AMD plans to build a multi-billion chip factory

[v_rr]

AMD plans moving forward
http://capitalnews9.com/content/headlines/111746/amd-plans-moving-forward/Default.aspx

AMD will work with HVCC to develop curriculum to prepare future workers
http://albany.bizjournals.com/albany/stories/2008/03/03/daily21.html?jst=b_ln_hl

[Calmatory]

with money? Seriously, can't they go bankrupt? Where do they get all this money from? They keep making loss over and over again every year, yet they plan on building and building more factories and then take even more losses, and keep rolling... Who has paid them so well?

[Marvin_The_Martian]

My guess is those who do not like seeing an Intel dominated market... could be the Euro's or the Chinese who knows lol...

And let's face it, they might loose money but they still have a good stake at the server market 2 up and more where intel just can't touch htt untill ics comes out. And their showcasing their 45nm procedure which they developed with IBM as partner ( another income source or atleast another party which carries it's own weight ).

[Periander6]

AMD plans moving forward
http://capitalnews9.com/content/headlines/111746/amd-plans-moving-forward/Default.aspx

AMD will work with HVCC to develop curriculum to prepare future workers
http://albany.bizjournals.com/albany/stories/2008/03/03/daily21.html?jst=b_ln_hl

Did you bother reading the articles? They have not made a final commitment.

"Certainly it would be nice if they were ready to commit," said Town of Malta Councilman Peter Klotz.

Looking at AMD's balance sheet, as well as the fact that they put their already delayed their Fab 30->38 conversion, and seem to have no difficulty filling orders with just Fab 36 running atm, this has about a snowballs chance in hell of ever being constructed. Right now the sole reason they are stringing this along is for political support from Spitzer and the New York delegation in terms of the anti-trust case. Meanwhile the New York taxpayers are being bilked out of hundreds of millions in taxpayer dollars to build infrastructure for a Fab that will never be built.

[Eldonko]

Im so happy I bought Intel and not AMD stocks.

[v_rr]

Did you bother reading the articles? They have not made a final commitment.

It´s news like many other news in this section. :rolleyes:
Why wonder why you don´t say that in every news topic opened....

[Vapor]

FYI AMD, most construction companies don't take Monopoly money. :p:

Time to grub more from UAE investors.

[dengyong]

Looking at the thread title, I thought they were constructing a factory out of billions of chips that no one wants. :rofl:

[Extelleron]

Looking at the thread title, I thought they were constructing a factory out of billions of chips that no one wants. :rofl:

Just because they aren't the best overclockers and enthusiasts don't want them, doesn't mean that no one wants AMD chips.

OEMs are gobbling up the tri-cores & Gateway seems to be using Phenom processors exclusively in their new PCs.

[Nedjo]

Looking at the thread title, I thought they were constructing a factory out of billions of chips that no one wants. :rofl:
Dude aren’t you glad that in your great country someone is opening more IT jobs??

[Scubar]

Dont you need to make decent CPUs before thinking about expanding your production line. As much as i love AMD their current dual and quad core CPUs are pitiful in comparison to C2D/C2Q. Partly the reason why i switched to Intel when building my new rig.

[thenrz]

Dont you need to make decent CPUs before thinking about expanding your production line. As much as i love AMD their current dual and quad core CPUs are pitiful in comparison to C2D/C2Q. Partly the reason why i switched to Intel when building my new rig.

If AMD was going for the performance crown, sure. They are in this price/performance trend, and it seems to be working.

[Spawne32]

AMD's new multi billion dollar factory

http://thymetocook.files.wordpress.com/2007/11/p1093633dt.jpg

"Because we cant build it with money!"

[superbowtie]

ahh great their running with The US economic financial approach. I bet pretty soon their going to start handing out chips to all those who can't afford them. :rofl: Universal Chip distribution.

But all joking aside I want them to Live and make good chips hopefully they do in this new factory.

[dengyong]

Needs more chips....chocolate chips.

[MrChubz]

Maybe AMD started taking-up making potato chips. Mayber each bag of chips can come with a free Phenom since they ran out of things to do with the CPU's they weren't selling.

I kid. Rootin' for ya' AMD>

[ryboto]

i thought the Arabian money was going to build this?

[JumpingJack]

My guess is those who do not like seeing an Intel dominated market... could be the Euro's or the Chinese who knows lol...

And let's face it, they might loose money but they still have a good stake at the server market 2 up and more where intel just can't touch htt untill ics comes out. And their showcasing their 45nm procedure which they developed with IBM as partner ( another income source or atleast another party which carries it's own weight ).

Actually, AMD hit a share high of about 24-25% and they are now down to the 12-13% range and the trend continues ... though B3 will likely help them stabilize it some.

[JumpingJack]

Did you bother reading the articles? They have not made a final commitment.



Looking at AMD's balance sheet, as well as the fact that they put their already delayed their Fab 30->38 conversion, and seem to have no difficulty filling orders with just Fab 36 running atm, this has about a snowballs chance in hell of ever being constructed. Right now the sole reason they are stringing this along is for political support from Spitzer and the New York delegation in terms of the anti-trust case. Meanwhile the New York taxpayers are being bilked out of hundreds of millions in taxpayer dollars to build infrastructure for a Fab that will never be built.

Well, Peri .. honestly, I don't see that they have much choice but to look for ways to finance it and make this happen. What I mean by that is a comment from a quarterly earnings conference (I think it was the Q4 06 report), Meyer said that (and I paraphrase) 'it has become to expensive in this business to stay small, we cannot survive as a niche player' -- this was in response to a question from an analyst asking why the agressive move to get big and not making money vs staying small and making money.

Meyer is basically saying, get big or fade away.... and he is probably right, AMD will need the cash flow the extra capacity can generate ... they just need a more compelling product to regain some pricing leverage. Besides, a 1 billion dollar check is hard to leave on the table for doing nothing more than building a fab (there's 33% of the costs right there).

[Periander6]

Well, Peri .. honestly, I don't see that they have much choice but to look for ways to finance it and make this happen. What I mean by that is a comment from a quarterly earnings conference (I think it was the Q4 06 report), Meyer said that (and I paraphrase) 'it has become to expensive in this business to stay small, we cannot survive as a niche player' -- this was in response to a question from an analyst asking why the agressive move to get big and not making money vs staying small and making money.

Meyer is basically saying, get big or fade away.... and he is probably right, AMD will need the cash flow the extra capacity can generate ... they just need a more compelling product to regain some pricing leverage. Besides, a 1 billion dollar check is hard to leave on the table for doing nothing more than building a fab (there's 33% of the costs right there).

To even come close to filling up this fab they would basically have to move up to a 35-40% market share, basically doubling their market share. How are they going to gain all that share at Intel's expense when they don't even have performance competitive products in most areas? Only with an apocalyptic price war that given their already broken balance sheet and Intel's near limitless reserves AMD has no chance whatsoever of winning.

Jerry was able to keep AMD afloat for decades without such a market share. If current management is incapable of doing the same then AMD is doomed. End of story.

[Shocker003]

Maybe AMD started taking-up making potato chips. Mayber each bag of chips can come with a free Phenom since they ran out of things to do with the CPU's they weren't selling.

I kid. Rootin' for ya' AMD>


We can trash AMD with all sorts of words for not coming up with a kickarse cpu. I totally agree with you guys but we will be doomed if they are gone from this race. Intel will name their next chip (Dracula QX9900), then suck our pockets dry. This can equally make Intel to relax thereby creating a vacuum in cpu tech = we might be runing the 65nm and 45nm intel chips for the next 25yrs:yepp: I own an intel rig, but am still waiting for the AMD come back to make my switch. The Athlon 64 cpu era was great and i hope they will still bring out an Ace chip.

[duploxxx]

Did you bother reading the articles? They have not made a final commitment.



Looking at AMD's balance sheet, as well as the fact that they put their already delayed their Fab 30->38 conversion, and seem to have no difficulty filling orders with just Fab 36 running atm, this has about a snowballs chance in hell of ever being constructed. Right now the sole reason they are stringing this along is for political support from Spitzer and the New York delegation in terms of the anti-trust case. Meanwhile the New York taxpayers are being bilked out of hundreds of millions in taxpayer dollars to build infrastructure for a Fab that will never be built.

pls give some concrete proof on there delayed fab 30-38 conversion.... fab30 is already offline for a long time now, i am sure they are already working on it.

[kiikkuja]

We can trash AMD with all sorts of words for not coming up with a kickarse cpu. I totally agree with you guys but we will be doomed if they are gone from this race. Intel will name their next chip (Dracula QX9900), then suck our pockets dry. This can equally make Intel to relax thereby creating a vacuum in cpu tech = we might be runing the 65nm and 45nm intel chips for the next 25yrs:yepp:

+1 On that. We are seeing this already in the GPU market. ATM the 9800GTX looks like big steaming turd IMO. People with older cards can match the performance or even get better results in benchmarks.

This industry NEEDS competition badly to go anywhere.

[Nedjo]

Meanwhile the New York taxpayers are being bilked out of hundreds of millions in taxpayer dollars to build infrastructure for a Fab that will never be built.

obviously you don't have clue how these tax things work! :rolleyes:

[donitsi]

AMD's new multi billion dollar factory

http://thymetocook.files.wordpress.com/2007/11/p1093633dt.jpg

"Because we cant build it with money!"

Aaahahhhaaa :ROTF: :rofl: Great! :D
That made my day

But i really hope AMD can pull this fab-rabbit from their hat
Would be good to their marketshare in 2012 -->

[Scubar]

If AMD was going for the performance crown, sure. They are in this price/performance trend, and it seems to be working.

Unfortunately it clearly isnt working, they are in a whole load of debt and have no products currently worth purchasing over an Intel counterpart.

[Periander6]

pls give some concrete proof on there delayed fab 30-38 conversion.... fab30 is already offline for a long time now, i am sure they are already working on it.

Let me clarify myself, the original plan was that they would continue production during the conversion. They changed that so that it would completely shut down during this year, not exactly suggesting they are hurting for capacity.

http://www.fabtech.org/content/view/3869/69/

AMD has made a significant change to its original plans to migrate Fab 30 in Dresden, Germany to 300mm wafer production and rename the facility Fab 38. The original plan was to remodel the cleanroom while undertaking a phased disassembly and removal of 200mm tools while continuing to have the facility produce microprocessors on 200mm wafers as new 300mm equipment was then installed and qualified.

In July 2007, M+W Zander FE GmbH was awarded the contract to undertake the conversion of Fab 30.

However, AMD executives said during a conference call with financial analysts last Thursday that Fab 30 would effectively be completely closed from a manufacturing perspective for the wafer migration work to take place. Little if any actual production activity is now expected from the facility in 2008, executives noted.

So right now they are meeting all their production needs with one fab, Fab 36. Fab 38 gets added next year. How are they going to fill that up, much less fill up a 3rd brand new Fab?

[GoThr3k]

after reading this thread i think there are some 12year olds active on this forum....

[Nedjo]

Unfortunately it clearly isnt working
how did you come up to that brilliant conclusion?:confused:

[BrowncoatGR]

Let me clarify myself, the original plan was that they would continue production during the conversion. They changed that so that it would completely shut down during this year, not exactly suggesting they are hurting for capacity.
Well maybe AMD deemed producing CPUs worth <150$ in a 200MM Fab at a 90nm node not worth the cost of keeping the facility running during the conversion...

[Pete]

AMD's new multi billion dollar factory

http://thymetocook.files.wordpress.com/2007/11/p1093633dt.jpg

"Because we cant build it with money!"

Nah it's built with all there left over stock and faulty stuff. They have an endluss surplie of it and will soon have more fabs that Intel but still producing jack . . . .

Face it AMD have lost and Intel are toying around. All we need is Sun to get on the ball and if they did it right could wipe AMD well and truly out and even give Intel a bloody good run as far im concered there is just about too much fanny farting going on at the mo

[Marvin_The_Martian]

Face it AMD have lost and Intel are toying around. All we need is Sun to get on the ball and if they did it right could wipe AMD well and truly out and even give Intel a bloody good run as far im concered there is just about too much fanny farting going on at the mo


You're so wrong on this in my opinion I would almost sig this so I can slap you with it in a year time :up:

Now if you're right, I would look foolish for a year.... that and my signature is big enough as is.

Not bashing your opinion, just stating mine btw :yepp:

And what has Sun to do with x86? If you paid more attention you would know the most likely candidate would have been Ibm not Sun

[***Deimos***]

:rofl: :rofl:
We can trash AMD with all sorts of words for not coming up with a kickarse cpu. I totally agree with you guys but we will be doomed if they are gone from this race. Intel will name their next chip (Dracula QX9900), then suck our pockets dry. This can equally make Intel to relax thereby creating a vacuum in cpu tech = we might be runing the 65nm and 45nm intel chips for the next 25yrs:yepp: I own an intel rig, but am still waiting for the AMD come back to make my switch. The Athlon 64 cpu era was great and i hope they will still bring out an Ace chip.

Oh dark lord, Intel-igence without bounds, master of the universe, devourer of galaxies, Krum2Duo the Konquerror, darkness everlasting I await your arrival. We are not worthy of your exhalted QX9900. Let the price not be measured in paper but by flesh and bone. We your loyal minions, slaves for billions of clock cycles, bound by your chipsets, guardians of your x86 software, are ready. Just say the words and it will be loaded, decoded, executed and written back into memory.

[LowRun]

You're so wrong on this in my opinion I would almost sig this so I can slap you with it in a year time :up:

Now if you're right, I would look foolish for a year.... that and my signature is big enough as is.

Not bashing your opinion, just stating mine btw :yepp:

And what has Sun to do with x86? If you paid more attention you would know the most likely candidate would have been Ibm not Sun

Play safe, just bookmark it :D

[Pete]

You're so wrong on this in my opinion I would almost sig this so I can slap you with it in a year time :up:

Now if you're right, I would look foolish for a year.... that and my signature is big enough as is.

Not bashing your opinion, just stating mine btw :yepp:

And what has Sun to do with x86? If you paid more attention you would know the most likely candidate would have been Ibm not Sun

Look closer, Sun and IBM are closer than you might like to think.

Oh and i am no fool, look at what ast 2 min years with Intel and AMD...who wins eh

[Marvin_The_Martian]

Look closer, Sun and IBM are closer than you might like to think.

Oh and i am no fool, look at what ast 2 min years with Intel and AMD...who wins eh

Pete, two people can have very conflicting ideas without thinking the other party is a fool :up:

Sun and Ibm are partners not the same company afaik.

[hollo]

well pete has a point... AMD are stuck on 65nm k10 this year while Intel have 45nm core2.
in '09 AMD have 45nm k10 but Intel will have nehalem.
they're not going to make a lot of money, if they make any, and they can only go on losing money for a limited period of time. what happens if they lose money all through '08 and '09?

Sun or IBM enter the PC-computer market with a non-x86 linux-based offering, that's what :hehe:

[Marvin_The_Martian]

hollo, press 'new posts' and look around, think you'll find information stating you're wrong in your assumptions. Amd is showcasing 45nm right now and planning on retailing them h2 this year ( iirc ).

[hollo]

well yeah, AMD should have early 45nm quads out this year. their goal will be to release them before the peak buying season, q4.
but overall they'll be a '09 product
not many people will jump onto the 45nm k10 bandwagon right before mainstream nehalem comes out in '09, and most of the 45nm k10 lineup will arrive in '09 too

[dengyong]

after reading this thread i think there are some 12year olds active on this forum....

And an 11 year old as well. :welcome:

[Donnie27]

And an 11 year old as well. :welcome:

:rofl: :ROTF:

I hope AMD sticks around all Geeks need them.

[Donnie27]

Let me clarify myself, the original plan was that they would continue production during the conversion. They changed that so that it would completely shut down during this year, not exactly suggesting they are hurting for capacity.

http://www.fabtech.org/content/view/3869/69/



So right now they are meeting all their production needs with one fab, Fab 36. Fab 38 gets added next year. How are they going to fill that up, much less fill up a 3rd brand new Fab?

QFT!

[SmaKKed]

SO it looks like alot here actually work for AMD and know what goes on in there books and behind the scenes? :rolleyes:
They might not be doing well but you still cannot run a buisness by saying "we are not making money now so we better not look to the future and we wont spend any money". Most buisness's when things are looking down will borrow money and try new things to see if they can gain ground again, not sit back and wait for it all to fall apart. Only time will tell on this, so sit back and wait and see.

[hollo]

you don't have to work for AMD to know what their financial situation and roadmaps are like
they need a new processor, not a new fab

[informal]

They will have not 1 but 3 new chips,called Shanghai/Deneb,Propus and Montreal.Last one is no brainer and will be consisted of 2x Shanghai in MCM to make it 8 cores in one package.This lineup will be good enough for 2009.

[hollo]

i bet you a dollar that the top-clocked octa-thread nehalem quad (server part) is faster than the top-clocked montreal in everything except floating-point heavy workloads :cool:

[informal]

"Top-clocks" and "8 cores" don't go hand in hand.I bet you a dollar that both Nehalem and Montreal 8 core devices will work at very low clocks(think 1.6-2Ghz maximum).Power consumption ,even at those lowly clocks should be quite high in full 8 core load scenario.It doesn't matter who has the upper Mhz(since the clocks would increase the TDP).WHat does matter is perf./watt ratio of said CPUs.Whichever has better perf./watt and lower system caost(board,memory etc.),will be a better chip.

[hollo]

that's kindof what i was thinking, but note i said 8-thread nehalem quad.
the TDP of the montreal will be high, and the clockspeed will reflect it.

i'm guessing the nehalem-quad may well beat the k10-oct in absolute terms thanks to its higher clockspeed and better core design with half as many cores

and of course perf./watt will massively favour hyper-threaded nehalem
system cost will... not

[informal]

that's kindof what i was thinking, but note i said 8-thread nehalem quad.
the TDP of the montreal will be high, and the clockspeed will reflect it.

i'm guessing the nehalem-quad may well beat the k10-oct in absolute terms thanks to its higher clockspeed and better core design with half as many cores

and of course perf./watt will massively favour hyper-threaded nehalem
system cost will... not

Wait for real figures from Hyperthreading before you call it a winner.I bet you a dollar that HT quad core Nehalem chip has no chance Vs the Montreal 8 core :).HT can surely help it,but to say it will best a real core with its "virtual" one is a long stretch at the moment.
Remember,Nehalem is Penryn base+IMC+CSI and hyperthreading.Also they added small L2s per cores with a big L3.

Montreal on the other hand is Shanghai based with HT3.0 connecting the cores.Still remains to be seen how efficiently 2 dies will communicate inside the package and between various packages(in 4 socket config. ie.)

[4X_FTL]

Let me clarify myself, the original plan was that they would continue production during the conversion. They changed that so that it would completely shut down during this year, not exactly suggesting they are hurting for capacity.

http://www.fabtech.org/content/view/3869/69/



So right now they are meeting all their production needs with one fab, Fab 36. Fab 38 gets added next year. How are they going to fill that up, much less fill up a 3rd brand new Fab?

Wouldn't it make sense at some point to bring the discreet GPU chip production inhouse and stop giving those dollars to UMC and TSMC?

[hollo]

Wait for real figures from Hyperthreading before you call it a winner.I bet you a dollar that HT quad core Nehalem chip has no chance Vs the Montreal 8 core :).HT can surely help it,but to say it will best a real core with its "virtual" one is a long stretch at the moment.
Remember,Nehalem is Penryn base+IMC+CSI and hyperthreading.Also they added small L2s per cores with a big L3.

Montreal on the other hand is Shanghai based with HT3.0 connecting the cores.Still remains to be seen how efficiently 2 dies will communicate inside the package and between various packages(in 4 socket config. ie.)

iirc intel said nehalem will be about 15% faster than penryn single threaded, and up to 200% faster multi-threaded. it was in a news-forum thread a while ago.

and i'd interpret the "200% multithreaded" as meaning twice as fast as penryn per core and per clock
ie the hyperthreading has been well implemented such that each "nehalem thread" will have about the same performance as a "penryn core" at a given clockspeed.

so, assuming i got those figures right, intel can be trusted, and i'm interpreting it correctly:
a 4-core 8-thread nehalem is twice as fast in a multi-threaded app as a 4-core 4-thread penryn
that makes a 2ghz nehalem quad equivalent to a 2ghz skulltrail setup.

let's say 45nm k10 is 20% faster than 65nm k10 per clock, and 10% faster than core2 per clock.
we'll have montreal at... 2ghz, maybe more, i think they might go over 2ghz if they're willing to raise the TDP
and that would be equivalent to a 2.2ghz skulltrail.... give or take a little depending on the app

and then there'll be... 2.5ghz? 3ghz?.. nehalem.
nehalem should win

well there's a few assumptions in there, but hopefully it's in the ballpark.
it'll be over a year before montreal comes out. maybe you'll owe me a dollar :D