Dell with Quad Core Xeons on day 1. That's some time to market and endorsement:
http://biz.yahoo.com/bw/061108/20061108005201.html?.v=1
This is where the rubber hits the road and all the talk about K8L disappears into thin air as the product still does not exist. At this point all the dudes who are fantasizing about how AMD's quad core will kick Intel's ass get a dose of reality. It doesn't matter if your architecture is superior. If the product isn't in the market you ain't selling it. By Q2 Intel will have introduced more flavours of quad core desktop parts and dropped prices significantly below 999$ which is the launch price. In Q2 (hopefully) or perhaps Q3 AMD will introduce their quad core part but will not be able to price it in the same range as Intel will have almost a 3 qtr advantage to stabilize yields and reduce manufacture costs on the part. AMD will then have to play catch up again as Intel begins 45nm late 2007 followed by new mArch in 2008.
On a seperate but related note, Apple introduces the new Macbook with Core 2 Duo.
http://www.apple.com/macbook/macbook.html
Welcome to the party. It's about to get interesting!
-----------------------------------------------------
UPDATE
The opening line was meant to be "Dell with Quad Core Xeons on MINUS day 1. That's some time to market and endorsement:"
The implication here is the official Clovertown launch is actually 14th Nov. Dell has jumped the gun and also jumped every other OEM. There's a message in here and it's a Dell message...not Intel or AMD. Dell needs to win in servers to drive their margins up. Specially since they are losing share to HP.
Subscribe to:
Post Comments (Atom)
13 comments:
Damn, this wasn't mentioned in Pervasive 64 Bit Journals? Quad core at 64 bit?
Did those bloggers get sh*tcanned along with Rummy?
You don't get it do you?
AMD doesn't need to catch up Intel. Having OEM contracts and market awareness they still struggle to sell this "old 90nm crap" - because demand is so high...
65nm conversion will help to increase marketshare. Even crazy price war and C2D didn't help Intel and AMD is bigger and bigger.
So for another year Pentium D will have to fight with Athlon64 X2 - soon faster and 65nm. And will lose this battle no matter how Core 2 Duo is fast...
My dear Mikolaj - unfortunately it is human nature to think you're right. Because if you think you're wrong it is also human nature to think differently. So whether I get it or not is decided by only one person - ME.
Intel is already seeing success with the sever and mobile Conroe parts in terms of shipments. They are struggling on desktop. As I've said earlier, this is due to the fact that there is not a wide enough range of lower cost boards to bring the system BOM down. Once this happens in this quarter, things will change in the channel. The other thing I've said is the Core 2 launch in Q3 was primarily channel and Dell. It's only in Q4 that we should see the other OEMs come on board in a signficant manner. Much of Intel's success here depends on HP globally and Lenovo in PRC.
Let me contextualize this. Unlike some other people...I don't see this battle having a single victor and time frame in which it will end. This is a swinging pendulum and the pendulum is weighted to swing back toward Intel at least till Q3 next year. But, to make that happen Intel has to do some things. Winning HP is one of them.
A Dell man quoted here http://www.theregister.com/2006/11/08/dell_delivers_quads/
told, AMD is up to deliver QC in Q2 07.
Anyhow with "dismissing" vPRO, Dell with the note about AMD QC lefts a taste of "taking wind out of intels sails".
Further after decades of "closed mouth" we see press releases with critisism to intel (vPRO).
I do not estimate, Intel will grab "all" QuadCore business of 07 in its "we are the only one quarter".
Although a operating temperature of 88 degrees (inqiurer ?) might be in normal range for silicon chips, we all have got more or less off of thinking about this "feature"; Will -or do we want- our T-monitor running up again, one eye on it ?
As a consumer / customer we all must conclude :
Because of AMD, just Dell's products shifted from "celeron only" to a lot of different type cpu's.
We see this all over on IT Industrie.
regards juergen
PS:
I can respect most kind of fanboy. But I agree to: if one tells to be journalist, he should proof it. If one has to tell, the paid days are gone long ago, he may ask himself wether he is acceptable any more.
May be he has a typo on his site with "Perversive" journals ?!
;-)
AMD doesn't need to catch up Intel....
oh yes they do, because Intel surely isn't going to wait around for them. Core2 will only be around for 2 years now as Intel announced their shorthened product transition.
Having OEM contracts and market awareness they still struggle to sell this "old 90nm crap" - because demand is so high...
Yes, which is why Intel still have P4 around until next year. The demand in the far east is very strong for low end processors. But what you don't realize is how Core2's performance have relegated all AMD's of products in the desktop to compete with P4's instead.
So for another year Pentium D will have to fight with Athlon64 X2 - soon faster and 65nm. And will lose this battle no matter how Core 2 Duo is fast...
You said so yourself. The problem is, as soon as Intel shifts the majority of its production to Core2, the bigger pressure it will get on its ASP.
Increased market share is just one thing. What part of the market that share is another. Core2 just removed AMD's share from the high ASP market share and moved it to mid to low end. While Intel now has 100% of the workstation/enthusiast market. That isn't good.
Roborat - your last observation is an astute one. I had said this back in August with the July price drop and moving Pentium down the pricing stack. AMD will get squeezed on pricing. We can see this by the fact that they have declining ASPs even while they are capactiy constrained. This defies the most fundamental law of economics.
all the talk about K8L disappears into thin air as the product still does not exist.
I see. Was this your point of view about Conroe back in May? Or is this another double standard?
By Q2 Intel will have introduced more flavours of quad core desktop parts and dropped prices significantly below 999$ which is the launch price.
Actually, no. Intel won't drop the price of quad core until it has to because this would cannibalize X6800 sales.
In Q2 (hopefully) or perhaps Q3 AMD will introduce their quad core part
Server in Q2, desktop in Q3.
but will not be able to price it in the same range as Intel will have almost a 3 qtr advantage to stabilize yields
Now you are dreaming. AMD's price will match whatever Intel does.
AMD will then have to play catch up again as Intel begins 45nm late 2007
I wouldn't get cocky if I were you. AMD had no choice but to wait until FAB 36 was built to start working with 65nm. However, this won't be a problem for 45nm. AMD should catch up a bit with 45nm.
followed by new mArch in 2008.
Is this the same fairytale about IMC and CSI? This won't happen until 2009.
isn't just like Intel has to shift its weight onto core 2 production like some lazy fat slob who just isn't doing so at the moment. Intel is pushing hard to shift to core2. Their current ramp rate is probably the fastest we'll ever see out of them, which is interesting.
You’re looking at the PC industry the wrong way around. Intel and AMD can only make CPU based on demand. They don’t ‘make’ the demand.
Intel (and AMD) makes new products and then entice the system builders to switch to the new product. This is a slow process and you have to put into consideration what your system builders have and what they have in their pipeline. You can’t all of a sudden stop making legacy products when your system builders have legacy parts (and their supplier is still making legacy parts). There is a long supply chain that needs to be cleared otherwise it’s a business loss for all of them.
AMD is going to ramp 65 nm pretty hard, and barcelona even harder. I'm pretty sure if you look at their previous ramp rates when they had less pressure, this looks good for AMD.
The rule isn’t different to AMD. It must only ramp and switch products based on the pace of the industry. Otherwise it will incur losses or inventory build up on both sides of the supply chain.
I don’t think you have any proof as to how much quicker AMD can ramp compared to Intel, but then again the need to ramp isn’t always the same.
There are 2 types of ‘ramping’ and the speed required to ramp is different to both. One is technology ramping, ie, from 90nm to 65nm transition. Typically it is done on a volume runner. This is the ramp that needs to be done quickly mainly due to cost. There is the cost of losing Fab capacity during the transition and the cost benefit of having the new technology.
The other ramp is the new uArch (Core2/K8L) ramp which needs to be done on a matured process. This ramp is market dependent and must ramp only slightly quicker than market demands. Historically this has always been slower than the process technology ramp.
I don’t see the point in comparing ramp speed since it is mostly based on the lowest possible cost and highest possible benefit depending on the company. Speed isn’t always a requirement.
Looking back, we’ve seen how both Intel and AMD have ramped quicker than the market (2000) and at times slower than the demand (1998). AMD today is capacity constraint at the high end which shows AMD is theoretically ‘behind’ in their volume ramp for Fab36.
Back in May, we had silicon Conroe that benched.. We have no idea how K8L will turn out. And he meant that K8L will have to compete with a product already established for 2-3 quarters before AMD competes.
He didn't say specifically drop the QX6700 price but that the introduction of more models mean cheaper entry into quad core. I don't think that Intel will fret about the loss of their low volume X6800. Though the cost is ~doubled, you cannot simply add cores, where with E6300+ you can clock up to X6800 levels. I can only see this as at least double the demand for Extreme.
You say Nehalem won't come until 2009 as if it were fact. If you mean Q408/Q109, ok. Otherwise, based on all the slides saying it's in 2008, you're being a bit vague as I'm sure S180 has not bothered to try read past the ~10K words before your 'CSI is dead' story. BTW s180, I like your sweet, to the point articles:)
Ashenman - I have got to disagree with your comment on low demand for Intel 90 & 65nm Pentium parts. Certain Pentiums at specific price points in emerging markets are selling above retail price. In fact, they are selling in the grey market at a 15-20$ premium...which is an opportunity lost for Intel.
The other comment on Intel's ramp is I believe they are sitting on a significant amount of capacity. Last year Otellini said he will build excess capacity so he's never in the situation of losing any customer's business. The down side I think is playing out for them as their factories are not full. I think the one thing they have to do is gain more traction with HP. My prediction earlier was now that Dell has gone AMD, HP will swing back to a heavier Intel wieghting simply because they must head in a different direction from Dell. Having said that, Intel's new product line up makes that decision more easily justifiable than it was when they had only Netburst.
It seems to me more that HP has been chasing after IBM's product line and Dell's prices.
In the server context...HP definitely competes more head to head with IBM than Dell. Primarily because the hardware is a cross-sell with their services business while Dell does not really get to play in those deals. However, in a broader across the board context a couple of things:
1. HP commands a similar ability to Dell to negotiate pricing with their suppliers. This is obviously better than most of the other smaller players so think of them roughly equal here (though Dell probably had better pricing with Intel which may have been part of the reason HP went AMD in a big way a few years ago).
2. Dell has been traditionally cheaper than HP due to their supply chain management efficiencies which allows them to reduce inventory and the costs around that. Which they pass on to their customers. Over time that advantage has eroded as the other's have figured out how to do this too. HP on the other hand does not pass on all the savings they make on pricing negotiations to their end users. Instead, they try to ensure they have robust margins on what they sell. My own beliefs on why HP is re-gaining share against Dell now that Dell's sole competitive advantage of cost is whittled away are:
a. Dell is a supply chain company with a screwdriver. They don't really innovate in their technology (at least not much that I can see). Secondly, their brand has taken a beating in recent times with some of the customer support issues they've been having.
b. HP on the other hand is re-discovering it's technology roots. It is trying to bring new innovations to end users and is also re-investing in what used to be one of the strongest technology brands around.
The big take away is HP is re-focusing on innovation. Because in this business that's what customers pay a premium for. They go Intel more strongly because they know that they will have more clout at the negotiation table now that Intel has lost Dell. They will get better pricing and better partnership at just the point when Intel's product line up has stopped sucking. my suspicion is they will then start doing what Apple is doing - demand product customizations from Intel so they can build cool and innovative products. They know Intel wants their business and they also know of the two companies it's Intel who has the resources to do this for them.
As a parting thought...the semiconductor business is going to evolve. It's going to start representing the auto industry. Can you figure out what I'm talking about...?
That's not completely true. There is a ton of R&D that HP specifically does. Dell does some too but not as much. The issue is when a product manager for a certain line decides what he wants to sell, he's really focussed on beating the competition...usually through cost. He's not thinking about how do I create or expand the category by adding new and innovative features. Because then he has to create the market into which other OEMs can walk in later and compete. For example...HP was the first to really promote fingerprint security on the notebook in a noticable way.
So what I think is happening is at the mangement level HP has revised their PC business strategy. They've decided they will bring innovations to market to lead the market and command a premium. Basically...back to the roots of the technology industry. I suspect this is why they acquired Voodoo. To have a premium gaming brand they can take more mainstream and yet sit above the regular desktops/notebooks in terms of price premium. Rahul can scream all he wants that Voodoo will remain a super premium brand but once he's in...he either gets on the HP train or gets left behind at the station. Expect to see HP at least on the consumer side worrying and competing more with Apple in the next couple of years and expect to see Dell get squeezed further down the stack purely as someone who sells on price thereby hurting margins. This is why I think servers are important to Dell. Same reason they were (and continue to be) important to AMD. It allows you to fund lower margin businesses where the volume is.
Post a Comment