 2+ years behind - are you comapring some specific architectural feature or actual performance?
I could (idiotically) say, AMD was at least 1-2 years due to delayed adoption of DDR2, but as that had virtually no major impact on AMD performance on Athlon vs P4 that makes as much sense as your comment above. (Unless you are referring to 4P+ systems)
Sharikou - why are you avoiding specifically defining what tou mean by market share (runrate)?
It would be good for you to do this before the time your prediction is evaluated.
Please define for your readers, again a helpful tool woud be to calculate this for Q2'06.
Also, I assume you mean OVERALL X86 market (server, desktop, mobile)?
And also not some crazy stuff like retail only?
"Core 2 is already beating K8 without MC/HT, the lead will extend when they do implement them "
AMD will be pushing K8 microarchtecture into Geodes by that timeframe.
Post a Comment
<< Home