Thursday, September 28, 2006

AMD won Dell 2 years too late - I'm not alone

I have mentioned in several places that I think AMD winning Dell would have had far more significant impact 2 years ago. Today Dell is struggling to find direction, they no longer hold a significant price advantage over their competitors and are still focussing on squeezing every ounce of cost from their supply chain and suppliers and this is exactly what they are using AMD for. To deliver ultra cheap systems. Unfortunately, in the process they are probably hammering AMD's margins. Meanwhile Intel has moved to a uniform pricing on desktop parts for all their customers. What this means is everyone get's the same price...from Dell...to HP...to their distribution channel. So Dell is unable to squeeze Intel and as a result they will suck AMD dry even more. AMD can't say no because they have been waiting to get entry to Dell for donkey's years.

My opinion is borne out by ThinkEquity analyst Eric Ross. Ross of course has access to more information than I do so he's got more facts to substantiate the argument:
http://biz.yahoo.com/seekingalpha/060928/17679_id.html?.v=1

Bottom line...Dell is going to be a drag on AMD's margins. AMD going to Dell will be counter-balanced by HP getting more comfy with Intel. You can read more about it at my earlier post:
http://sharikou180.blogspot.com/2006/09/new-bed-fellows.html

45 comments:

Randy said...

2 years too late... patience is a virtue after all.

Why would it have been better for AMD to go bargain basement, for the past two years they have been selling their chips at a nice premium due to their obvious superiority. If you have the best chip, you don't sell it to Dell and let them haggle your profits away.

At the same token, if your in a price war, you act like it. You can't win a price war by keeping high margins, such is the definition of a price war. Now is the time to go for market share, the best way to do that is with Dell. AMD is behaving like a Co. that feels it can sustain a price war longer than its competition(the deal with Dell, buying ATI all the while Intel lays people off). Whether or not that's the case remains to be seen, but that's what it looks like from the outside.

By the way, AMD has tried to one-up IDF again with a little help from MS.

Scientia from AMDZone said...

It's a nice theory but it doesn't make any sense. If AMD wanted to move more bargain priced chips they could do it through eMachines and Tiger Direct. They really would have no need of selling discount desktop processors to Dell. I guess if your theory is correct then AMD should see a big drop in revenue in the 4th quarter. As I've said, I'm certain that Intel's revenue, volume share, and revenue share will all be down compared to 2005 whereas I expect AMD's to be up in the same categories.

Anonymous said...

Say old chap. Why dont you make an entry of the latest news on that Sony battery scandal. Seems Toshiba also had to return a few hundred thousand becuase they sucked. That must really put a huge spike in Shirakous cofine.


Keep it real man
/MOrK

Scientia from AMDZone said...

Remember that Intel gets half of its income from the Asian market. It looks like AMD is pushing into this market now.

http://www.cio.com/blog_view.html?CID=25293

BTW, have you considered turning off the word verification? It's off on my blog and I think it's much easier. I still have comment moderation on but not word verification.

I put a counter on my site at the bottom of the page; I got one from bravenet.com. I've had 582 visitors to my blog since I added the counter. I've had 26 visitors today. I really can't tell otherwise how many people look at it because most people don't post comments. It might be because my articles to be somewhat involved.

I post here because, while I think that some of your assumptions are wrong, I don't think you actually have an unreasonable perspective. I've asked Sharikou several times to explain his Intel bankruptcy theory and he just ignores my posts.

Scientia from AMDZone said...

BTW, have you seen your listing at rubyworks?

blogs

Courtesy of Mad Mod Mike.

The Sheepshagger said...

"It's a nice theory but it doesn't make any sense. If AMD wanted to move more bargain priced chips they could do it through eMachines and Tiger Direct."

Let me correct one mistatement: AMD doesn't move bargain priced or any priced chips when the want, nor does Intel. What moves demand is what end users desire. Needless to say, AMD could shovel all the bargain chips they want for e-Machines or to Tiger Direct but if nobody is buying there why would either company build an invetory of chips that aren't in demand? Seriously... e-Machines and Tiger Direct will order all kinds of chips when it matches the demand.

I have more to say on 180's theory that actually backs it up but it will wait until I get more time.

BTW... I had already left a message for Sharikou about Toshiba's announcement this morning. What is even better is Lenovo announced a recall yesterday, too. Naturally, I had to share that with the lad as well. I can't wait for his side of the story now that 4 OEM's (Dell, Apple, Lenovo, and Toshiba) have recalled products with Sony batteries. My prediction is he'll point out the obvious that Intel was to blame and that Sony must have drinking their own bong water when they mistakingly posted a recall notice accepting they had a manufacturing issue. What sayeth Ye?

180 Sharikou said...

Scientia - thanks for the pointer to MMM's listing. I've turned off the word verification - let's see how it works.

WRT Asia - AMD has been attacking emerging markets as a focus for two years now. Their key identified areas of focus are China, India, Latin Amercia...to name a few. It's not that they are giving up other business but they are pushing hard into these markets. AMD's biggest success was winning Lenovo in their bid to penetrate deeper into the smaller tier markets in PRC. Not winning this was a big miss on Intel's part and Lenovo has been the principle reason for their share loss in PRC...a market they were in before Lenovo was really a sizable force. Founder is going AMD because Lenovo has successfully used AMD and penetrated deep into China. Now Founder is hoping to replicate and compete. Don't be surprised if other's follow. But the reality is in PRC, Lenovo is king with the end user. They represent the might and the power of a Chinese company that has become a global brand and symbol of what China can achieve.

On the Dell issue - when you're neogtiating pricing you don't always get to say no. You gotta have give and take. Unfortunately Dell as technology company in my mind is basically bean-counter with a screwdriver, I'm inclined to believe there's more give and less take for AMD. Here's what I really believe. AMD is literally giving away the low end stuff like Sempron/Athlon 64. They are also getting hit on margins more than they'd like on the lower bin X2's like the 3600/3800. However, they are keeping their margins robust on the higher bin X2 SKUs in the hope this will balance out. My problem is Dell seems intent on positioning AMD on price...not technology. look at their web site - when they talk about AMD it's messages are affordability, robust, etc. Contrast this to how they talk about Core 2 - "lightning strikes twice", breakthrough performance. Dell is using AMD to try and service the low end of the market which is not a good place to be with Dell. This was exactly what resulted in their first ever qtrly miss on targets last year. They are ruthless in squeezing their suppliers and after year's of being pampered by Intel, they will expect even more from AMD. They are fully aware of the symbolism of giving AMD their business and will take every opportunity to remind AMD.

The entire idea of this blog is to spark debate so I do appreciate you coming in here with a different but I think balanced POV. At this point in time, I believe Intel is heading to an upswing at AMD's cost and you are seeing that reflected in my thinking. Who knows...6-9 months from now I may be calling it differently.

Anonymous said...

This blog is not as good as Sharrikou's "Pervasively Gay 64 Bit Computing".

AMD 'winning' Dell. Talk about undertones. Overtones. Innuendos.

Hector steals Michael from Pat.

What happens next?

Tune into Sharrikou's "Pervasively Gay 64 bit Computing" for the next exciting episode of "Desperate Chip Titans".

Accept no imitations. Only industry insider Sharrikou flits from bed to bed, from city to city, to get the insider knowledge that will quench your lust.

The Sheepshagger said...

The thing with emerging markets is there is huge growth potential for both Intel and AMD. The number of chips sold in Asia is growing at a staggering speed. I don't have access to the year-over-year percentage but for the time being and in the years to come both chip makers should be doing well.

Three years ago Intel had a 90% market share in China and AMD only had a 5% share. The current year projections show Intel with nearly a 3/4 share of the market and AMD growing to 18% of the remainder which despite being in the Intel camp, I can't help but applaud AMD making those sorts of in-roads.

The bottom line is the growth rate and demand in China alone is such that even if AMD gets 30% of the market share in China, Intel will still be selling more chips in China than in the previous year. Or to phrase it differently, despite the potential of losing market share in China, Intel won't be selling fewer chips in that market by any means.

Mathematically speaking, let’s say the Chinese market was 1 million units in 2003. At 90%, 900k would be Intel chips and 50k would be AMD. With a 20% year over year increase in demand over 3 years, the new demand in China is now 1.728 million units. At ~75%, Intel would move 1.296 million chips and AMD at 18% would move 311k chips. Clearly the market segment share and growth is in AMD's favor, but Intel would still have sold ~300k chips more during 2006 than during 2003.

Obviously the example above was just for ease of numbers for proving a point. The danger Intel faces is losing market segment share in 'mature' or chip saturated geographies like the Americas where the growth rate is only single digits.

As an example let's say the US demand is 1 million and year over year unit demand is only 5%. Assuming the same percentages of segment share as China (which is obviously going to be factious) where in 2003 Intel would have 90% and would move 900k chips while AMD would have 5% and would move 50k chips.

With a 5% growth rate through 2006 the new demand is only 1.158 million chips. With Intel declining to 75%, they would only ship 870k chips (roughly 30k fewer). AMD's growth to 18% would mean 208K chips sold.

I don't want to take away from Intel or AMD's successes but I do feel that certain people (the original Sharikou and not 180) don't have a grasp on these sorts of things. While much is to be made about superior processors in the previous 3-4 years by AMD, this isn’t really the sole reason AMD's worldwide acceptance has come about. People often overlook Intel's Q4 financial report during 2002 or 2003 where then CEO Craig Barrett announced that Intel had underestimated the market. While there was a significant downturn in the US in the years following the corporate refresh prior to and immediately following Y2K, coupled with the dot com burst, Intel in the 4th quarter of that year alone left $1 Billion in unfilled orders on the table and it would be long before AMD would successfully capitalize on it. If you want my opinion that is what forced the OEM's to take a look at Opteron.

Again... I can't say anything bad about the chip. I believe AMD had the best chip on the market and they should have been able to make money hand over fist. AMD went about selling the Opteron cheaper than what it probably could and should have commanded for the chip but they wisely chose to sell the chip at a reasonable price so that could keep their foot in the door and pull their way into the market soundly.

What is ironic about AMD's lawsuit for unfair practices against Intel is that they are suing for money lost when Intel practically gave AMD the keys to the kingdom with their misstep.

Okay... I've droned on long enough. I had a few other thoughts but I shall save them for another visit. Let me know what you think and have a great weekend!

The Sheepshagger said...

PS... I like the lack of verification word.

Anonymous said...

"BTW... I had already left a message for Sharikou about Toshiba's announcement this morning. What is even better is Lenovo announced a recall yesterday, too. Naturally, I had to share that with the lad as well. I can't wait for his side of the story now that 4 OEM's (Dell, Apple, Lenovo, and Toshiba) have recalled products with Sony batteries. My prediction is he'll point out the obvious that Intel was to blame and that Sony must have drinking their own bong water when they mistakingly posted a recall notice accepting they had a manufacturing issue. What sayeth Ye?"

Good lad :)

I think he will say something in lines of "Dual Core does cause dual explotions" .. even thou it aint possible from a scientic point of view. Then again that man has more issues then G.Bush...

180 Sharikou said...

Sheepshagger - your assessment is a good one. Yes - Intel made a big mis-step in 2004 when they did not produce enough chipsets. When Otellini was questioned about the inventory build up in Q2, he said it was a mistake he did not intend to repeat again. Bryant took the rap saying he advised Otellini not to build more. I think AMD was gaining traction with a superior product. This supply issue allowed them to gain momentum even faster.

The biggest problem however when you cannot supply what your customers want is they lose faith in you because they are losing business themselves. This is why AMD has been scaling back their 1MB caches parts...to maximise supply.

The Sheepshagger said...

Yes... 2004 was a chipset issue. What I was referencing was 2002 in which not enough CPU's were made to meet the demand -- hence the $1 Billion in sales left on the table and from where I think AMD really became an option for OEM's who had orders to fill.

Greg said...

I fail to see how the hammering of AMD's margins matters compared to how Intel is doing. If AMD's margins are being hammered it has to be just as bad or worse for intel, seeing as they have to be competitive too, run at a higher operating cost than AMD, are losing business, have a higher production cost per processor, and have more processors sitting in inventory.

I realize Intel has more assets it can eat into than AMD, since it's made more money longer, but it's using alot of those to upgrade all of its remaining fabs to 65 nm, and build up the resources to convert to 45 nm. This may not seem like as much of an expense as buying an entire other company (AMD/ATI) but seeing as fabs have operating costs in the Billions, and that intel has a ton of them, and you've got upgrade costs that are probably in the tens of billions of dollars that's eating into their assets as well.

I also see tighter margins as a lesser issue for AMD being that their processor sales are growing, meaning that they can make as much if not more money than they did before while operating at those margins. We should also not forget all the other businesses AMD is either breaking into or growing in, such as HP, Lenovo, and Foundation.

180 Sharikou said...

Greg - let me start with one simple fact. In a good Q2, AMD made 80 million $s. In a bad Q2 Intel made 800 million $s. Heading into red from 80 to 800 is much easier.

Please also note that I am calling a market share reversal in Q4...if not earlier. When you couple share loss with margin hit AMD has a problem. Just like Intel has had last 2 qtrs which has sunk their qtrly profits from ~2 billion to only 800 million in Q2. In the semi business, due to high fixed costs of factories, once your factories are running at full capacity almost every incremental $s of revenue drops straight into the bottom-line. Hence, the price wars are not just to gain share, they are to ensure both Intel and AMD can keep their factories at full capacity.

Now Intel had an abysmal 49% margin in Q2. But they are calling it as strengthening to 55%+ by Q4. Presumably on the strenght of Core 2 manufacturing ramping and a lower cost per unit to manufacture. Let's see what happens in Q3 but I'm pretty sure AMD will see margins drop by Q4 on lower ASP. Their only hope is 65nm ramp to stabilize. Just as Intel's is Core 2.

Scientia from AMDZone said...

Obviously, it is silly to suggest that a failed battery has something to do with the processor. Sometimes Sharikou just seems to be reaching.

Also, it makes no sense to talk about Intel profits or AMD profits; these terms are flexible and don't really tell you anything. Intel's profits depend on how much it decides to spend on stock buyback which is flexible. Also, you are adding cash to profit which is counting money that was earned in the past. This is far too sloppy. It is better to go by gross revenue and margin on just the cpu sales.

I'd say that Intel's gross revenues, volume share, and revenue share will all be worse than the same quarter in 2005. I'd say that AMD's will all be better than the same quarter in 2005. So, overall, Intel will see the worst shrinkage since 2001 while AMD will see a gain over 2005 and no shrinkage.

I don't think that AMD is being squeezed by Dell. The lowest priced AMD system uses a Sempron 3400+ which is probably the lowest chip that AMD currently produces. The additional cost for adding a higher processors to the base configurations also seems consistent with the price spread on places like NewEgg. The lowest priced unit that Dell makes that uses an X2 is $700. However, if you compare the price of this unit with the lowest priced C2D system the costs seem consistent. When I configured two systems the same with an E6300 and a X2 4200+ the AMD system was only $13 less.

180 Sharikou said...

Also, it makes no sense to talk about Intel profits or AMD profits; these terms are flexible and don't really tell you anything. Intel's profits depend on how much it decides to spend on stock buyback which is flexible. Also, you are adding cash to profit which is counting money that was earned in the past. This is far too sloppy. It is better to go by gross revenue and margin on just the cpu sales.

Profits are used to either fund future growth or provide return on investment to shareholders. Both are important to the stock. Not sure how cash is being added to profit? Pls clarify. Gross revenue and margins are important too but my point remains...going from a quarterly profit to a loss is a big deal and AMD will go there by Q207.

I'd say that Intel's gross revenues, volume share, and revenue share will all be worse than the same quarter in 2005. I'd say that AMD's will all be better than the same quarter in 2005.

Quite possibly this will be true. Specially since 2005 was a very good year for Intel and we already have 2 quarters of relatively poor performance to see. But it's not important when I'm trying to assess whether momentum in the market-place is moving back to Intel in terms of market share QoQ.

Cannot comment on your Dell statements. You need to give us a more thorough analysis like I did when I compared the system specs.

Anonymous said...

Come on Sharikou 180*... You simply must writ something about this:
http://theinquirer.net/default.aspx?article=34827

That article must really make Shakira cry,,

Scientia from AMDZone said...

Profits are used to either fund future growth or provide return on investment to shareholders.

Negative. R&D and capital expenses are subtracted BEFORE determining profit. If you want to count R&D and capital expenses you have to use gross revenue and margin as I suggested. Also, some of Intel's money is transferred to shareholders via stock buyback and not by dividends. Again, this is subtracted before determining profit. Profit tells very little.

Not sure how cash is being added to profit? Pls clarify.

Okay. Let's say you break even but you spend $1 Billion of cash. This would then show up as a $1 Billion profit. Intel offset its revenue drop in Q1 by spending more than $2 Billion cash. As I said, this isn't accurate.

Gross revenue and margins are important too but my point remains...going from a quarterly profit to a loss is a big deal and AMD will go there by Q207.

It isn't a big deal. AMD had a small loss in one quarter in 2005.

Quite possibly this will be true. Specially since 2005 was a very good year for Intel

Actually, the best year Intel has ever had.

and we already have 2 quarters of relatively poor performance to see.

Right, this will make 2006 much worse for Intel.

But it's not important when I'm trying to assess whether momentum in the market-place is moving back to Intel in terms of market share QoQ.

Yes, I see what you are trying to do but you won't get the numbers you are looking for with a simple quarter to previous quarter comparison. The 4th quarter is typically the biggest quarter so an increase in revenue in the 4th quarter does not indicate a market trend. Similarly, increases in revenue share in a falling market are not accurate indicators for AMD. There are ways to determine the actual trend but not by looking at profits and not by doing a simple quarter to previous quarter comparison.

180 Sharikou said...

I'm not talking about just revenue increase as an indicator. Pls note - I am specifically calling a market share reversal coupled with a gross margin upswing for Intel.

ashenman said...

I don't see how you can say AMD will be losing market share in q4. Yes, Intel will have core 2 and qcore 2. But qcore 2 will have an even smaller ramp than core 2 (only 1 million before quadcore opteron in 6 months?). Not only that, but only 25% of Intel's production will be core 2. AMD will have rolled out 4x4 by then, and will have its new 65 nm parts. This will be before Christmas (and seeing Intel's "amazing" ability to make its products available when they're released, I bet quad core 2 wont be).

It wont matter as much if these parts wont largely outperform core 2 parts by then, they'll have continued to increase performance, meaning they'll have kept very competitive in the portion of the market that's largest while having a performance part to toss around for pr. Intel hasn't said anything about releasing new core 2 parts that will increase performance, so they can only lower the prices by that time, which means they'll be making even less money on their primary money maker that still makes up such a small portion of the market.

Meanwhile, Intel is still producing 75% parts that are pretty much obsolete, including celerons, which have lost competitiveness due to cheaper AMD options. All of this in a season where average Joe is looking for a computer to surprise his family with, without surprising his wife too much when she goes to the bank.

If Intel were doing something to increase competitiveness that didn't only increase that competitiveness for the gamer, then maybe it wouldn't matter that AMD is consolidating its resources in areas that are selling best, while Intel diversifies into every nook and cranny without care for what can actually make money or not. OEMs are going to want to update their product lines with new processors just before christmas season, and Intel barely has any. This only bodes well for AMD, that is producing all of its volume in the areas that sell best.

Sorry for the long post, but you'll find that all of this is stuff that comes straight from the horses mouth. Intel has said this is exactly what it's doing, and AMD has said this is exactly what it's doing. You can only help but put two and two together.

180 Sharikou said...

Dude - you are all over looking one critical factor. That is the strength of the Intel and Pentium brands. Do not assume because the technology is inferior that means the business is a dud. Netburst has made billions for Intel in spite of being the inferior technology. And now with the Pentium brand down from 130 all the way down to 70 bucks, suddenly many more average Joes can afford this. Just because a few techies on this blog say the technology is obsolete doesn't mean millions of consumers are not going to continue to lap it up. Betcha this christmas many families surprise their wives with a nice computer with a Pentium inside that doesn't break the bank.

Core 2 is NOT the game in Q4. It is Pentium...Pentium...Pentium.

And in Q107 they bring in single core Core 2 under the Pentium brand name to continue to mop up the low end of the market. As I've said before, Celeron is dead. At 50 bucks and less, this is only a fighting part to do tactical deals.

Pentium is going to eat the bottom out from under AMD. OEMs and channel love it because it's so easy to sell a Pentium at these low prices. Scan through the comments on some of my older posts and you'll see my opinion on this.

cheers...

The Sheepshagger said...

ashenman said...

I don't see how you can say AMD will be losing market share in q4. Yes, Intel will have core 2 and qcore 2. But qcore 2 will have an even smaller ramp than core 2 (only 1 million before quadcore opteron in 6 months?). Not only that, but only 25% of Intel's production will be core 2. AMD will have rolled out 4x4 by then, and will have its new 65 nm parts. This will be before Christmas (and seeing Intel's "amazing" ability to make its products available when they're released, I bet quad core 2 wont be).


I' not sure where you are getting your ramp information but I would like to clarify a thing or two, in particular your statement about Intel's production being only 25% in Q4. In case you missed last weeks announcement, Intel has already announced that they have transitioned > 50% of there volume over to 65nm. As a referesher... the only thing on 65nm is Core 2 material so this includes former codenames Conroe (desktop), Merom (laptop), and Woodcrest (server).

Intel hasn't said anything about releasing new core 2 parts that will increase performance, so they can only lower the prices by that time, which means they'll be making even less money on their primary money maker that still makes up such a small portion of the market.

Much has already been made about the successor for Woodcrest which is due to be released in November

As for as everything else, you're talking about new products (Conroe - released in July and Merom - released in August). As it is only the beginning of October, I wouldn't expect much 'new' news because Core, Core 2, and Core 2 Duo are new 'brands'. Too many changes when introducing new brands can only cause confusion for those people unlike us that don't follow computers so quickly. That being said, I can foresee speed subtle increase in the near future but wouldn't really expect major changes until after the holiday season. My advice? Keep this link handy.

Meanwhile, Intel is still producing 75% parts that are pretty much obsolete, including celerons, which have lost competitiveness due to cheaper AMD options. All of this in a season where average Joe is looking for a computer to surprise his family with, without surprising his wife too much when she goes to the bank.

Again... revert to my previous reply about the transition from 90nm to 65nm. Your 75% figure has to be way off. Even so, if there is a demand still for Celeron and Pentium 4 material still... why would Intel completely shut off if they can still make a sale while producing it cheap?

OEMs are going to want to update their product lines with new processors just before christmas season, and Intel barely has any.

I would guess based on this statement that you don't follow the industry too much or haven't followed for many years. The holiday buying period (Q4) is the season most computers are sold. You can bet your sweet pippy that OEM's are chomping at the bit to release new product lines and that they in fact have had engineering samples since the beginning of Q3 to prepare for such a launch in time for Christmas.

The Sheepshagger said...

A couple other things to add before I go off on a rant:

greg said...

I realize Intel has more assets it can eat into than AMD, since it's made more money longer, but it's using alot of those to upgrade all of its remaining fabs to 65 nm, and build up the resources to convert to 45 nm. This may not seem like as much of an expense as buying an entire other company (AMD/ATI) but seeing as fabs have operating costs in the Billions, and that intel has a ton of them, and you've got upgrade costs that are probably in the tens of billions of dollars that's eating into their assets as well.


Contrary to your opinion, Intel won't convert all their fabs to one technology (ie 90nm, 65nm, or 45nm). Given the fact that the the newest of the new computer chips running on the smaller geometries (currently 65nm with 45nm due in the next 1-2 years), it makes no sense for technologies that were designed and yielding properly on larger geometries (90nm and above) to be converted. There is no bang for the buck.

Old factories don't go away, they tend to get transistioned to other products like chipsets and communication chips. While Intel has in the past converted partial portions of other fab to expand neighboring fabs, this sort of transition has been typically a few hundred million, if even that.
The point is, an entire factory doesn't need to be gutted and re-tooled to go from one technology to the next. Semiconductor companies in general purchase new equipment when building new fabs that should and will enable them to produce 2 or more jumps in technology.

I would also like to encourage you, Greg, to look into Intel's financials as intel publishes how much of the budget is spent on expansion or construction. I used to track that myself over the last few years as an avid investor. Unfortunately, I don't have a link for you to qualify the following statement but Intel in the last 10 years has spent $3B - $6B annually on new fabs and construction in general vice the tens of billions you alluded to.

180 Sharikou said...
This comment has been removed by the author.
180 Sharikou said...

Sheepshagger - agree with all your points. Just one clarification, there are flavours of Pentium D and Pentium Extreme on 65nm:

http://indigo.intel.com/compare_cpu/default.aspx?familyID=1&culture=en-US

The TDP of these products is down to the 65-90 watt range and if you look at the reviews on Newegg, as long as you're not using the cheap plastic cooler that comes from Intel, the product is great.

Having said that, it doesn't change the logic of your argument and only strengthens your contention (which is mine too) that there is demand for Netburst...specially now that the Pentium brand is available so low down the pricing stack.

-:)


-:)

ashenman said...

There is demand for netburst now, but OEM's will want "fresh" marketing incentives. If they have a new line that they can push on their customers, then they will. Ya, the really cheap pentiums will eat up the really low end portion of the market, but most people exist in between that and the high end, and this is where AMD has more product, and where OEM's can get good volume from AMD.

To address the whole, you must not watch this sector alot comment, I was saying that about ramping new product lines for sale for christmas, just before christmas, meaning I was saying the exact thing you were accusing me of missing. That's why Intel will hurt, because this is where all of their competition is, and this is where the big money is. Ya, OEM's love how easy it is to get cheap pentium d's. But people don't just go, hmm, pentium, that's a good one, or that's the best. They look at a price point, a shallow list of features, and find what best fits that. From there on the question is, what OEM's want to offer what features with what processors. Which will depend on what's available. Yes, I've seen that 50% of their line is 65 nm, but 50% of that fifty percent is producing 9 series pentium d's, that are still more expensive than AMD's concurrent offerings, have a more expensive platform, and have a higher tdp and resulting failure rate.

Obviously AMD wont start selling more systems than intel in this area, but they are an extremely viable option, meaning they will take a good amount of sales from intel in this area.

Scientia from AMDZone said...

in particular your statement about Intel's production being only 25% in Q4. In case you missed last weeks announcement, Intel has already announced that they have transitioned > 50% of there volume over to 65nm. As a referesher... the only thing on 65nm is Core 2 material

I don't know where you got this idea from. Most of the 65nm production is dual core Pentium D and will be even in Q1 07.

revert to my previous reply about the transition from 90nm to 65nm. Your 75% figure has to be way off. Even so, if there is a demand still for Celeron and Pentium 4 material still

Your previous reply is incorrect. Even in Q1 07, C2D units will still be less than half of production.

Getting back to market trends. There are no numbers that you can compare for one quarter to determine a market trend. For example, AMD's revenue and volume share were both up in Q1 and Q2 of this year. However, the market can be elastic for one or two quarters so these are not necessarily genuine gains. I've charted AMD's and Intel's revenues back to 1999 and have a fairly good idea about how they've done. If Intel has indeed suffered a drop then their general trend should be upward. However, the 4th quarter can be a bump (a temporary high) as it was within the trend in 2001. AMD's revenue's were much more volatile in 2001 - 2003. However, with AMD's production increases their revenues should be much less volatile today.

The general trend I would expect from Intel is upward. The general trend I would expect from AMD is upward. Specifically, it wouldn't surprise me if AMD's revenue share dropped. This would only surprise me if it dropped below 15%. It is possible that Intel could gain in north american retail but this should be offset by stronger mail order, corporate, and sales in the Asian market for AMD. I don't see AMD's volume decreasing. However, if the market increases, Intel will be able to move more of its inventory and this would reduce AMD's volume share even if it were selling more chips.

180 Sharikou said...

Scientia - let me introduce you to the concept of market share which is what will reverse in Q4 if not Q3. Yes - both can grow revenue and profits but the point I'm continuing to make here is Intel will re-take market segment share. That's a black & white number...which we will know after the quarterly results.

180 Sharikou said...

Yes, I've seen that 50% of their line is 65 nm, but 50% of that fifty percent is producing 9 series pentium d's, that are still more expensive than AMD's concurrent offerings, have a more expensive platform, and have a higher tdp and resulting failure rate.


Ashenman - some bold statements. Let's dig in:

1. I think you're going to have to prove the failure rate with a link.

2. On the pricing issue, a search on Newegg will show the cheapest dual core from Intel is 90 bucks while the cheapest dual core from AMD is 160$. Merry Christmas old boy!

Obviously AMD wont start selling more systems than intel in this area, but they are an extremely viable option, meaning they will take a good amount of sales from intel in this area.

We shall see. My hypothesis remains - there will be a market share reversal from AMD to Intel in Q4...if not earlier.

Scientia from AMDZone said...

Scientia - let me introduce you to the concept of market share which is what will reverse in Q4 if not Q3. Yes - both can grow revenue and profits but the point I'm continuing to make here is Intel will re-take market segment share. That's a black & white number...which we will know after the quarterly results.

I will try again. There are two numbers for share: revenue share and volume share. In Q2 06 AMD's volume share was 22% and the revenue share was 18%. However, this was in a falling market. The point I keep trying to make is that share in a falling market is misleading. If the market volume increases I would expect AMD's revenue and volume shares to drop but this does not mean that AMD is selling fewer chips. This would only be a problem if AMD's gross revenues also dropped. The second mistake would in assuming that if AMD's share drops that this indicates a general trend. This could simply indicate a correction of a temporary decline in total market volume. As I've already tried to explain, AMD's revenue share was 14% but increased to 18% with the market drop. As long as AMD's revenue share stays above 14% this would be fine as long as the gross revenue and margin stay up.

Scientia from AMDZone said...

My statement remains the same:

You cannot tell a true market reversal from a single quarter's numbers because one quarter can be temporary.

A decrease in both revenue and volume share in Q4 06 would not actually be a market share reversal unless AMD also shows either decreases in total revenue or big decreases in margin.

In other words, as long as AMD continues to sell all of its chips without drastically cutting its prices its trend will be upward.

180 Sharikou said...

Scientia - I am talking about volume market share. Neither AMD nor Intel nor the market are static. Nor is sales and marketing a statistical exercise. Both companies fight for every sale they can make. At this point AMD has 22% voume share and has been on an upward trajectory. Starting Q4 (if not Q3) that trajectory will start to reverse. They will lose market share and go below 22%.

It doesn't matter whether the market is growing or shrinking and it doesn't matter if AMD is selling more chips than before. The point is of the available market they now have less than 22% which is a reversal of their momentum in Intel's direction. The only extenuating circumstance could be if they were sold out at 22% and couldn't provide a single more chip into the marketplace. I imagine this is not the issue with Chartered in place. However, even if this were the issue it's still a problem for them. Just like the chipset constraing in 2003-2004 was a problem for Intel.

I repeat - it is not ok to lose market share. Should I worry about gross margin and profit declines. Sure...but I'm not because right now both companies are in the middle of a price war. Intel to regain the market share they lost and AMD to get even more. And market share is the indicator of momentum because in this business market share does not shift back and forth on a monthly or quarterly basis. It heads up and down over several quarters.

Do you understand where I'm going with this now?

cheers,

180 Sharikou said...

Scientia from AMDZone said...

My statement remains the same:

You cannot tell a true market reversal from a single quarter's numbers because one quarter can be temporary.

A decrease in both revenue and volume share in Q4 06 would not actually be a market share reversal unless AMD also shows either decreases in total revenue or big decreases in margin.

In other words, as long as AMD continues to sell all of its chips without drastically cutting its prices its trend will be upward.


Momentum my friend - the word is momentum. See my comment above this one. In this business market share does not move back and forth QoQ. It shifts over several quarters. Just like AMD has been gaining over the last 3-4 quarters. And both Intel and AMD have already drastically cut their prices. The question is will ASPs drop. I thinkt the answer is yes - but that's just my opinion for now. Fact once we hear the Q3 results.

ashenman said...

Scientia was actually talking about momentum, which isn't as much of a black and white issue as you paint it to be.
Failure rates (being capacitor failure, product overheating, and instability) are all related to the core temperature of the processor. The 900 series operates in the 100-200 watt range, with the ee running at 230 watts (confirmed by toms hardware).
The 25% core 2 statement comes from IDF, where Intel said that's what it's product apportionment would be.
My statement about concurrent products still stands. Because what you're talking about are inferior products. Concurrent means equal in performance and volume. Thus, Intel's only concurrent processor to the x2 line would be the upper 900 level processors like the 950 and 960, which still underperform and have poor thermals. Thus increased failure rate due to dead caps, or failures on the OEM's part.
With lower tdp, the amount an OEM can be forgiven for poor cooler selection and assembly greatly increases, which OEM's love. Thus, they'll be happier with AMD products.

Scientia from AMDZone said...

Let me see if I can make this simpler.

AMD had a volume share of greater than 21% for Q4 05, Q1 06, and Q2 06. Therefore we should be able to say that AMD has lost volume share if its volume share drops below 21%.

Unfortunately, revenue share is not as straightforward and depends on total market revenue. At this point I should be able to assert that if AMD drops below 15% revenue share (which was the Q4 05 value) that this would be a definite loss. However, a higher number could still be a loss depending on the total market revenue.

180 Sharikou said...

Scientia - I am not disputing your calculation but I think I haven't made my point well enough. From a "running the business" standpoint, the first thing Intel wants is market share reversal because it is only once they have won back SKUs with their OEM customers can they think of influencing ASPs by selling them up. The last thing AMD wants to give back is...market share.

The battle for revenue share will follow the market share reversal in momentum. This is the nature of this business. Revenue share does not necessarily accompany market share...it follows it. Which is why when Intel and AMD announce qtrly results they highlight market share won & lost to Wall Street. It is the leading indicator of where the momentum is heading.

Let's agree to disagree on the importance of this and let time & quarterly results adjudicate.

Scientia from AMDZone said...

As I said, a drop below 21% would be a loss of volume share. Your 22% figure is not accurate for several reasons. First of all, the 22% figure is rounded up from 21.6%. Secondly, there is some quarterly fluctuation which does not indicate real changes so you need to look at the value over a longer span. For the last quarter of 2005 and the first two quarters of 2006 we have:

21.4%, 21.1%, and 21.6%.

This is basically flat. I wouldn't see a change from 21.6% to 21.4% or 21.2% as significant.

You are also placing way too much stock in the revenue share. For the same three quarters AMD had:

15.1%, 17.2%, and 18.1%

Which appears to be an increasing market share. However, the actual revenues were:

$1.307, $1.299, and $1.172

The volume share and gross income are flat (with a slight decrease in revenue). This is not surprising to me since AMD was volume limited by FAB 30 during this time. The apparent increases are due to lower demand for Intel chips rather than increased demand for AMD chips. Intel had much greater drops in revenue.

This is complicated but perhaps by Q3 and definitely by Q4 AMD's volume will increase due to production from FAB36 and orders from Chartered. I believe you were trying to say that we couldn't compare this Q4 with Intel's Q4 05 numbers because these were high. Actually, the Q4 05 was up only slightly from Q3 05, much less than Intel historically sees.

If AMD followed pattern they would see a 20% increase in revenue in the 3rd quarter with another smaller increase in the 4th quarter. If Intel followed pattern they would see the same revenue for Q3 as Q1. I can go over the actual numbers to see what is going on but it isn't as simple as you suggest.

Scientia from AMDZone said...

I forgot to mention that the Q2 06 volume share is skewed because of higher than normal volume from VIA as it was closing out its C3. Can you tell me what the adjusted number should be if we factor out the VIA interference?

180 Sharikou said...

Scientia - I'm not trying to gyp anybody and claim a moral victory if AMD loses some decimal point market share to Intel. I agree - going from 21.4 down to 21.2 is hardly a loss.

I'll try and put something as black and white as I can so that you don't feel I'm not willing to put my money where my mouth is and wiggle out through some back door excuse.

Here's what I'm thinking - AMD will lose 1-2 points of market share in Q4 (if not earlier) to Intel. What this means is when the quarterly results come out - AMD should declare they lost market share and Intel should declare they gained it. Now 1-2 points in one quarter is a bigish swing that I think will indicate a change in the direction of momentum.

I'm not focussed on revenue share at this point because the first thing Intel is trying (and needs) to do is regain market share.
If the swing is not compelling enough, I have no compunctions in saying I was wrong.

I'm really not trying to nit pick on nickels and dimes - as long as folks on this blog understand the "intent" of my thinking, I'm not into protracted arguments on the words. Fair...?

-:)

BTW - I'm not disputing your calculations and you're right, it does get extremely complicated. Unfortunately, sales teams are simple guys, they understand single minded focus and the focus for Intel is market share. Which is why I'm trying to focus on that one thing which is what matters most from a business management POV.

Scientia from AMDZone said...

I assume you don't read AMDZone. It's my general understanding that you would see a good 4th quarter as a rally for Intel and the start of a good 2007 with Intel pulling back substantial marketshare from AMD. If you had been reading AMDZone then you would know that I've been saying for quite some time that 2006 was Intel's last chance for a strong comeback. Basically, I've said that if Intel doesn't make gains against AMD in 2006 then the competition changes. AMD will go into 2007 with two FABs and a second supplier, plus a strong presence in servers, plus an increasing presence in notebooks.

I said this long before Intel showed Conroe and before AMD bought ATI. I still don't see this changing. Intel has no real hope of much change for 3rd quarter because the volume of C2D has been too low (less than 10% of all processors). The volume will still be too low on the desktop even in Q4. I honestly don't see Intel's making a big reversal on the desktop even though this is where they are strongest (because of the low Conroe volume). To me it should be interesting to see if Intel will have any effect on mobile. However, the biggest question will be if Intel can stop the erosion of the server market (which is where the volume of C2D is the highest). Their best chance for this is Q4 but I doubt this will happen. AMD seems to be well liked in HPC and this should be a mirror of the general server market. Secondly, Intel will have its hands full trying to replace Presler Xeon with Woodcrest and trying to hold onto 4-way with the same Presler Xeon.

I just don't see a big reversal happening. AMD was hurt in 2002 when they lost share due to delays with 130nm. They lost 1% share from about 9% and that hurt. Today, AMD has at least 15% share; they are far stronger than they were in 2002. AMD's downturn occured in the 2nd quarter when volume is normally down and Intel could take more share with extra capacity. In contrast, you are suggesting a downturn in the 4th quarter which is very unlikely as this is normally the largest volume quarter of the year. It has been suggested lately that K8L will launch in the 2nd quarter of next year, and, if this is true, this would be a very smart move on AMD's part. I assume your argument would be that Intel would stretch its 4th quarter capacity by selling off its overstock. This is possible; I'm just not sure that Intel can reverse a year trend in overstock in one quarter. I'm thinking Intel will have difficulty creating enough interest to use the overstock to displace AMD sales.

180 Sharikou said...

It's my general understanding that you would see a good 4th quarter as a rally for Intel and the start of a good 2007 with Intel pulling back substantial marketshare from AMD.

Yes to the marketshare question. Not necessarily to them having a good year. I have not made up my mind on that one yet until I see some more indicators on the fall out of the price war, inventory position on Netburst and one more quarter on Conroe (all segments) transition, etc. However, AMD's 2007 will not be as good as their 2006. Intel has a better chance of having an improved 2007 since 2006 was crap for them.


Intel has no real hope of much change for 3rd quarter because the volume of C2D has been too low (less than 10% of all processors).

As I said elsewhere in this blog, Intel will come in on track for their Q3 projections...so the impact of C2D is already in that projection. However, I am extremely curious to know what you think the right number for that ramp in one quarter should be? If you think Intel or even AMD build their business plans expecting to ramp their new architecture to a very significant volume in one quarter then pass me whatever you're smoking dude.

I honestly don't see Intel's making a big reversal on the desktop even though this is where they are strongest (because of the low Conroe volume).

First - I think Intel is strongest in mobile but I understand what you're saying so I won't quibble. I think they will win share back in desktop...not because of C2D but with a mainstream Pentium offering. And this will be a relatively bigger upswing for them in emerging markets...not mature ones like the US.

However, the biggest question will be if Intel can stop the erosion of the server market (which is where the volume of C2D is the highest). Their best chance for this is Q4 but I doubt this will happen. AMD seems to be well liked in HPC and this should be a mirror of the general server market. Secondly, Intel will have its hands full trying to replace Presler Xeon with Woodcrest and trying to hold onto 4-way with the same Presler Xeon.

So we know:

1. 2P is about 90% of the server market by volume.
2. In 3 months Woodcrest has become 40% of overall 2P market and 50% of Intel's 2P shipments as Otellini told the world at IDF.
3. If 2P is 90% of the market then I think it unlikely HPC reflects the server market in general.

Based on these, I'm inclinded to disagree. It feels like Woodcrest is ramping well.

AMD's downturn occured in the 2nd quarter when volume is normally down and Intel could take more share with extra capacity.

This doesn't make any sense to me for a couple of reasons:

1. How do you use capacity to take more share when overall total available market is shrinking? It's ususally the other way around - when you mis-call the market growth and hence build less capacity than the market needs or you have a manufacturing problem. Could you clarify this point further please...

2. Actual share loss in the market is a lagging indicator of loss of SKUs at direct customers. You lose design wins first so I don't think it's as "point in time" as your comment feels.

In contrast, you are suggesting a downturn in the 4th quarter which is very unlikely as this is normally the largest volume quarter of the year.

No - am not suggesting a downturn. I am suggesting that Intel will begin to re-gain market share it has lost in the last 3 quarters. As I said above, this is an indication that they are winning back SKUs at their customers...and possibly traction with the unbranded channel.

I assume your argument would be that Intel would stretch its 4th quarter capacity by selling off its overstock. This is possible; I'm just not sure that Intel can reverse a year trend in overstock in one quarter. I'm thinking Intel will have difficulty creating enough interest to use the overstock to displace AMD sales.

Actually, if they still have excess Netburst inventory I think they will be forced to write it off the books by Q2. They have announced EOL for almost all Netburst products by Q2. They have also shown us a single core Conroe to slide into where Netburst is right now and I can bet there will be a lower bin/price C2D below the 6300. I also think AMD's capacity problem short term is they don't have enough to meet their dual core demand. Hence, the dialing back of cache from 1MB to 512, EOLing certain product lines, etc. 2 months ago I thought this could be a big deal but some of the actions they have taken are clever. If they don't run into serious customer issues through Q4, they should be ok.

ashenman said...

180, you don't seem to be looking at this as an issue of who OEM's want to sell, and thus, who will actually sell, but more of a, these are the market trends, and thus, the world has to follow the rules of those trends.

You also like to assume what Intel assumes of its low end chips, that they will eat the bottom out from under AMD. If you'd look at your own blog, you mentioned that AMD based systems sell for less than their equals in pentium systems. How could Intel eat out the bottom if its still more expensive than what AMD is offering, performs more slowly, and is an older product. People who don't read your blogs or other think the AMD thing is some sort of new product, and thus, that it's better.

Like Centrino or Viiv, what sells the platform isn't the fact that it has better or more compatible and thus more solid hardware, it's the fact that there's a label on it that sounds cool that makes people think it's better than things without that label.

Last but not least, you don't point out that Intel will have to deal with the bad press of writing off the surplus they have, or that Intel didn't make any mention of a low end core product at IDF (though if they did, I'd really like a link, cause it would be interesting), which would lead me to believe that they would still be dependent on pentium for a while longer. Plus, the AMD scaling back the cache size was (according to AMD) to increase yields, and thus become more profitable. While this might lead to people thinking that AMD is having yield problems, it would make more sense for them to lie and say it was to meet demand, thus inflating how confident AMD appears to be. Plus, with AMD's track record, it probably wouldn't be a bad idea to believe them.

TheKhalif said...

I hav to disagree with your assessment of the AMD\Dell deal. I think that the price war affected them more than anything.


At this point, Dell needs AMD more than AMD needs Dell. AMD has been profitable for every quarter that Dell has lost revenue.

And Intel's new pricing strategy gives them no advantage but a helluva lot of NetBust.

They would be unprofitable for several more quarters waiting for Intel to even out the inventory.

Anonymous said...

Nice Blog. I will keep reading. Please take the time and visit my blog about: Internet Marketing and Making Money Online

MajorEnterprise