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.
Re : "Athlon 64 was 6 months late, too, but Opteron was early."
Dan, i sold my AMD stock at $74 today morning and bought MSFT for $7
Re : "It's amazing what AMD can do with it's limited resources."
this, my friend, is beautiful about you. AMD stock price is lower than it was 2 decades ago. If you count inflation, then it's equivalent to going through 2 or 3 reverse split over last 20 years. And all this during a time when the semiconductor industry has (probably) grown faster than any other industry in the history of business.
And yet, you find reasons to be amazed about AMD year in and year out.
Maui.
Article : http://www.businessweek.com/magazine/content/04_12/b3875079_mz063.htm
According to the IDC report, Big Blue had the fastest percentage growth for the past eight quarters in the market for Intel servers, which cost $1,000 to $25,000. One key reason: IBM sells servers with 8 to 16 Intel chips, while Dell has stopped selling those more complex systems. IBM is using its knowhow in 16-chip servers to deliver other innovations: Customers can rent supercomputing power from Intel clusters in IBM's new hosting centers.
Maui
Article : Intel damaged the competition with the Itanic.
http://www.theinquirer.net/?article=14550
----------------------------------------------
Maybe inquirer is getting the big picture now!
Maui.
Intel to Strengthen Itanium 2 Processors for Workstations
http://www.xbitlabs.com/news/cpu/display/20040306094234.html
New Itanium 2 DP CPUs to Come in Q2 2004
by Anton Shilov
03/06/2004 / 09:43 AM
Intel is preparing a rather surprising move to substantially strengthen performance of Itanium 2 processors for DP applications. Without waiting for the Fanwood chips to arrive in the third quarter, the chip giant will roll-out new IA64 chips with 3MB cache already next quarter.
Santa Clara, California-based corporation is currently shipping two Itanium 2 flavours for 2-way applications: Itanium 2 chip at 1.40GHz with a 1.5MB L3 cache and the Low Voltage Intel Itanium 2 processor at 1.0GHz with a 1.5MB L3 cache, which consumes approximately 62W – half the power of other existing Itanium 2 processors. Such microprocessors are suitable for technical and scientific computing systems; various clusters; entry-level, front-end enterprise systems as well as network edge and software engineering workstations.
To strengthen the current DP IA64 offerings, Intel had planned to introduce cut-down versions of its forthcoming Itanium 2 “Madison” – code-named Fanwood chips with 3MB L3 cache – in the third quarter of the year. However, according to the company’s latest plans, there will be two new DP Itanium 2 “Deerfield” processors with a 3MB L3 cache at 1.40GHz and 1.60GHz speed bins available already in the second quarter! CPUs with Fanwood core and 3MB L3 cache are expected to hit the market in the third quarter at 1.60GHz and 1.20GHz (low voltage version) speeds.
The Intel Itanium 2 processor at 1.60GHz with 3MB of L3 cache and the Intel Itanium 2 processor at 1.40GHz with 3MB of L3 will be quoted at $2408 and $1172 in 1000-unit quantities, respectively.
Representatives for Intel Corporation did not comment on the news-story.
Re : "Madison and the more recent revs were done with other folks in Oregon Duhhhh!"
Madison was designed by santa clara team (same team that did Merced).
Maui.
Article : Intel is Heading Towards Dual-Core Desktop, Mobile, Server Processors
Dual-Core Era to Start Next Year?
http://www.xbitlabs.com/news/cpu/display/20040229173249.html
Maui.
re : "I recall that overall AMD ASPs reached a high of approximately $95 in early (maybe Q2) 2000"
As a reference point, current ASPs are listed out (for both Intel/AMD) in the BW article at :
http://www.businessweek.com/magazine/content/04_10/b3873005_mz001.htm
A $95 ASP will be a nice boost to AMDs bottom line and their ability to compete in low-end server/WS work in their favor.
The factors that will hurt them are :
1. Intels manufacturing capacity this time around (compared to 2000) may not result in overall shortage, as in 2000.
2. Intel cost structure will continue to get better relative to AMD (Intel expect 25% reduction in 05 and AMD will have to come up new capacity/FAB) and will put pricing presssure in selective market.
Maui.
Re : "Actually, that wasn't the argument being advanced."
but that was the argument you attacked tecate on.
No ?
IBM rises, Sun sinks in server market
http://news.com.com/2100-1010_3-5165213.html?tag=cd_lede
Intel/AMD related :
The market for servers using Intel processors and Intel-compatible processors--Intel's Xeon and Itanium and Advanced Micro Devices' Opteron, for example--grew 18 percent to $20.3 billion, Gartner said.
Here, HP maintained its No. 1 rank with $6.8 billion in sales. Its market share dropped 0.4 percentage points to 34 percent, however.
In Intel servers, No. 2 Dell grew 22 percent to $4 billion, with its market share increasing 0.7 percentage points to 20 percent. IBM's revenue increased 21 percent to $3.4 billion, or 17 percent of the market.
The logic being pointed out to you is that Ashok Kumar is just not good at predictions.
If you step away from your attacks, you might see more possibilities.
Article : Intel, AMD comfy during fourth quarter
http://news.com.com/2100-1006-5152255.html?tag=cd_top
Maui.
Re : Perhaps Intel could have done a quick shrink of Northwood to 90nm? With Prescott struggling, this should have been a no-brainer."
there is no quick shrink anymore. The process shift is so complex that it is often easier to redraw. The days of easy shrink to get performance are over for good (since 180nm to 130nm shrink).
Maui.
smooth, re : "this business has a very low profit margin"
lower ASP than a CPU, but not much lower margins. Dont forget these chips are manufactured on older processes and in FABs that are almost fully depreciated.
Maui.
facsnotfiction,
Re : "Article claims Intel 100k Itanium shipped is false"
Your skepticism is undestandble given that there are still people who think humans never landed on the moon (-;
seriously though, what's so hard in believing that Itanium will do alright in enterprise space ? HP is putting all muscle behind it. They are getting out of chip design business and have spent years to move everything on Itanium and the move is beginning to happen and will continue next yearand beyond. There are other companies who have bet the farm behind Itanium and Itanium strength is beginning to showing (FP performance) after so many years of investment.
I wouldn't get too hung up on opteron vs xeon/itanium or Intel vs AMD. The way I see it is Intel and AMD jointly has pushed everyone else (excet perhaps IBM) out of chip making business and all those OEM are lining up behind Intel or AMD. So, there is no reason to believe why both Intel and AMD will not sell more of their chips. What the OEMs choose (itanium, Xeon r Opteron) is a detail that involves so many parameters (performance, furture roadmap, Ability to scale up, manufacturing ability, positioning of their own competitor, ability to negotiate on price/suport, backward software compatibility, governmenet policies! etc etc)) and each of Intel and AMD chips win in some and lose in other. The most likeley scenerio is that both company will prosper when times are god, and both will go after one another fiercely during a downturn.
Maui.
Intel Analyst meeting :
http://media.corporate-ir.net/media_files/webcast/MS/2003/nov/intel/PSOwebcast.pdf
On one of the Itanium slide (more than halfway in the presentation) say more than 100k units (sold?) in 2003.
Overall both Barrett and Paul O are presenting a reasonably optimistic picture for semi (and Intel) growth.
foils at :
http://media.corporate-ir.net/media_files/Webcast/MS/2003/nov/intel/download.html
Check out the cpu cost trend. Lots of cost reduction moving forward. Realy good news, it seems.
Maui.
Joe,
from the article :
At Intel, spokesman Howard High says that Itanium is actually doing better than most people realize. The company will announce at its analyst meeting later this week that analysts' estimates of its sales "are off by an order of magnitude," he says. "This year and next are very big years for Itanium,"
--------------------------
This is quite a statement!. What the street estimate for Itanium ? 15k ? If Intel has sold 150k instead, that translates into 300mil additional revenue. Most of which might show up in Q4. Add to that the centrino craze and current low cost structure, Intel (and AMD due to flash and Opteron) can make a big splash next 4 to 5 quarters.
Maui.
Keith,
Re : "If we were to ship Opteron Hardware it would put pressure on Intel to bring out its Yamhill chip," he said. "As soon as they play that card, it makes it more difficult for them to sell Itanium-based systems."
This doesn't sound like a strategy to follow, if SUN wasn't afraid of Itanium.
From this article :
Multithreading, 24MB Cache in Itanium 2's Future
http://www.eweek.com/article2/0,4149,1382547,00.asp
In the second half of the year, the industry began seeing the adoption of Itanium crossing over from the high-performance computing space into the enterprise, fueled by a domino effect that began with Microsoft Corp.'s release of Windows Server 2003 in the spring.
------------------------------------------
Any idea when the report for the numer of servers shipped in Q3 comes out ?
Maui.
Keith,
re : "No, they´ve just confirmed what they have said for a while now."
I must have missed that. I hear all over the place how Intel has slipped their 90nm. But AMDs slip on 90nm didn't seem to get as much attention.
Edit : I just saw your other post, which mentions the slip.
http://www.investorshub.com/boards/read_msg.asp?message_id=1705597
Nonetheless, AMD pushed out by a few months the production schedule for chips produced with 90 nanometer process technology to the second half of 2004. The company had previously targeted a transition to 90 nanometer production in the first half of 2004.
---------
Maui
Wbmw, Keith,
here's the foil :
http://images.visualwebcaster.com/17861/19392/slide42.jpg
You can view other pages of the presentation by changing the slide numbers.
Lot of (good) stuff on flash. Not much on CPU. The key thing was more emphasis on server and less on desktop.
It talks about 90nm product piloting and yielding well. It mentions at a couple of places that 90nm product will release in 2004. The skeptic in me wonders if AMD is backing off the earlier 90nm date of 'mid' 2004.
Maui
HP plugs Itanium server gap.
http://news.com.com/2100-1010-5101062.html?tag=cd_top
From the article :
Itanium is on the brink of broader use, said Vish Mulchand, director of business-crititcal systems marketing. "We expect 2004 to be a very strong watershed year," he said.
Although Dell, IBM, NEC, Unisys, Silicon Graphics and others build Itanium servers, HP has the most at stake with the new processor. It's unifying several server lines to use the processor. While it's not clear whether HP will succeed in its goal of having the chip as widely used in the industry as Intel's Xeon and Pentium processors are today, it appears to have settled comfortably at HP, Haff said.
"Whatever questions there are around Itanium gaining broad acceptance in the market as a whole, there's no real reason to think it won't serve HP well as a PA-RISC replacement," Haff said.
-------------------------------------
of course, the author say HP started working on Itanium since 1988. So, that certainly hurts his credibility as a reliable reporter.
Maui.
Server Sales Surge for HP, Dell, IBM
http://story.news.yahoo.com/news?tmpl=story&u=/nf/20031031/bs_nf/22603
Winning the horse race was Hewlett Packard, which sold 408,000 units in the quarter, enjoying a 21 percent increase over last year.
HP (NYSE: HPQ - news) tends to have an advantage in the "units shipped" category, because it is the leading vendor of Intel-based servers. The market has been gravitating toward Intel-based servers at the expense of Unix (news - web sites) servers for some time, as the less-expensive machines grow ever more capable.
Intel-based servers are "starting to see greater and greater traction on the backside of Windows and Linux (news - web sites)," Forrester analyst Brad Day told NewsFactor. Additionally, "what you're seeing is much larger acceptance of Intel (Nasdaq: INTC - news) servers with 8 processors than ever before," he said.
---------------------------------------
Maybe its a sign itanium is getting traction now.
Maui.
Elmer, Re : "[Use extra metal layers] Why else do it?"
For extra performance. Chip designers seldom shy away from extra available layer. You can always use an extra layer to distribute/minimize cross-cap on a given ayer, allowing you to jack up frequency without functional failure. You can put more deCaps in, more routing to connect to repeaters, all allowing for higher performance.
If intel want to fight the performance battle only, with AMD, thats a card they can play and get on par with the number of layers AMD uses. Of course, like you mentioned before, it comes with a cost to yield.
Maui.
sgolds,
Hard to argue that 90nm is not late based on historic standard. Anythings 90nm release later than Q3 03, is more than historic norm. The point was, can anyone (AMD) else transition to 90nm in less time than Intel and narrow the gap of 6 months ?
About the absolote need of 90nm Presscott to counter Opteron, I dont know. It would be nice from Intels perspective to have presscott to counter 130nm Opteron. But, note that AMD has quietly backed off from the initial expectation that 130nm Opteron will ship in volume. There have been saying that the real ramp will be on 90nm.
Maui.
Re : I would say that 90nm at Intel is already late... 130nm started shipping in May 2001, and products were announced in July. It is now about 27 months after that announcement, and no 90nm announcement yet. Intel has stated 24 months between process introductions. 90nm is now late. Not really late, but it must be considered "behind schedule" by Intel."
What is the reference point of comparison ? 24 month cycle between process generation in the past ? Hasn't Gordon Moore mentioned more than once in the last few years that Moore's law will slow down ? Wouldn't that automatically translate into a longer cycle between 2 process generation (either that or the new process will not give as much head-room as the past processes). For someone (like me) involved in Physical Design, its a given that there is no easy process shifting at these processes and unless companies throw enormous extra resources to solve physical design problems (over and above the physics/chemistry problem), these 2 years cycle is bound to expand in the near future. The improvement in design tools and methodologies is lagging the increased process complexities significantly.
Maui.
Dan3,
Re : "On .18? Sure! Nice call, eh?"
not sure about your call. but, got to admire your ability to duck and slime around any mispredictions!
Maui.
Dan,
Re : "Trying to run a PIII core at 1.8ghz may not be as easy as you guys would like to think."
Do you remember your prediction a few years ago on SI, that P3 would not scale above 1.1GHz ?
Maui.
Article : Intel Caught Off Guard by AMD Release
http://story.news.yahoo.com/news?tmpl=story&u=/nf/20031007/bs_nf/22429
--------------------------------
And Rau questions AMD's overall strategy. "This segment (64-bit chips) alone is not enough to save the company. They're weak in mobile computing, for example -- and it's something they really need to address."
"The Intel-AMD war is over," Rau adds, "Intel won."
---------------------------------
Is this guy (Shane Rau) any good ? Sure is not an AMD fan.
Maui
AMD Athlon May Not be Path to Profits, Firms Report
http://www.reed-electronics.com/electronicnews/index.asp?layout=document&doc_id=126780&space...
Maui.
Re : " I haven't done better than break even on my AMD holdings, even as Intel shareholders have suffered."
How have you managed to stay break-even in an investment that has done more or less the same as Intel over the last 1, 2 and 5 years, at a time Intel shareholders have suffered ? Have you selectively chosen your timeframe to make that statement ?
Dont mean to be attacking, but your above statement makes me wonder if you are really interested in an honest, intelligent discussion.
Maui
sgolds,
I agree that the top management dont disclose leak information aquisation information intentionally, the clues are often in place for the middle management and employees to watch. The level of engagement of high management in day to day operations, the body language and the tone when issues of fund-raising are brought up, etc.
I have been through a projects where just about every one knew that the product did not make sense and the employees kept bringing this up. The management 'convinced' us why it is important, but we never really bought it (although we kept working hard (-:). Eventually, for about a month or two their energy level dropped and they started having 'clandestine' meetings. When the 'important' meeting to announce cancellation came, the only surprise was when the management found out that we were not surprised.
Maui.
Wbmw,
Re : "Which ones do you like?"
I am looking at buy oct 12.5 if AMD goes above 13. And I wont buy till later this month.
Maui
chipguy,
I was talking to an AMD employee last evening at the gym and he mentioned that they strongly believe at AMD that IBM would buy them out. Not sure if they are reading the same rumors/speculation that we all do or if it trickles down the management chain. Maybe that (and opteron announcement and overall semi demand improvement) has something to do with it.
Beside, even at $12 stock, it's not even a $5B company. With all their IP and 'potential' (and debt!), it could still rise with an overall recovery.
Having said that I am eyeing some AMD oct puts.
Maui
Lehman: Best Q3 Since 1999, Recovery Under Way
http://www.reed-electronics.com/electronicnews/index.asp?layout=article&articleid=CA321122&s...
Maui
Re : Can anyone give us any indication about what it would take to move K7 to 90nm, possibly with SOI?"
From physical design perspective, migrating data from 130nm to 90nm is a lot of work. The design rules (esp of lower layers like Poly, diffusion) are much worse than any previous generation and thus proess-shifting is not straight forward. OPC (optical Proximity Correction) begins to kick in a lot more requiring more effort. Cross Capacitance/Inductance and Noise issues would force lot more shielding. If the design is (re-)synthesized (as AMD is more inclined to do), it might throws off your timing convergence and can throw off your performance. And I haven't even seen the complexiity from SOI yet. Productivity can be improved somewhat, if area is traded off. But overall efoort required would not be insignificant.
Maui.
AMD revives Duron line
http://news.com.com/2100-1006_3-5067917.html
---------------
From the link :
The new Athlon XPs, meanwhile, are coming to market to compensate for delays to the upcoming Athlon64 lineup. The first Athlon64 chips are set to hit the street in September, but won't start coming out in massive volume until the debut of 90-nanometer manufacturing, Marty Seyer, vice president and general manager of AMD's Microprocessor Business Unit, said in an interview earlier this month.
---------------
Maui
Dan3,
Re : "IBM simply buying AMD"
I have wondered that myself. Here are my thoughts :
1. Why would IBMs competitor (HP mainly) want to buy chips from IBM anymore. Microsoft certainly would decrease support of AMD/IBM chips. Doesn't all of these isolate IBM and end up increasing Intels market share and hurting IBM in the long run ?
2. Hasn't IBM gotten badly hurt fighting Intel before ? Why would they want to fight Intel, in the area of Intels strength ? The fight between IBM/AMD and Intel could go on for years with each eating into others revenue. Why would IBMs shareholder invest in such a deal ?
About Microsoft buying AMD, again why ? Just because they have the money ? They have bigger fish to go after together (cellphones, PDAs, high end server market etc).
For AMD, life revolves completely around how Intel does. For larger companies like Intel, Microsoft, IBM etc, issues like, growing the overall IT economy, getting into new markets (most likely together) make a lot more sense than fighting each other.
Having said that, I do see why Microsoft, HP, IBM etc would want to have multiple supplier and would like a good competition between AMD and Intel. But, I am not sure any of them would benefit from buying AMD.
I would appreciate thought from you and others.
Maui.
Hi Keith, So, you had expectations that Opteron wouldn't translate into any meaningful ASP and revenue growth for years ?
What is your expectation with regard to opterons contribution to bottom line and ASP (sorry i dont have a link to your earlier post on that) ? 2006 ? 2007 ? 2008 ? Do you think AMD can afford that ? if yes, how ? More importantly, can you, as a shareholder afford that ?
Maui.
sgolds,
I also got the same impression from the CC that the growth to profitibility will come from participating in all markets (and hence entering new markets, where they dont participate now). If true, their tone of "Opteron will change the world" (though they didn't say that today) is not expected to translate into any meaningful ASP and marketshare growth for them for years.
Except that AMD is at a rock bottom price, things dont look good for AMD shareholders, IMHO.
Again, thanks for the great CC notes.
Maui.
sgold,
thanks for the notes.
Here are few more blurbs (paraphrasing, of course) I heard :
- Count on revenue of any significance through Opteron Sales only in 2004.
- FASL will add $20 mil in loss right now (will increase 200 in operating cost and generate 180m in revenue).
- Unit shipment down in Q2 and ASP down too (although slightly less).
- Expect modest, if any, processor revenue growth for years (Can someone confirm if I heard this right ? If true, they are really not counting 64-bit CPUs to help revenues growth for years.
Thnaks.
Maui.