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,
I was just repeating (paraphrasing) what Andy Bryant said.
Joe
wbmw,
Not sure what data you are using in your statement.
Just the fact that Intel has to discount them so deeply, while at the same time complain (during the CC) that these are expensive to make (2 die vs. 1) and package.
Joe
wbmw,
Just look at their prices right now. It's a rip-off for an AMD X2 processor. Their 2GHz, 512K L2 cache X2 has a list price of $303. For only $13 more, you can get a Pentium D 950 with much better performance at a reasonable 95W TDP.
I think Intel screwed up very badly by releasing the 8xx line of processors. It left such a bad taste all around that all of the Pentium D products are considered junk, and Intel seems to be having hard time giving them away even at fire sale prices. That includes the 9xx processors, that on merit may be a lot more competitive.
With 8xx and Paxville, Intel just tarnished its reputation for no reason, no gain.
Joe
In addition, its Intel-only stance has cut Dell off from the healthiest part of the x86 server market, which is for gear based on AMD's Opteron chip.
http://www.theregister.co.uk/2006/05/08/dell_q1_cut/
chipguy,
Itanium allies have pledged to spend $10 billion advancing the Itanium technology and market in the next five years, and $1 billion a year of that is from HP, he said.
There is quite a difference between concepts of:
- advancing the Itanium technology
- R&D line on a quarterly statement
Under the loose standard of "advancing the Itanium technology", most of SGI's expenses would be comparable, not just the narrow R&D line.
re: Why are HP figure fake? Because nobody invests 29% of R&D in part of the business that brings in 1.19% of revenue. HP certainly does not.
Yeah sure Joe. Hurd is a blatant liar.
I am saying that HP is not spending 29% of its R&D in area that generates 1.19% of revenue. Paraphrased Hurd statement said nothing about R&D dollars, only about vague "advancing Itanium" dollars.
At this point, it would be a good time for you to concede one of 2 points, either the vague 10B figure, or that comparable SGI vague figure is far beyond the strict R&D line.
Since the $10B myth is something you will have to maintain in the future, I figure that conceding that SGI pledge was a lot more than $400M would be a lot more convenient for you...
Joe
chipguy,
If AMD couldn't do that in the last few quarters being a
pure x86 MPU play vs Intel's x86 MPU plus chipset plus
flash plus telecom plus embedded product lines it would
have been a huge embarrassment.
Funny for a guy with Itanium on his mind all the time, you somehow forgot Itanium as one of the things dragging down blended Intel margins vs. pure x86 MPU margins.
I am sure that despite the fact that it fits in the pattern of half truths, in this case the omission was accidental.
Joe
chipguy,
Yeah it's clear how much you much resent having your
BS shot down.
You brought up a counter point, that happens to be invalid. That's a far cry from shooting anything down.
Your counter point is that HP is investing $1 billion per year (using fuzzy numbers) - in order to lend credibility to $10B figure
On the other hand, you said that investing only $80m / year, using strict figures.
These 2 figures are not comparable. You can either use strict figures, such as what you did with SGI, and in that scenario HP figure and 10B figure shrink considerable, or you use a fuzzy figure, based on which SGI pledge is much more. It would not at all surprise me if SGI is (or should I say was) #3 on the list of that Itanium alliance.
Why are HP figure fake? Because nobody invests 29% of R&D in part of the business that brings in 1.19% of revenue. HP certainly does not.
Joe
Joey,
a cut-throat PC pricing environment will do wonders for AMD's margins, dont you think??
Interesting you bring it up just when AMD margins surpassed Intel margins for the first time ever...
Joe
chipguy,
BTW, SGI spends about $20m per quarter on all R&D.
Presuming that is all going to IPF development that is
only about 4% of the $10B over five year pledge.
One day you pump their up, such as:
For example SGI alone bought over 12,000 I2s
in its first year of selling the Altix 3000.
http://www.investorshub.com/boards/read_msg.asp?message_id=3349070
The strong get stronger:
http://www.investorshub.com/boards/read_msg.asp?message_id=2564904
And now, suddenly, on the day they file for bankruptcy, they are negligible. How convenient...
Joe
chipguy,
only ~7% of total shipments in Q1 were X2 and that represented an increase of ~3% from the prior quarter
In line with my expectations:
http://www.investorshub.com/boards/read_msg.asp?Message_id=10683045&txt2find=x2
chipguy,
If you aren't joking then all alternative explanations for the
utterance of such nonsense are extremely unflattering.
When I questioned the $10 billion pledge, I was not joking. A lot of creative accounting likely went to this figure. Do you have a list of who pledged how much?
It would be interesting (if there was such a list) to see SGI contribution. SGI can probably say that most of their expenses were part of their contribution.
Joe
mmoy,
It's a bit surprising that they couldn't engineer a sale to
a bigger player.
A potential acquirer would have to be someone who wants to become stronger in this field, which would limit the choices to HP, IBM and potentially Dell. HP already has plenty of its own investment in Itanium, IBM is out of Itanium business and Dell does not seem to have appetite to get in. So th universe of potential buyers is extremely small to non-existent.
Not sure who'd be in the market for something
like this but they'd be attractive just for the Itanium expertise
to one of the companies spending the $10 billion.
Nobody is "spending" $10 billion. There was a pledge to spend that much. SGI probably accounted a goo chunk of that pledge, which they will most likely not be able to honor.
Joe
Chris,
Meanwhile incompetent management continues in place, with huge new stock options issued. That's the American way........
It would not surprise me if if the management got their zero worth old stock / options replaced by newly issued stock...
Joe
mas,
Im tempted to buy at 6 cents !
Don't!
All of SGI's existing common stock and the unsecured subordinated debentures will be cancelled upon confirmation of the plan by the court and receive no recovery
http://www.siliconinvestor.com/readmsg.aspx?msgid=22427821
Joe
cruzbay,
Speaking of which, SGI just filed for bankruptcy. This generally includes some disadvantaging of the shareholders.
Yeah, high probability of being completely wiped out can certainly be called a "disadvantage".
Joe
chipguy,
Wow what an astoundingly ignorant comment. Total ARM sales
last year was ~1.6 billion processors. Intel's share was about
30m units although at much higher than industry average ASP.
So how is it that Intel is not making it? It has the same advantages as Itanium vs. competition - use of depreciated Intel fabs at variable costs, free use of Intel process technology...
Joe
Intel to cut XScale group
If it ain't a cell phone part....
http://www.theinquirer.net/?article=31476
I am starting to feel sorry for chipguy. Another one of his favorite computer architectures is on life support.
Joe
wbmw,
How many processors do you think AMD sells through Cray in a year?
Wasn't there an announcement just recently that Cray will sell 20,000 quad core processors in a single system?
Cray is certainly no HP, but it probably is within top 10 Opteron customers.
Joe
RGood,
Actually, I would be more worried about INTC's 65nm and 45nm progress.
Well, and I would be more worried about holding an ever shrinking stock.
Then Flecks reliability predicting INTC's future...
Now let me see what you advised us to worry about in the past and your predictions for the future:
First INTC doesn't have to pay IBM for the FAB use.
Second - SOI is a well known as being an expensive process.
Third - Out of every chip they sell right now will be costs to pay off debts. Huge debts.
Of course I speculate. But I speculate with 3 facts. But of course all of the AMD cheerleeders are shaking there Pom Poms right now. So go for it.
http://www.investorshub.com/boards/read_msg.asp?message_id=1336418
That was after Opteron launch, before Athlon 64 launch. Looks like you 0 for 3. Or should I say 0 for 2 and 1 unintelligible gibberish?
How about another prediction of yours:
Like I said before - It is gonna cost AMD a heck of a lot more to produce the Opti than INTC is in making the Itanium.
http://www.investorshub.com/boards/read_msg.asp?message_id=1336089
Are you starting to see a pattern here?
Joe
wbmw,
This is not a matter of whether Intel *has* been a bad investment. That's just hindsight. This is about whether Intel is *going* to be a bad investment, and the facts are *not* in for that.
Nobody knows that now, but there is great deal of humor in Ronster and rest of this thread attacking Fleck's credibility, when in fact he was right, people on this thread were wrong and are holding a stock that has performed very badly.
You can't point to a broken clock and claim it's usable, just because it happens to be pointing at the right time.
I think a compass would be a better analogy for prediction of stock movement. He has been pointing in the right direction.
Come later this year, he's going to start looking quite foolish if he doesn't close his short positions. _That_ is my forecast.
Could be. I don't know if he discloses his positions. He may even be trying to get out of his position as we speak.
But it will certainly take a big move on the upside (while Fleck is pointing south) for Intel longs to equal Fleck's credibility.
Joe
RGood,
We all here at the INTC thread know enough that Fleck is so full of BullSh!~ that's not even funny - he consistently has been wrong on INTC.
FYI: Intel is near 2 year low
He may be a good hedge fund guy but his bend on INTC has been consistently negative.
The facts are in: He is right, you are wrong. Intel has been a terrible investment for years now.
A good thing for you to think about is this: What is it that Fleck saw that I missed? What can I learn from Fleck to become a better investor?
Joe
chipguy,
(old post)
This is all meaningless drivel and FUD. Why don't
you do some research on the bin spead of modern
CMOS processes across plus and minus three sigma.
To get the complete picture, you need the center of distribution (coming out of your fab) and center of distribution of what market demands.
I don't think you even understand what variable cost
means. It is the incremental cost of materials, labour,
energy etc to manufacture one additional unit. Design
costs, capital depreciation, advertising etc do not
come into play.
I chose Banias because it is manufactured in the same
process as Intel's current leading edge chipsets. The
variable cost of Banias today would be basically the
same as it was when it was at peak sales.
I understand variable costs very well, and I understand that they are a fraction of total costs, which you are conveniently disregarding just to make a particular point.
The cost is "basically" the same, while you swept so many relevant things under the carpet that the carpet looks like a mountain.
Joe
Chris,
• If it delivers as advertised, ZRAM, could potentially reduce AMD’s die size by up to 40% and reduce AMD’s cost per chip by up to 50%. Additionally, ZRAM also has the potential to materially improve both the performance and power footprint of its MPUs.
For this to happen, AMD would first need to grow the die size by some 100% with SRAM L3. Only then it could be reduced by 40% by replacing SRAM L3 with ZRAM L3.
Joe
chipguy,
He is a former Intel employee whose group designed x86
MPUs and during the 1990s competed with the IPF group
for money and resources and apparently still holds grudges.
As an Intel shareholder, you should be thankful for all the battles he won.
Joe
morrowinder,
You getting nervous now that AMDs at 31?
I certainly admire your calm demeanor, taking it up your ...
Joe
chipguy,
Montecito was delayed for nearly a year? So you expect it
not to ship in volume until early 1Q07?
No, H2 2006 vs. H2 2005.
The fact is SGI was the author of its own misfortune in Q2.
Madison 9M 667 MHz FSB SKUs have been available for
a while but SGI simply stumbled on the intro of the 4700
series Altix entirely on its own.
Seems like an answer of the socialist, when faced with failures of socialist experiments: "They were not socialist enough".
"Since I joined the Company less than 90 days ago, we have redefined our strategies and approaches across markets, geographies, and products. The intent of these efforts is to reposition the Company to address 80% of our customers' spend versus the 20% we address today...
http://biz.yahoo.com/prnews/060425/sftu137.html?.v=28
Could the new CEO be thinking beyond Itanium?
Joe
wbmw,
Two new brands, each contributing $1B to the bottom line in 12 months.
Revenue is called the top line, profit is called the bottom line.
Joe
SGI blamed the revenue shortfall on delays in winning final customer acceptance of some computers and a decline in revenue from computer visualization products following the company's announcement in March that it was ending further development.
Hmm... No mention of Montecito delay making SGI products uncompetitive for nearly a year?
I guess it is not advisable to bite the hand that feeds to (free or discounted Itanium processors).
Joe
chipguy,
Yet another example of PC myopia. Perhaps you should
review the history of the computer industry to see how long
new ISAs took to grow to prominence, especially in the
server market. The RISC architectures around today were
introduced 14 to 20 years ago.
That was at the time RISCs were competing with very limited PC based servers. There was a plenty of room above that for RISC architectures to exist, prosper and multiply in numbers. IPF missed those good old days.
These days, if you went to a veture capital firm with a proposal to start a new high end CPU with new ISA, I think a secretary would call the paramedics, and would have you escorted out to a room with padded walls...
IPF is an example how bankrupt that idea is today, despite the fact that IPF has a nearly decade long head start, from the time when the idea of starting new ISA was not completely crazy.
IPF itself is bankrupt many times over, only kept alive because Intel can move large sums of cash from its profitable PC business to unprofitable Itanium business.
IPF didn't start to get taken seriously until McKinley was
introduced in 2002
Well, it took a long time to get it out of the lab. We can discuss if it was:
- poor execution
- ill conceived idea to start a new ISA
- EPIC itself being less than a briliant idea
I expect IPF to overtake SPARC this year and POWER
before the end of the decade.
Very possible, as long as Intel keeps the life support going. But I am more interested in how this market (of boutique CPUs) will continue to shrink vs. x64 servers.
Maybe we should start a contest of which processor will be the last one to turn out the lights on this segment: SPARC, IPF or POWER
Joe
chipguy,
What I see is IPF servers growing revenue at 90+% while x86
servers grow at under 7% and SPARC shrinks at around 10%.
IPF's main competitor, POWER, went from ~10% growth in 4Q05
to falling 9% in 1Q06 while IPF sales growth is reaching triple
figure percentages in important high IT spending growth markets.
In Japan IPF sales were already running 20% ahead of POWER
and 50% ahead of SPARC in late 2005.
You have been (selectively) quoting these numbers for years, how Itanium is outgrowing this and that. But there seems to be some disconnect. After some 5 years of these (selectively picked) growth rates, the fact remains that as Itanium is still fighting to move up from the last place when it comes to 64 bit processors. POWER is ahead of it, Xeon and Opteron raced past it. Sparc may still be ahead of it, despite the fact that Sun is gradually moving away from SPARC.
Joe
chipguy,
You better tell HP. The biggest seller of Opteron systems
is under the mistaken belief that Montecito based systems
will outperform contemporary Opteron systems 2x socket
for socket.
Once I got past the key word in that sentence: will, the question immediately popped popped into my head: If it does, for how long?
IPF bashers make me laugh alright. Year after year IPF
sales continue to grow rapidly despite every new reason
they throw out why it is doomed. First it was AMD64,
then it was Intel's own EM64T line. Now it is son of HT.
Yeah whatever you say Joe. :-P
If you are happy with Itanium progress, who am I to disturb your happy state of mind. But if you look around, you will realize that in sales, Itanium is looking at the tail lights of just about all of the competing 64 bit processors still on the market: POWER, Opteron, Xeon, and probably Sparc too. Considering what we were told that Itanium is the one 64 bit processor to rule them all, it is not entirely happy situation...
Joe
chipguy,
ROFL. SGI has its Numaflex technology which scales to levels AMD can't even dream about for HT.
I don't mean to spoil your day or anything, since you are in such a good mood and laughing. The slight problem is that your analysis is incomplete.
Whether you like it or not, Opteron is Itanium competitor. In all areas where these processors overlap, Opteron (and Xeon) won, Itanium has lost.
The next generation of Direct Connect Architecture will increase Opteron capabilities, it will increase the areas of overlap, which will result in Itanium losing out, and retreating to higher ground.
Ok, you say that Numaflex is the Mount Everest. You may be right. But you have lost India, Bangladesh and China. All you have left soon will be Nepal.
Joe
mmoy,
We had rumors about January 2006 to May 2006 and now Q3.
Q3? I must have missed something. Do you have a link?
I seem to recall 2 dates: May 23 and June 6, and May was supposed to be for Turion X2.
Joe
chipguy,
In scale up machines SGI's competition is overwhelmingly
IBM, not Cray. In HPC the competition from x86 is mainly
low priced clusters using commodity boxes. Cray faces
the exact same problem with cost based competition from
x86 clusters for its scale up systems as do IBM, SGI, and
HP.
SGI must have realized that they can't beat the competing x86 boxes (which are very competitively priced) with Itanium based boxes, so if you can't beat them, join them.
The K8L doesn't change a thing except perhaps the
performance/price flavour of the month in the x86 cluster
commodity box market.
It may turn out to be quite a bit more than that. It will be sitting on top of what AMD calls Direct Connect Architecture 2.0, which will likely help commodity x86 boxes scale up... Maybe that is what SGI is really after, and they are just getting their feet wet.
We will see what happens between now and then (arrival of DCA 2.0), if SGI does not "expire" in the meantime. But regardless, if it turns out to be true, it is an interesting death bed conversion by SGI...
Joe
mmoy,
I'm not crazy about AMD and their approach of slipping DC Turion again and again as well.
I think DC Turion is Rev F, and Rev F slipped.
Joe
chipguy,
It could also be that SGI got a peek at quad core K8L that one of their competitors (CRAY) will be selling...
Joe
wbmw,
Intel did mention in one of the warning that Intel lost market share, but since AMD units were down on the quarter, some 7% IIRC, I don't think Intel units were down more than that. There were some positive stories about market in general.
My guess is that INTC did not lose any market share, and if they did, it is not much more than 1%. I think INTC earnings would have to even below the lowered guidance to lose more than that.
My speculation is that Intel sold some processors at very low prices in March, keeping the unit market share, lowering ASPs, and losing some dollar market share.
Joe
Alan,
On market share in unit terms, my expectation is for only minor changes. AMD units were down 7% in Q1. Intel units would have to be down more than that, which I don't think has happened (unless the inventory situation skewed things in a way that channel inventory of Intel stuff got sold, and Intel sales were reduced by that much).
I expect Intel ASP drop, roughly the same unit market share, the result of which should be some dollar market share gain for AMD.
Joe
wbmw,
Keep in mind that Joe already lopped off 25% to the numbers before coming up with his brand averages. So another 10% brings the total discount up to 35%.
It's possible that the largest OEMs get that kind of discount. But, I think the discounts are much more controlled than before. In the past, you could always get OEM part cheaper. Now, the availability of OEM parts is greatly reduced.
The pricing on Pricewatch, for example, is close to list price, but if you look at some of the PCs at HP, you can get something with an Athlon 64 3400+ for $350 and upgrade to a 3800+ for $40 more. They couldn't offer this if AMD was charging them anything close to list price.
I recall the same practices going on on the Dell site, BTW.
Joe
wbmw,
My take: I happen to think your product mix is a little off. Here's my take, assuming your sub-product mixes are correct:
Regarding your product mix my guess is that your point of view may be too US centric. It may be even better than that in the US, but beyond US, Western Europe and rich Asian countries (Japan, Korea, Taiwan), the mix is much more "value" centered.
Joe