A very quick note. There is some wild gesticulation from the AMD "fannies" that AMD's CPU margins and cost structures are better than Intel's based on some ridiculous calculations (which I have already pointed out are baseless) from my good friend the Doctor. Here's why this is inaccurate and just a bunch of hand-waving from the AMD fans:
1. In Q3, Intel and AMD's gross margins were almost identical. However, AMD's margins are only for the high margin business of CPUs while Intel's are for CPUs + other businesses with significantly lower margins like flash, chipsets, mobos, etc. Furthermore, businesses like flash are actually a drag on profits recording a 116 million $ loss in Q3. On a CPU revenue of 5.8 billion out of a total 8.7 billion Intel's CPU margins would be significantly higher than the ~50% it recorded.
2. If AMD's margins are so good, why is it that it's net income for Q3 is only 1/10th of Intel's in spite of the fact that it's CPU revenues and market share are roughly 1/4th Intel's. And this is after Intel's net income being reduced by ~740 million due to Flash and Other losses.
The answer to both these is simple. Not only is Intel's cost structure better, it's ASPs are higher in the CPU business.
Subscribe to:
Post Comments (Atom)
10 comments:
Check out some of this ROFLCOPTER stuff:
http://forumz.tomshardware.com/hardware/Journal-Pervasive-64-bit-Computing-ROFL-ftopict207692.html
I'm wondering who thinks that AMD has better ASP and margins. With 90nm 200mm production it is not possible.
It will have better profitability in two quarters introducing K8L, having FAB36 converted to 65nm with 300mm wafers. This time AMD is going to have better ASP and more money. Q3 2007.
Mikolaj said...
..."It will have better profitability in two quarters introducing K8L, having FAB36 converted to 65nm with 300mm wafers. This time AMD is going to have better ASP and more money. Q3 2007. "
--- 0 ----
better than whom? Certainly not intel who's transitioning to 45nm Q1'07. K8L isn't even going to be competetive to demand such a high ASP. Not to metion volume limited.
Actually, your assumptions are incorrect.
First of all, you don't seem to bother asking why Intel's margins were fine in Q4 05 even though it still had all of those margin dragging side investments.
I noticed too that in trying to build this logical house of cards you are quick to point out these side investments for Intel but blindly overlook that fact that AMD produced its margins with an older 200mm FAB running at 90nm against Intel's much lower cost 300mm FABs producing 65nm. AMD's costs will steadily improve all during 2007 and unfortunately this drop in costs will be at least double what Intel's is. Your margin arguments are nonsense. Even if Intel fully divests of everything including flash memory it will still be at a disadvantage in costs on the same process at the same FAB size. I assume this is why Intel fans keep talking about Intel's switching to 45nm in late 2007. However, Intel won't have any serious production of 45nm until 2008.
BTW, Intel's ASP is higher than AMD's. However this has been narrowing and should continue to narrow as AMD moves more into mobile.
Scientia:
1. I know why Intel's OVERALL margins were fine in 05. Because their factories were running at higher capacity and they had a higher microprocessor mix in their sales. However, that is irrelevant to the high level point I'm making here which is purely on CPU margins as they are today. The implication is that the wild calculation Doctor Sharikou cooked up on how much it costs Intel and AMD to make a CPU respectively is wrong. Nobody outside a few people in those companies know the true number and the folks who are blindly believing him need to learn to think for themselves.
2. Second - I don't care if AMD is on 200nm. Let me repeat - I'm talking about CPU margins today. Being a process generation behind is a cause for lower margins but that's life. You're missing the point I've made in this post in your defensiveness. Intel is not on the same lithography as AMD and that's their competitive advantage = margin upside.
3. Your ASP comment is a fair reflection of the marketplace. Intel's ASP is indeed on a declining trend while AMD's is increasing. I suspect both went south in Q2/Q3 due to the price war but in general Intel's overall ASP is heading downward while AMD's is improving.
Finally - I've asked you a few times before to please substantiate your POV with logic or facts. You're more than welcome to call my arguments nonsense if you can disprove them. But so far all you've done is be rude except for your ASP comment.
I've now pointed out three of your articles that have been bogus: the AMD layoffs story, the $500 notebook story, and the channel partners fleeing AMD story. As for being rude:
There is some wild gesticulation from the AMD "fannies" that AMD's CPU margins and cost structures are better than Intel's based on some ridiculous calculations (which I have already pointed out are baseless) from my good friend the Doctor.
I don't recall that I've ever gotten any financial information from Sharikou. He tends to cherry pick stories just like you do even when they are wrong just like you do. And, like you, he never admits that any of these stories were wrong.
Now, your point is that Intel's margins would be better if they fully divested of all of their side investments including flash memory; this is absolutely true.
However, this point is countered by the fact that AMD's margins will be better on 300mm with 65nm production.
Now, why is your point invalid without the AMD point? Yes, theoretically if Intel were divested today it would have better margins, better than AMD's. However, it is impossible for Intel to divest overnight; this process takes time. Assuming that Intel got serious about divesting today it would still take awhile for this to take effect. During this time, AMD will shift to lower cost production.
Now, the important point is that Intel has not shown any indication of fully divesting and I doubt they will before early 2008. However, AMD is definitely ramping FAB 36 and will definitely begin scaling back FAB 30 in Q2 07. AMD's plans are definite as will be its lower costs; your talk about Intel is more wishful thinking.
I've now pointed out three of your articles that have been bogus: the AMD layoffs story, the $500 notebook story, and the channel partners fleeing AMD story. As for being rude:
http://sharikou180.blogspot.com/2006/11/doctor-is-making-money-of-intel.html
There's your response.
Hey, I posted a comment on your blog with things you haven't backed up or explained(Intel's great Jan06 CPU, Intelliots that say 'this then 'that', why AMD won't forecast, why you're ranting about Tigerton with "just 266" compared to Tulsa 200), yadayadayada. Still haven't heard.
However, this point is countered by the fact that AMD's margins will be better on 300mm with 65nm production.
They're continuing to introduce new 90 nm models in Q307. Somehow I don't think that they'll be 'ramped' by mid07 enough to have an impact on margins, as they could still be hit on yields(2.9 max on 65 nm?)
http://sharikou180.blogspot.com/2006/11/whose-cpu-margins-are-better.html
There's your response.
Scientia:
The Doctor has recently reiterated that his earlier calculation that Intel's CPU manufacturing costs are higher than AMD's and that they make a loss when they sell a part below 60$ while AMD's mfg cost is ~40$.
This post is to poke a hole in this outlandish theory. The reality is the cost per CPU is something very few people actually know and those people work for AMD & Intel respectively. Read nothing more into this post.
Second - pls note the title of the post says "CPU margins". Not margins...
Ashenman - chipset business does not have the same kind of margins that CPUs do. But you're comment on what platforms like Viiv and VPro are intended to do...increase Intel's share of system level revenue. The sum of those discrete components become greater than the whole in terms of margins only if they are able to build those into brands like Centrino. Otherwise, it's a commodity business based on whether someone actually has a product differentiator and price. At which point Intel has no signficant advantage once AMD offers equivalent solutions. Simple fundamentals of business strategy at work here.
Couple of things:
The intent of this post was to point out that Intel's CPU margins are better than AMD's. That means excluding the chipsets and other businesses. From a gross margin perspective AMD is at a strong advantage right now wrt the context of my comment for this post having dumped their flash business. We're not talking profit here...gross margins which means % of profit after removing cost of sales. The advantage of platforms to Intel is a higher share of revenue and hence more profit. But this drags down margins because inside the platforms they are selling more components with lower gross margins.
What I was really trying to do is debunk Sharikou's completely fabricated formula to calculate Intel's cost per CPU as well as AMD's. Fact is - this is a closely guarded secret and outside those companies nobody...not even Wall Street with access to more information thatn you or I know what that number is. We can only guess.
Chipsets are a significantly lower volume than CPUs. Intel probably does about 30-40 million chipsets per year.
Post a Comment