Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
wbmw,
Maybe they will. But it is still going to be regulated below the Core 2 premium price range, and be competing with Pentium D in the $100-200 range. AMD Sempron will also have to go down in price, since a P4 is still more competitive in the $75-100 range.
AMD already have some price cuts scheduled and when Conroe is out in force, there will likely be more price cuts, so that DC Athlon 64s start from $100 to $150 range. There will likely be a 3 GHz FX-64 (probably for more realistic price of $600 to $800) and a 2.8 GHz 5400.
Once that transition (to DC) takes place, AMD can stop disabling part of 512K cache on Semprons, and Semprons of Q4 and Q1 will be what SC Athlons are today.
AMD's ASPs will probably start dropping by $3 to $5 per quarter, hopefully offset by increased volume.
Joe
wbmw,
I don't think the number of parts from AMD has any relation to the ratio of parts. Just look at the number of dual core skus, back when they were <10% of the market. It covered nearly half of all AMD products, and yet only constituted a small portion.
I am not sure I understand your post. Intel's percentage of Pentium vs. Celeron was better than AMD's Athlon vs. Sempron. I don't think the ratios will change in the future for either company. All I was saying was that while Intel will go nearly 100% DC in mainstream to high end line, so will AMD by the end of the year. The "value" lines will stay SC.
Joe
wbmw,
It used to be 70% P4 and 30% Celeron. Now, by the end of the year, it will be 30% Core 2, 40% Pentium D, and the rest single core P4s and Celerons. What percentage of AMD's parts will be dual core? Maybe 25%? That means that 25% of AMD's production will be excluded from the upper 30% of the market, and 75% of AMD's production will be excluded from the upper 70% of the market. It's back to being relegated to the low end. And you cannot see this?
I think the trends of product introductions indicate that AMD's entire Athlon 64 brand is moving to dual core. There may be some SC lingering in Q4, but since AMD has not introduced any SC chips faster than 4000 model, and even 4000 1MB model is discontinued, it seems that vast majority of Athlon 64 will be DC. I doubt that DC vs. SC will be the distinction, or any advantage to Intel. The advantage will be the performance lead of NGA and the monopoly position in the market.
Joe
mas,
which one was it? The same where Opteron launched or something bigger?
Joe
Alan,
I expect closer to 500k... quite a gap in expectations there.
Is there any basis for your belief in 10K? I get my 500k from a 1MU market size times the Intel statement that 50% of Q3 server shipments will be woodcrest.
That's a fast ramp. Whatever happened to Dempsey? What's the point of even launching that? Woodcrest is way better.
Joe
mas,
Intel Corp. announced Monday that it will begin shipping its new Xeon processors for servers next week.
Weird. Why not just move back the announcement one week? Not that AMD has not done this before, but it is one thing to, for example, announce something, such as new Opteron 2xx and 8xx, and say Opteron 1xx will follow in certain time, but the announcement without shipment should just be postponed - for either company.
Was there some self imposed deadline to announce / release / whatever in Q2?
Joe
chipguy,
re: I meant both of Itanium customers,
Corporate and institutional?
HP and SGI.
chipguy,
re: Maybe the customers should camp out in front of Intel packaging plant to make sure they get their Montecitos. Both of them.
Given the rate that Opteron ASPs are falling two Montecitos
might be worth more than AMD's Q4 server MPU sales.
I meant both of Itanium customers,
Joe
wbmw,
Pretty weak, Joe. The 'Droids used to come up with all sorts of good reasons why IPF is doomed (well, at least along the lines of architectural philosophy, or the steep ramp towards adoption, or the lengthy and expensive process of porting software....)
The argument in the past was that Itanium was not going anywhere (or that the market would not go where Itanium was going).
The results are in, and there is no need for further arguments. Itanium went nowhere, that's where it is, and that's where it is going to stay.
In the past, one had to argue that SGI management's bet of the company (or the shareholders equity) on Itanium was not a good idea. Today, the shareholders are already wiped out, so what is left to argue?
From the SGI management: "some of you (SGI shareholders) may not make it back from our Itanium quest, but that's the sacrifice we are willing to make".
All you have now is a weak argument about low volumes, which isn't factually correct, and doesn't even make sense outside of a bitter tongue-in-cheek perspective.
Not bitter, delighted if anything. First Opteron put Itanium in a cage, and next, Intel's own AMD64 processors will drive the stake (woodcrest) through Itanium's heart.
Joe
chipguy,
The issue now, as far as SGI is concerned, is getting Montecito into the field, which should happen in about three weeks. "We are trying to get as many Montecitos as we can," says Matzke."
Maybe the customers should camp out in front of Intel packaging plant to make sure they get their Montecitos. Both of them.
Joe
wbmw,
It seems that SGI held its breath for a competitive x86 system a little too long. The shareholders "expired" during the wait.
<edit>
Or was it the wait for comeptitive Itanium that caused the company to go titsup?
</edit>
Joe
wbmw,
The cost argument is irrelevant in this case, since all of these are low volume enthusiast platforms. Kentsfield is nothing more than an Extreme Edition part, with some possible penetration into the top couple of mainstream slots. 4x4 is an uber enthusiast sku that no one will be able to afford unless AMD drastically reduces the pricing of their CPU stack.
AMD is reducing the prices in July, but I am still having hard time seeing the benefits of 4x4. Heavily multithreaded apps, such as rendering, visualization workstations - maybe. But for gaming, it seems of extremely limited use.
There was a press release of some software companies supporting this, but the benefits will arrive only when more games are rewritten to take advantage of multithreading. Even there, going from 1 to 2 cores gives you the biggest benefit, and each additional core provides less benefit.
I can see 2 reason for 4x4:
- if the Reverse Hyperthreading is actually a near term possibility, and if it speeds up execution materially in enough apps
- as a way to have more memory memory DIMMs, to add memory on the cheap
Joe
chipguy,
Anyone who knows anything about scalable computing
systems knows that compared to the orginal, the EV7, the
Opteron's CC scheme *is* a toy suitable only for low end
commodity systems. I guess we know where this leaves
you.
Considering that:
- AMD share of the server business was nearly zero
- Opteron HT solution is superior to competion in commodity end of the server market
- the commodity server market is > 50% and growing
I would say that AMD made the right decision.
BTW, the area you chose for your nitpicking only shows you are unable to see the forrest for the trees.
Joe
chipguy,
AMD will do what it has always done throughout its long
history of facing superior products from Intel. Slash prices to
a level below which Intel will refuse to go and then sell to the
bottom feeding part of the x86 market.
Intel management must think it is a good strategy, since Intel is now employing the same strategy (selling inferior Prescott Celerons to the bottom feeding part of the x86 market for $35).
Joe
wbmw,
He's either a fool for believing such nonsense, or a snake for using such a canard to generate sympathy from his supporters.
He identified Intel's problem. You are just trying to shoot the messenger, because the message is hard to accept.
Joe
Durl,
I always meant to ask you, but I am not sure if you will be able to answer: After all the drugs and the effect they left on your brain, do you think the drugs were worth it?
As I said, you may not be able to answer this, since you are operating with greatly reduced capacity to reason.
Joe
mas,
There's a 65W version of the 2.6 GHz 5000+ too.
That's great news. It means that 2.8 GHz at 89W (or is it 95W) should not be a problem and 3 GHz is not too far out of reach.
Joe
wbmw,
Intel's mainstream part still has nearly a clean sweep lead over AMD's FX part.
You have been talking about future Intel parts as if they were current for so long that one has to wonder if July is going to be the introduction or retirement of these parts...
Joe
UpNDown,
Re: reverse Hyperthreading
I don't think it was mistranslated. Conceptually, the wording makes sense. The label "Hyperthreading" corresponds to 1 core CPU acting like 2 cores, and is seen by the OS as 2 CPUs. Reverse of that is that 2 cores would act as 1 core.
Joe
Mike,
Weird that Distributor 2 has only 1 Sempron speedgrade in stock in stock (AM2) and none in S754.
Joe
mas,
I doubt very much that this will surface anytime before the successor to K8L as this is a very big change but OTOH it is the 2nd time we heard this rumor now .
I would think so too, but what do you make of this:
According to our personal information, obtained from the familiar with the plans AMD of sources, already in this year the company will open access to technology, reverse on its action Of hyper-Threading.
Also, 4x4 is aiming at the end of this year, which would make more sense with the reverse HT than without it.
Joe
Science to get a power boost
http://www.knoxnews.com/kns/local_news/article/0,1406,KNS_347_4775570,00.html
chipguy,
Cray wins $200m government supercomputer development contract.
Thanks for the link. Nice to see the official PR. The info what somewhat leaked some time ago:
http://www.investorshub.com/boards/read_msg.asp?message_id=10416730
Joe
gb,
One benefit is slightly lower power consumption. As far as performance benefit, there is only slight, since:
a) most apps are not bandwidth limited
b) according to Hans, there is a bottleneck in K8 memory controller allowing the full benefit of the DDR-800 memory to be realized only at 3.2 GHz clock speed, a little bit down the road, and even when fully realized, the performance benefit will be limited because of a)
Joe
wbmw,
It's more than I expected in such a short amount of time. I have a feeling that the Street is feeling similarly, which is why I'm backing out. If the stock goes back up a few bucks I might re-enter.
I would not be surprised to see the stock go to 30 and back. BTW, after selling my INTC calls, I bought some puts this morning, since I expect another visit to the lows. Longer term, I am bullish on INTC though.
And FWIW, I think you and others are seriously undercalling Intel.
The contest is all inclusive, meaning, including one time charges. I guess that's the expectations of people with negative entries.
There has been very little info on the restructuring. I think I heard that something may be announced today (on 15th) but there was nothing so far. Possibly after the market closes. That could be bullish if it is extensive, bearish if it looks to be only cosmetic.
Joe
Snow,
Yes, I was talking about microprocessors only. I have 7900M for total revenue. You guys (Intel investors) should enter your estimates when the next phase of the contest opens (beginning of July).
Joe
wbmw,
Congrats on the gains.
Joe
Alan,
so you expect CPU revenue of around $4B?
I expect < $5.5B. My EPS Contest entry is 5210M:
http://epscontest.com/06q2eps/mfc_entry_intel.htm
Joe
Dan,
What does it cost Intel to FAB, package, and test one of those die, then put it in a box with a good quality heatsink and fan and ship it to one of the dozens of on line retailer who sell it for <$40 after making some markup?
The packaging and testing these is the same as $200+ SC Pentiums or mobile Pentiums. Packaging the box + HSF is probably in mid single digits. What Intel receives for the processor alone may be in low 20s, which is roughly at cost (zero gross margin), possibly even below cost (negative gross margin).
Obviously, they aren't recouping any general, marketing, administrative, or FAB costs, but they can't even be recovering marginal costs.
Nothing left for those.
Intel can probably cover the variable costs, but only a fraction of fixed costs. Recovering even a fraction of fixed costs would be a good thing (compared to unused capacity) if these units sold were truly additional units. Some of these may be new units (that would otherwise go to VIA) but some of these may be cannibalizing more expensive Celerons...
Joe
AMD pushing tri gate my ars!!! Yeah thats a good one like they had it first this is what I call delusional!!!
Or should I say prove it!!!
???
I think you hit the reply button on the wrong post.
Joe
Alan,
LOL.
Wouter,
Does that matter? The value is in the chips themselves, not in the barcode that identifies it as product X.
I am not questioning the capability of these chips. They are capable of much higher clock speed. They are heavily donwbinned.
The surprise (to me) is the length Intel goes to achieve its goal to keep unit market share - fighting bottom feeder VIA for the scraps.
Joe
Tenchu,
LOL, I thought PR was only an "invented reason" by me ...
I will go to the barber to have my hair cut, and while waiting there, I will check out the latest issue of Newsweek (or something with more pictures than text).
Did I go to the barber to read magazines?
Joe
Alan,
We will just have to wait and see...
Joe
chipguy,
So either
1) you don't believe NGMA is huge a step forward
for Intel and a major setback for AMD in its bread
and butter desktop segment.
or
2) you have a brain pan sizzling case of cognitive
dissonance.
Which is it?
3) A trade. Bought on a dip, sold calls expiring on Friday. If I am right, I will not have the position on Monday morning.
BTW, I agree that NGA is a huge step forward for Intel from the position Intel found itself, but its near term impact is dwarfed by the impact (on both Intel and AMD) of Intel price cuts. That's where the action is for now, that's what will impact financial statement for current quarter and the next quarter.
INTC + AMD revenue is shrinking. AMD + INTC market cap is, and will continue to shrink. AMD + INTC profits are shrinking. That's what is relevant for now.
It reminds me of Linda Fiorentino (in the Last Seduction) making her passenger unbuckle the seat belt (among other things) and then crashing their car into the tree. Intel hopes to be able to walk away, but that is not exactly guaranteed, and there may be a plenty of bruises and broken ribs along the way.
I am mostly on the sidelines watching the spectacle, you are another passenger.
Joe
Alan,
If it was really about this they would not have prominently posted it on their web site. They would not have used layman friendly language in the lawsuit. They would not have taken out newspaper ads. The lawsuit is ONLY about PR.
PR accompanied the submission of the law suit, but it certainly is not ONLY about PR.
PR can certainly shame Intel into stopping the most egregious practices, and if not shame, then at least it puts the price tag on the risk of continuing with illegal practices.
Joe
wbmw,
Old inventory is old inventory, no matter what Intel calls it when they sell it to the market.
No, it is a brand new SKU, and Andy Bryant said that they have more of the "old" stuff in the pipeline.
While you continuing effort to find a way out of the dead end of your argument is admirable, I am losing interest in responding to your flailing on this subject.
Joe
Chipguy,
Their tireless, altruistic missionary excursion to this thread
to convince us heathen INTC longs that Intel's upcoming
beating of AMD to a bloody pulp in the desktop segment
isn't in our self interest is truly touching.
No, I think you are way more altruistic than I am, holding the bag, going down with the ship(s). It does not particularly interest me. I invest to make profit, whether it is AMD or Intel.
Speaking of which, I just sold my INTC leaps from 2 days ago for a nice 19% profit. I may buy INTC again, as it continues to make fresh new lows. And I have re-established a small position in AMD.
Joe
wbmw,
Ahh, but they are claims, Joe. The observations are easily explained as inventory reduction discounts,
How many times did I tell you that these are NEW never sold before SKUs.
Joe
Alan,
Good points,
If conroe ASP is about $300, and overeall desktop stays at $140, then the rest of desktop (besides conroe) needs an ASP of about $117 to maintain the overall INTC ASP of about $150.
Yes, but to reach $117 is not that easy, from portion of the market that does not contain high end desktop, mobile, and servers. AMD has not reached $100 without those exclusions.
My point was that Intel can't just slash and burn all of Netburst prices. There still needs to be segmentation, so that there is something to offset $50 Celeron prices, meaning higher priced Netburst, SC and DC CPUs.
Joe