I don't know guys.. I think its the right thing they are doing but for some reason the pit of my stomach says that its the end of AMD CPU game.
I had the same prediction for 3dfx.
Printable View
I don't know guys.. I think its the right thing they are doing but for some reason the pit of my stomach says that its the end of AMD CPU game.
I had the same prediction for 3dfx.
Just a reminder that TSMC has been shipping 55nm for a while now, you need to see the long term better, sure when AMD releases 45nm TSMC will be nehind but then we get 40nm and they are back on top again. I'll say they are quite close in manufacturing process but TSMC is far better as a Fab.Quote:
Originally Posted by Nedjo
Another Good read:
http://www.fabtech.org/editor_s_blog/_a/amds_gamble/
http://www.fabtech.org/news/_a/found...hics_business/
This pretty much shows what some people refuse to accept.Quote:
Originally Posted by metro.cl
The investments will be lower than previous, processnode advancement go slower etc. And this is even with tooling and construction.Quote:
Foundry Company’s 2009-2013 capital spending budget is forecast to be $4.8 billion (when using the mid-point of its announced $3.6-$6.0 billion spending plan). The funds will be used to install capacity to produce devices for AMD as well as to enter the IC foundry business. IC Insights believes that this financial commitment is uninspiring at best. It should be noted that AMD’s capital expenditures over this and the previous four years (2004-2008) are expected to total $6.4 billion, and look at what shape the company is in today.
Intel is expected to spend $25.6 billion dollars in capital expenditures from 2004-2008, four times what AMD will spend over the same time period. IC Insights believes that Intel’s capital expenditures over the next five years will total about $30 billion, which would be five times the Foundry Company’s most optimistic plan of investing $6.0 billion over that time!
Given that Foundry Company’s capital spending budget for the next five years is forecast to be less than AMD’s capital spending outlays over the previous five years, it is highly unlikely Foundry Company will help AMD gain additional market share. In fact, AMD will be lucky to sustain flat market share in the MPU segment. Moreover, a significant portion of the Foundry Company’s budget must be allocated to the new foundry business, and this, of course, will take away precious spending needed for MPU production!
The diference is that with ATIC the investments are made and start as soon as 2009. Without ATIC the investments where simply stoped because AMD don´t have money.Quote:
Originally Posted by Shintai
And as far as the deal ATIC will invest $3.6-$6.0 billion + pay 1.2B debt + 700K to buy 55% of foundry shares + 300K AMD shares.
In total ATIC will invest 5.8 - 8.2B.
being from Chile I presume that you didn't have chance to visit AMD fabs in some of press tours... unfortunately if you had been able you would have different opinion.Quote:
Originally Posted by metro.cl
:rofl:Quote:
Originally Posted by Nedjo
omfg that post nearly made me fall of my chair. :ROTF:
what a stupid comment in the fabtech article
Given that Foundry Company’s capital spending budget for the next five years is forecast to be less than AMD’s capital spending outlays over the previous five years, it is highly unlikely Foundry Company will help AMD gain additional market share. In fact, AMD will be lucky to sustain flat market share in the MPU segment. Moreover, a significant portion of the Foundry Company’s budget must be allocated to the new foundry business, and this, of course, will take away precious spending needed for MPU production!
as if this has anything to do with market share...the quality of a product is more important then the latest technology. K8 has proven this for many years.
and regarding the Intel investments.... if you would calculate the amount of fab space they require to produce there market share it would be normal that they have to invest more (not just for there huge amount of cpu's but also there crappy chipsets all over the place :D on there old fabs), amd fab36 on its own can take the current volume amd require even 1-2 year back when they had a higher marketshare. Not to mention the fact that intel has no co-operation like amd with ibm/samsung.
Only been to the HQ and as far as i know getting into a Fab is pretty much a no go for most of press, so i think is pretty weird that you have been there as press.Quote:
Originally Posted by Nedjo
Also been from Europe or US have you been to taiwan and looked at TSMC fabs?
bunch of EE press members was in the Fabs two and half years ago!Quote:
Originally Posted by metro.cl
http://pcekspert.com/articles/416-1.html
no I haven't been in TSMC's fabs. You?
@ Hornet331
what have you find so funny in my previous post?
55nm is the half-node of 65nm.Quote:
Originally Posted by metro.cl
The funny part is your post sounded very arrogant.Quote:
Originally Posted by Nedjo
Also have you seen anything else but AMDs facility? Just because AMDs facility impressed you doesnt mean its better or worse than the others. Since you havent seen any of those. So basing it all from what you have seen one place is...well lets say it wouldnt matter if Metro.cl have seen it or not.
Its like tasting one cookie and declare it the best in the world. Yet you havent tried any of the other cookies. Or even have a foundation of knowledge to base on if a cookie is actually good or not.
that never crossed my mind. I've just assumed most logical fact (knowing AMD's policy regarding the fab visits, and metro's geographical location)Quote:
Originally Posted by Shintai
completely agree, that was my point in quoting this statement by metro:Quote:
Also have you seen anything else but AMDs facility? Just because AMDs facility impressed you doesnt mean its better or worse than the others.
Well' this is far better answer to metro's post than my "arrogant" one:Quote:
Originally Posted by metro
again excellent quote to reply on metro's post:Quote:
Since you havent seen any of those. So basing it all from what you have seen one place is...well lets say it wouldnt matter if Metro.cl have seen it or not.
Next time I'll take you as my personal spokesperson ;)Quote:
Its like tasting one cookie and declare it the best in the world. Yet you havent tried any of the other cookies. Or even have a foundation of knowledge to base on if a cookie is actually good or not.
C'mon nedjo, you work for AMD. Say it now ;)
That is quite nice but mostly what i said only if AMD invites a bunch of guys (not sure if any other time they did the same) you can't enter it's fabs as press, we have asked for this since we travel a lot and the answer was imposible.Quote:
Originally Posted by Nedjo
And my post about TSMC been a better Fab is because as a fab you need to handle many needs from many customers in many diferent lines and sync that with your own needs and roadmaps. AFAIK AMD has never done this ;)
The fact is they never had to do this so we can't tell they couldn't :rolleyes:Quote:
Originally Posted by metro.cl
then if it is your first time you wont be the best or better than a fab with 20+ years of experience on this that is the point.Quote:
Originally Posted by AbelJemka
Why not?
Not like AMD is a rookie. ? More than 30 years making processor give you enough competence to deal with complex process.
Not for arguing Metro i think you can't be affirmative about your statement
What you say is exactly the point 20 years making CPUs doesnt make you the best making everything in semiconductor. You believe they will be the best i think they will need a few years to be in the game and see how good they are but i dont expect them to be the best any time soon.Quote:
Originally Posted by AbelJemka
Not arguing either here just diferent POV. I'll just add that most websites that are focused on semiconductor production think closer to my POV.
OK......I work for AMD............There you a Happy Camper?Quote:
Originally Posted by STaRGaZeR
I never say that but i believe with their competence and experience they will be able to climb to the top so i don't understand why everyone underestimate them so much :shrug:Quote:
Originally Posted by metro.cl
Uh?Quote:
Originally Posted by Brother Esau
Even AMD and ATIC aren't expecting themselves to be competitive for at least another 2 years. They plan to be competitive in the established semiconductor market by about 2010-2011, which seems realistic to me at least.Quote:
Originally Posted by metro.cl
It's true though that a bit of experience can help, although that does not have to make you better than the anyone else by default.
http://www.reuters.com/article/marke...0081007?rpc=44
Seems they forgive a little point?
aargh, what a load of BS in this thread. sad
lets just hope things turn out well, and some day AMD can scream out to the intel guys to BRING IT ON
You don't really understand how the foundry business works, do you ? You think its enough to build the latest and greatest FAB and you're king of the hill ?Quote:
Originally Posted by AbelJemka
Here's the revenue breakdown for TSMC :
>.5 micron around 4%
0,25/0.35 micron around 10%
0.15/0.18 micron around 23%
o.11/0.13 micron around 17%
0.09 micron around 28%
0.065 micron around 17%
Ooopsie.. 80+% is legacy ; looks like there isn't a lot of demand for a new processes like 45/32 nm.
No pool of legacy customers to move forward; need to build the entire customer design interface ; design tools and libraries.
That is anything but easy.
Source?
Just wondering the source, your sig somehow does make me wonder about things heh:yepp:
www.TSMC.com.tw dude really check the company websites and less Inq and Fudzilla.Quote:
Originally Posted by Rammsteiner
Q2 2008 report has those numbers on P7.
http://www.tsmc.com/uploadfile/ir/qu...esentation.pdf
Also UMC, check P12
http://www.umc.com/English/pdf/UMC08...entation-E.pdf
Chartered, scroll down
http://ir.charteredsemi.com/phoenix....ar&id=1179205&
lol, chill man:shrug:Quote:
Originally Posted by metro.cl
I just wanted the sources? Besides that, after other posts it ain't that strange Im asking for sources?
Anyhow, thx though:up:
I do not understand for sure but i believe that AMD understand.Quote:
Originally Posted by savantu
But I understand you may emit doubt about their capacity to manage but I do not understand why you doom them even before they start ...
The point is the foundry will not pace for the highest tech. Or even if, then the volume will be very low. Unlike AMD/Intel today.Quote:
Originally Posted by AbelJemka
If I remember correctly you have no relationship with AMD or new foundry what so ever. How then you conclude the strategy? Last time I checked this was not xtremeprediction board.Quote:
Originally Posted by Shintai
So what you say is. The foundry will do unlike ANYONE else. Rapid pace the latest tech and then? Go bankrupt in record pace? In the foundry business you keep using and reusing your equipment till its basicly dead. Also your customers rarely want the latest and greatest.Quote:
Originally Posted by Cooper
And the capital expendures laid out by ATIC shows exactly that. The foundry will use less money, including consutruction. Than AMD did in its previous years.
So honestly, there really aint much prediction in it.
So AMD don't want the latest and greatest?
You never cease to amuse.
Are we talking AMD now? Or the Foundry where AMD is a minority shareholder?Quote:
Originally Posted by Eson
AMD is the Foundry's only customer, of course they will have to listen to AMDs demands.
Thats only today. If thats the future then the foundry is doomed. They need more customers. Right now AMD only uses 80% of the capacity of FAB36. FAB38 is empty and the foundry wish to build another FAB in NY aswell as tool FAB38. So I really do hope they get other customers...Quote:
Originally Posted by Eson
yeah right! like we believe you ;)Quote:
Originally Posted by Shintai
anyhow arguments that some people on the Net are pointing out that fab contracting manufacturers are bound to produce majority of their product in outdated tech are simply shortsighted... hell dose self proclaimed experts know already all about new player in contracted manufacturing business that they don't bother to read press release, and they actually think that the name of the new company is "Foundry Co." :rolleyes:
Actually Shintai has a great track record on predictions. Just look at the signatures of other forum members here.. :DQuote:
Originally Posted by Cooper
Yep,No L3 cache in Core i7,HD4850 as fast as 3870...Quote:
Originally Posted by azraeel101
As for the prediction on the manufacturing front,we'll have to wait a bit on that :).By 2012 we will know how things stand with the new player.
2012 the favorite year of all doomsday conspirators.Quote:
Originally Posted by informal
Well i don't think it will be any different from 2010 or 2011,but everyone is free to think whatever they want ,it's a free world :yepp:
Source: http://www.thestreet.com/_yahoo/news...FREE&cm_ite=NAQuote:
Originally Posted by TheStreet
Damn he sure gets a lot of money from these kind of deals, something that I do not consider very logical...:rolleyes:
ATIQuote:
Originally Posted by Shintai
AMD will:Quote:
Originally Posted by Shintai
* Have equal voting rights with ATIC in The Foundry Company
just reminding :ROTF:
What happens when ATIC starts pouring money into the venture ? Do you think this status quo is anything more than a smokescreen ?Quote:
Originally Posted by eric66
Or when AMD says..ohh lets quickly advance to 32nm. And ATIC says..no. Lets milk this 65/45nm nodes.Quote:
Originally Posted by savantu
hmmm this one can happen but i don't think that arabs will stand on amds way they know that amd is much more experienced and probably will agree with what amd will sayQuote:
Originally Posted by Shintai
The arabs wanted a foundry. Not a black hole for losing money.Quote:
Originally Posted by eric66
Also even tho AMD got equal rights at the board. What happens when the arabs share is big enough to have a shareholder electrion and basicly put AMD out of power. Then the equal rights mean nothing.
This investment group is here to make money. And am sure they intend to do so and cheaply. They got a cheap 65nm factory and easy way to get a 45nm aswell with an empty and ready for equipment factory. All they need to do is harvest the fruit.
Its a fools dream to think that the arabs will pay the bill to follow TSMC/Intel in process nodes for the long term. If AMD couldnt make money with their current methods. Why would it make money split in 2? There will be one winner and one loser.
October 16th should also be interesting.
Hector Ruiz is the new President of the foundry company so it´s him that say what to do.Quote:
Originally Posted by Shintai
Also 32nm are on the route. It´s in the presentation on AMD site.
No, they have more money than they can spend and now they want braggin rights. Like, Dubai has all the skyscrapers and all, Bahrain has the F1 circuit, so they want something to brag about too. When they go to Dubai and the local sheik shows them the latest underwater restaurant with sharks as waiters, the Abu Dhabi sheik will say: "so what? See all those laptops and computers you use? I made them. I made the chips, I made the video things, yea, that's right, I iz tha Sultan of Chipz yo! In.your.FACE!!"Quote:
Originally Posted by Shintai
So yea, I think the Arabs care more about being known in the region as having the best technology (think about the fab said to be built in Abu Dhabi) than about profits. I mean, they surely knew AMD has big debts and won't turn really profitable in the short run. So, as I see it, they will rather spend ludicrous amounts of cash to get the best fabrication process than worry about profits.
If you would know arabs as I do, you would know why I say these things. They are worse than rappers when it comes to flashin and braggin.. :D
Yes, APIC will ignore it’s bigger client, and in the process not make a lot of money… ATIC will say no, just like chartered said no to SOI.Quote:
Originally Posted by Shintai
Even more because, as everyone knows it’s impossible to advance to a new process and milk the old ones, it's just impossible and it's really clever that you point it that out as the main reason for ATIC no going to a new process.
Not even when someone points you the obvious, you get it:
http://www.xtremesystems.org/forums/...7&postcount=64
Yes they aren't even going to build a new fab in NY to accomodate a new process node, why would anyone think that they will invest in new process nodes. :confused:Quote:
Originally Posted by Shintai
AMD is member of a consortium that had for two months now a functional 22nm SRAM cell, unlike Intel and TSMC. AMD being a member of that consortium means ATIC is also a member.
But you're right ATIC will not pay the bill to follow, they will pay the bill to lead.
Yes, owning 90% or 10% it's all the same, having equal voting rights or having no rights at all, it's all the same. :clap:, if someone tells different it's just a smokescreen.Quote:
Originally Posted by savantu
They can pour all the money they can, rivers of money, if AMD don't sell their shares, nothing will change, period.
Intel fans in panic... priceless.
Doug Grose is the CEO of the foundry. Hector is a chairman in the foundry.Quote:
Originally Posted by v_rr
There is no president in the foundry.
The question is if Hector is an executive or non exectutive chairman.
You pointed it out yourself. Look at what the arabs got and what they left behind. Including the majority of the debt and a failing CPU division.Quote:
Originally Posted by azraeel101
The arabs focused on the healthy parts.
This is what they wanted:
http://www.newglobalfoundry.com/
Not:
www.amd.com
hmm i think the same way they aren't interested in money that this company will make for them they are after the fame oil business is enough for them to make ridiculous amount of money :DQuote:
Originally Posted by azraeel101
This way they are also making sure that they still have a ridiculous amount of money in 100-200 years, as by then they will simply have some share in every imaginable part of the global market. That's their main goal, or at least I think that's their main goal, simply ensure that you will always make money by spreading your risks so to speak.Quote:
Originally Posted by eric66
Will you make up your mind already. A few days ago you were telling us that the foundry couldn't possibly compete with the older more established ones. Now the foundry is the healthy part? :shakes:Quote:
Originally Posted by Shintai
Do you mean this?Quote:
Originally Posted by BrowncoatGR
http://www.xtremesystems.org/Forums/...&postcount=136
Quote:
Thats only today. If thats the future then the foundry is doomed. They need more customers. Right now AMD only uses 80% of the capacity of FAB36. FAB38 is empty and the foundry wish to build another FAB in NY aswell as tool FAB38. So I really do hope they get other customers...Quote:
Originally Posted by Eson
Has I said NY fab will be build and go online very fast because AMD have prepared everything since 2006.Quote:
New 300mm Fab at the Luther Forest Technology Campus, New York
Construction is expected to begin in 2009 and is expected to be completed in roughly two years, with production beginning in the 2011 timeframe.
Maybe you should also provide a source to where you got that from...Quote:
Originally Posted by v_rr
You actually claimed that we would have volume production in 2010, so you where off by about 2 years. That quote of yours says that production will begin in 2011, which will equate to volume production in 2012 and that to me is quite a bit later than your 2010 claim.
http://www.newglobalfoundry.com/techsvcs.aspxQuote:
Originally Posted by Helmore
I never said volume production in 2010.
The fab starts to build in midle 2009 -> finish build and start producing in midle 2011.
Actualy I was very accurate because everyone pointed to 3-4=3.5 years and I pointed only to 1.5 years. Missed for 6 months.
I knew that construction will be very fast because since 2006 the deal is on the works. I was too optimistic by 2010 time frame but still the Fab in NY should be the Home of Buldozer by 2011?
Are there any probability that Fusion be delayed to be build in Germany when equiped with Bulk 32nm? Initially Fusion was thought to 45nm SOI CPU part and 40nm Bulk GPU part but it can delay to 32 nm Bulk CPU+GPU part.
The design is also a bit responsible for the clockspeeds, if they design something with a few hotspots on the chip, and some transistors get a lot hotter then their neighbouring trannies they will not reacht the same frequency. AMD just should add an extra step to the pipeline, allowing higher clocksQuote:
Originally Posted by LOE