Citigroup analyst Glen Yeung reiterates his Buy rating and $29 price target on Intel (INTC), citing channel checks that revealed strong demand for INTC's Atom processor:
"Field checks from Japan suggest Intel has ordered 20M Atom flip-chip packages for 3Q08 and 25M for 4Q08. This is orders of magnitude more than our modeled 2M and 2.5M units in that timeframe. These figures do sound aggressive to us (and some is likely inventory build) and so we believe a substantive haircut is appropriate. Nonetheless, even at half these levels, Atom is running well ahead of our expectations. We conclude that Intel's relatively aggressive 8.8% 3Q08 revenue growth guidance, while still prone to macro factors, looks more achievable in light of Atom strength".
I may have to give anecdotal evidence in support of Glen Yeung’s report. The lead time for ordering Atom processors from distributors have now gone up to 6 weeks! It's beginning to look like last year with Barcelona but only different. Atom demand has just gone through the roof as if every single technology firm has found some clever use for the low power microprocessor. Hopefully, Intel meets demand for the coming Christmas season. The concern about sales cannibalisation should be limited to computers below the performance range of Core 2 Duo, which as it stands today, equates to all of AMD’s SKUs.
8.07.2008
Strong Demand for Intel's Atom
Subscribe to:
Post Comments (Atom)
290 comments:
«Oldest ‹Older 201 – 290 of 290"There really is not another significant improvement in lens or numerical aperture on the horizon,"
This is true. I can not remember a node in recent history which has *not* had a lithographic improvement. Although, IBM/consortium is about a generation behind intel.
IBM/Intel
32nm/22nm = Significant double patterning
32nm = Immersion (rev2, 1.35NA)
45nm/32nm = immersion (rev1, 1.2NA)
65nm/45nm = 0.93NA tools
90nm/65nm = 0.85NA tools
130nm/90nm = ?? -- mass adoption of 193nm? -- can't remember back that far.
Sadly, after the rev2 immersion tools, there isn't a rev3. We'd really like a higher index of refraction liquid, but that's not coming soon.
The answer really is in the synthesis of the layout -- and seriously restricted design rules and either an explosion in layers or a significant amount of double patterning.
And nobody knows how to do it. Or if they do, they're keeping it secret for competitive reasons. =)
What does this mean? It means that 22nm is going to be HARD. And the companies which can tune a process along with a set of compatible design rules are going to have a HUGE advantage.
The problem is IBM just seems scattershot right now. 3 months ago they were talking up EUV as if that was getting closer to being a possibility for 22nm. Now they're talking up computational techniques. Tomorrow it will probably be a new medium for immersion or nano-imprint or some other "researchy" idea.
The problem with these announcements from IBM is that when they are this far out they are just noise, there is no ACTUAL DATA and there is no detail - yet it gets picked up by many 'tech' sites and those folks interpret it as something that could go into production near term.
I'm sticking with immersion extending to 22nm and then EUV/nanoimprint/some other exotic technology for the 15nm node.
The problem is IBM just seems scattershot right now.
I'm sure if you had the guys in Intel's development group each publishing on their individual projects Intel would seem to be disorganized too. The issue with these IBM announcements is that they lack context. IBM (and everyone else doing serious research) is looking at a number of possibilities, because no one knows which ones will pan out.
But IBM needs to be seen making progress, both from a company culture perspective and from a PR perspective. I suspect this is just these groups trying to keep their pet projects alive.
"The issue with these IBM announcements is that they lack context."
ITK - I guess this was my point... You have all these announcements, which are essentially activities and intermediate steps and then it gets circulated around in the press as if it actually means something.
Then you get the blogger/fanboys re-interpreting it to mean mainstream. Between EUV, vacuum airgap, computational litho, strain, SOI... you'd quickly forget the overall technology node gap has remained constant and Intel's process is outperforming even for a given node.
IBM's reserach is top notch - but they do kind of blur the line on these announcemets. My opinion is they leave these things vague intentionally so that they can be misinterperted - much like the games AMD plays now regarding introduction/launch/shipping to OEMs/sampling dates (when was the last time you saw an AMD roadmap with product available for purchase on date X?).
The 22nm IBM SRAM "announcement" was a classic example - they made no mention of an actual SRAM chip (i.e. probably a working 6 trransistor cell). They claimed <25nm gates! This is a 22nm SRAM...it should be less than 22nm! I wonder if it was anywhere even close to litho targets! If this was truly a 22nm SRAM the gates probably should have been in the 18-20nm range. There was no mention on planar transistor vs double gate/FINFET (which would theoretically be implemented around then given pat IBM announcements)... and suddenly this announcement is interpreted as if they are closing the gap /have caught up to Intel.
Lot's of bluster, little follow-though:
http://www.eetimes.com/news/semi/showArticle.jhtml;jsessionid=540P41TNONCESQSNDLSCKHA?articleID=210603477v
During the presentation, IBM revealed that it will not deploy high-k dielectric materials for logic chip production until the 32-nm node. It was widely believed that IBM and its partners, namely Advanced Micro Devices Inc. (AMD), could bring out high-k as early as the 45-nm node.
Any guesses? Perhaps the integration and yield was not as straigtforward as the "we've got high K at 45nm, too" announcement lead folks to believe. Perhaps the gate first was not simply an easy "drop in" to the 45nm process? (as I may have suggested not too long in the distant past)
Later in the article IBM claims they had it working on 45nm and simply chose not to do it. Apparently there is no truth to the rumors they used the same PR spin firm that AMD used when they claimed we "chose to release low clocked K10's at launch" (you know because customers were demanding it). Perhaps IBM's customers were demanding less efficient, non-highK 45nm solutions?
Seems rather curious - they had the option to put a 35% performance with 30-45% less power (at least that's what they claim), which is (theoretically) easy to put in as it is a gate first technology and they CHOSE not to do it?!?!? For the folks who believe this I have some shares in Fannie Mae and Freedie Mac I would like to sell you (I hear those companies are going to take off!)
AMD's quarterly report is due 3 weeks today - any guesses as to how it's gonna read?
Hopefully with the 48XX graphics cards and with K10 at where it should have been 10 months ago, plus the sale of some of the ATI assets, their losses will be under $200M for the quarter.
Considering the global credit crunch, I doubt AMD is gonna be able to float more debt at this point, so look for more layoffs and reduced R&D. Bobcat got killed off already according to the Inq, I wonder what is next...
For the P2P enthusiasts amongst us:
http://arstechnica.com/news.ars/post/20080924-thomas-verdict-overturned-making-available-theory-rejected.html
"Jammie Thomas is off the hook—at least for the time being. Judge Michael J. Davis has overturned a federal jury's copyright infringement verdict and award of $222,000 in damages to the RIAA. The verdict was handed down last October after a three-day trial and a few hours of deliberations."
anon: "Later in the article IBM claims they had it working on 45nm and simply chose not to do it."
That sounds like a baseball player who struck out to end a ballgame explaining that he could've hit a home run but chose not to.
Interesting - there are reports on significant peak power consumption on some of the high clocked AMD bins for 45nm (at least 'high' for AMD).
I think this may be the final nail in the coffin to the cliff that the 65nm process was on. The only data point really needed is to see peak power consumption on the lower bins between 65 and 45nm
If they dropped as significantly as the high bins did - then AMD probably significantly improved their process from 65nm to 45nm (or did some real good work on the design side with the shrink).
If, as I suspect, the lower bins show a marginal improvement, this would (in my view) clearly show that they were just running at a power cliff on 65nm (at around 2.6GHz) and AMD managed to push the cliff out a bit on 45nm.
I know there was a lot of speculation and theories on the board thrown out by numerous folks, but as the data continues to come in it looks like the predictions and theories turned out to be pretty close to reality.
45nm? Top bin quad core 3.0, maybe 3.2GHz (assuming highK is indeed 32nm). Power is hard to say but eventually the 3.2GHz may be a 125W bin.
Moose-
“Hopefully with the 48XX graphics cards and with K10 at where it should have been 10 months ago, plus the sale of some of the ATI assets, their losses will be under $200M for the quarter.”
That’s a tough one. They’ve got nothing to write down this quarter. (Well, except maybe the depreciating value of their 65nM tools) Last quarter they wrote down the only thing that’s making them money presently, ATI (888B). God forbid they should admit that Pheromones is a drop dead dog, let alone 45nM, Deneb, Shanghai, Triple Cripple, whatever.
About the Bobcat, there is this-----
http://www.xbitlabs.com/news/mobile/display/20080924115649_AMD_Bobcat_Processor_Still_in_Development_Despite_Rumours.html
It seems the AMD spin machine is still very active in any event.
The 3Q will be very indicative of how AMD’s asset light strategy, although not officially implemented, has run it course since last quarter. With that in mind, 4Q will be a very tough nut for them. INTC has October 19 price cuts just in time for the holidays.
http://www.hexus.net/content/item.php?item=15521
The article specifically mentions making room for Nehalem.
I believe the holiday season won’t be good for anyone although INTC has plenty of cash to weather this storm. For AMD however, with big credit in the toilet and very weak product lineup, banks (with money) are certainly not going to make the same mistakes that Lehman Brothers did. Getting 1.6 billion in cash isn’t going to be so easy this time around. Ironically, that’s the exact amount AMD holds in cash, currently.
(Hey ‘G’, do you still think the 1.2B NYS of Mania deal is still going to go through after this credit debacle?)
All this, plus the above process folks analysis, AMD hasn’t really done anything but cut their losses and put lots of lipstick on a very ugly pig.
Then there is 5.3 billion in debt on a company that’s only worth 3.17B. Tell me like I’m a two year old, what’s wrong with this picture?
SPARKS
Hey ‘G’, do you still think the 1.2B NYS of Mania deal is still going to go through after this credit debacle?
I'd still say slightly better than 50/50... Asset removal/smart/light will pawn off the debt and 1.2Bil in subsidies is still a 30% off coupon. Keep in mind if AMD splits (and of course if they survuve) and has a majority ownership of the foundry they may be able to use the cash they get from the minority stake they sell to get the ball rolling in NY (it's probably cheaper intermediate to long term then ramping F30)
I'm assuming the deal NY cut will survive the asset smart outcome - that is 'AMD foundry' will still be eligible (I have no clue if this is true but I have to think if AMD is the majority owner it would be).
Still thinking Intel will be in the high 20's by year end? As I think I've mentioned before Intel's stock price has little to do with competitive position with AMD these days and is more related to macroeconomics.
“Still thinking Intel will be in the high 20's by year end?”
Cute, real cute.
Allow me to preface my answer by saying I had no idea that the big bank bust would be this catastrophic. (WaMu, Et Tu?) I’m in pretty good company. Hence, the macro economic relationship you wisely suggested. But then, we both knew this all along.
Conversely, all the President needs to do is fart in the right direction (like today’s impromptu press conference) and INTC gets a 4 percent bounce while the NASDAQ market index takes a small hit.
Therefore, the short answer is no, not by the end of this year. However, given the companies position in product lineup, execution and mountains of cash, I can’t think of a better future investment vehicle in the long term, especially with the current market conditions.
Besides, I like you too much and am wise enough never to debate you, I’d lose. But in this instance, I think I’m on the money and I suspect you may agree.
As far as the NYS of mind is concerned, 50/50 is a long shot, and that's a lot of "ifs".
The state is close to 4B in debt. Risky ‘investments’ on a weak AMD whose position is in diametric opposition to INTC is not risky, it’s nonsense. I believe It will be perceived to be foolish in the not too distant future when they try to get a budget passed. The money ain’t gonna be there, as you say, if they survive.
Footing the bill against a powerful INTC is just plain stupid. Throwing good money after bad is just as ridiculous. Someone up in Albany will come to their senses, despite the millions of infrastructure already ‘invested’.
LILCO, LIPA, now this? Paaleeese, give me a f**king ba-rake! Why don’t we just build another nuclear power plant, say, in Garden City?
SPARKS
The state is close to 4B in debt.
4 Billion? That's less than AMD's debt! (and with a much larger revenue stream). 4 Billion just means a few less lap dances for the governor! Or a few more earmarks for Hillary should Obama win the election. (Yes I'm happy that some of my federal tax money is helping to refurbish an Opera House in Illinois and fund oyster studies in Delaware, but I digress)
Someone up in Albany will come to their senses
I think it is completely out of NY's hands right now - the offer is already out and AMD has until mid-09 to commit or decline it. I don't see how NY could pull it without legal or financial repercussions.
I think you will see AMD try a very sophisticated 3 card monte game around the end of the year with Asset light. The reason for the clock/deadline and "sooner rather than later" speak is, I believe, it needs to get done ahead of the NY subsidy option. You will see AMD sell a minority stake or raise capital via stock offering for the new foundry spinoff and use that money to seed the NY fab and get the payola. I'll reiterate if AMD were really smart they would settle the lawsuit with Intel for 500mil-1Bil and then the NY fab would be a done deal.
I'm not saying any of this will make AMD competitive but it would probably get the NY fab done.
“lap dances for the governor”
This gives a new meaning to the ‘immersion process’!
“I'm not saying any of this will make AMD competitive but it would probably get the NY fab done”
Marvelous.
Hmmm, raise capital via a stock offering based on the ‘new’ spin off. Hey, if a deal this bad sounds this good, it probably is, hmm, and it just might work.
Just for the sake of argument, let’s assume (when) they ram this fiasco down the tax payer’s throats. (Read: Me, Revenue steam, $120 this week alone in NYS IT). What process would the plant be built on, on what node, 32 perhaps? Who’s done the research, E-mails from INTC’s ex-employees?
Clearly, PDSOI is everything INTC (and others) said it wouldn’t be and more. Would they go back to a bulk process with Hi-k? Would they try a long shot for FDSOI? Sure, in the best conditions the 30 percent off coupon is great but who’s going to fund the rest, Arab Micro Devices, maybe? The banks aren’t going to do it, not now. Christ, Uncle Sam owns half of them anyway. Can they build a budget 32nM plant? (What a laugh!)
From where I sit (QX9770 kicking down the doors @ 4 gig, plus I’m going to be dethroned in November) AMD’s 45nM hasn’t really shown a significant improvement. In fact, the thing is still bleeding like a pig. And, at 3 GHz, it’s spinning pretty fast, but it ain’t going anywhere. How the hell can they keep going with loss after loss, quarter after quarter, with die the size of Cheese Nips that bleeds juice like 100KV mains and can’t compete with INTC’s 18 month old bread and butter chips?
Basically, what MAGIC can they make even if this son-of-a-bitch in Luther Forrest does get built? Or will this be another Dresden, all show and no go, which never paid for itself, let alone the other FAB that’s been mothballed. Hell, by the time they get it built INTC will trounce them with 22nM with the advanced technologies we’ve discussed previously.
Surely, money ain’t the only answer here, where’s the beef? Why aren’t those scheming lechers in Albany asking these questions?
Obviously, the Germans are getting the shaft, the NYS taxpayers are next! Hey, this is getting personal!
F**ken’ A bubba, I rather INTC, cut margins to nothin’, take a $10 a share hit, and crush the maggot/leeches where they crawl. The share price will come back, my f**king TAX DOLLARS ARE LOST FOREVER, GODDAMN IT!
I DON”T WANNA WORK FOR AMD!
…AND I DON”T WANT A COURT SETTLEMENT! FIGHT THE BASTARDS! THEY’YE GONNA GET THE MONEY ANYWAY AND IT WON’T BE FROM A SUPERIOR PRODUCT, THAT’S FOR SURE!
“G”, your logic, as usual, is impeccable, and I capitulate.
Now my nuts are really in a twist.
SPARKS
Screw it! I'm treating myself to a nice Christmas present!
CORE i7 EXTREME
X58 w/SLI!!!!! (I'm done with ATI)
TRI CHANNEL DDR3 @ 2000+
http://www.intel.com/consumer/game/index.htm?iid=gg_play+home_game
SPARKS
What process would the [NY] plant be built on, on what node, 32 perhaps?
That is hard to say. It depends a lot on the time line for actually building the facility. If they start in 2009, then figure it will come up in 2011, best case. Let's say 2012 on 22nm as a guess.
Another consideration is the 450mm wafer size conversion. Intel is pushing for that in the 2012 time frame. (I don't think they will hit that target because the equipment vendors are dragging their feet, but that is neither here nor there.) The smart move would be to size the factory to eventually accommodate the larger wafer size.
I really believe that AMD's whole motivation in building this fab is to get a facility close to the source of the R&D.
To understand why this is a big deal you need to look at Intel's copy exactly program. What doesn't get mentioned when people look at this is all the behind the scenes training. Engineers from the converting fab spend 6-12 months working in an existing fab on the new technology node before returning to their old fab to direct the transition.
Intel brings people from Ireland and Israel to the US for this all the time as well as shipping US employees between states. I don't think AMD can afford to relocate a large number of engineers from Dresden to East Fishkill for months at a time. I don't even know that IBM would allow them to work in their facility for that long. So the solution is to move your operation closer to the development fab where you can visit and observe a day at a time where needed.
Well, I did something yesterday I rarely do. I wandered over to AMDZone to see what they made of the whole Bobcat/Atom situation. If there is one thing this board lacks, it is a contrarian viewpoint. So on occasion I go looking for it.
While there, I took a look at their Nehalem vs Shanghai (or some such) thread.
I found a number of assertions by their resident self-proclaimed process expert that I though should be corrected. Since it is pointless to correct the misinformation there, I will do so here.
There are basically two issues to address. The first is the concept of yield and defect density. Next, we will cover Intel's process lead.
The statements were made that
Actually AMD should have a lower defect density because it is using a higher resolution process.
You are incorrect. Intel had a lower defect density at 65nm because of DFM. However at 45nm AMD has DFM as well but is also using immersion.
I guess the best place to begin here is to look at yield. Yield has two components.
First there is failure due to defects, typically small particles if you will that don't belong on the wafer. They cause shorts or block subsequent layers causing gross failures in the circuit. Missing features due to litho errors and poor filling of features by subsequent film deposition also fall into this category. So defects are things you can see with visible (under a microscope) inspection.
The second component is parametric failures. These are results of process marginality. They come from things like layers being too thick/thin, under/over dosing implants, metallic contamination, etc.
Particles are responsible for the vast number of defect issues. Both litho and DFM are designed to reduce parametric issues. So you can see that neither DFM or your litho processes have a big impact on defect density. Armed with that information you can see that claims that AMD's defect density issues on 45nm will be reduced by DFM and immersion litho are simply wrong.
Since bringing a new process online involves introducing new tools and new layers, there will be a learning curve to resolve the particle issues brought on by these new tools and process steps. For a number of reasons many of these issues don't manifest themselves until they leave development and move into high volume manufacturing.
Another issue with parametric yields that I won't address now is the idea of integrated failure modes. These are (usually) parametric failures that are caused by combining multiple steps of the process. DFM is intended to help with these, but these will be issues that AMD has to address as well as dealing with each process independently.
And for the record, AMD has always had some level of DFM. Every manufacturing process does. It isn't something they magically turned on for 45nm.
Moving on to Intel's process lead, another poster said
It is almost certain that Intel's 45nm process as of the Bloomfield launch will be more mature & have lower defect density than AMD's 45nm process at the launch of Deneb, given that AMD is just bringing out its first 45nm products and Intel has had 45nm CPUs on the market for about a year now.
The reply was
Well, that is a reasonable guess but Intel's process was still poor in early 2008. It's process time lead is far less than a year, in fact it's not even 6 months. Secondly, by early 2009 Intel's process maturity lead will be zero. Again, you are generalizing based on 65nm and this is incorrect. BTW, my information on 65nm comes from Intel itself.
I wonder when Bob Baker and Bill Holt (Intel's manufacturing VP's) put the folks at AMDzone on their mailing list. They forgot to include me.
Actually by late 2009 Intel will be bringing 32nm on line. Compared to late 2008, early 2009 for AMD's 45nm. That puts Intel a little over 1 year ahead on process nodes.
There were also these comments:
No. AMD launched 65nm a year after Intel but even though they converted quickly they continued to struggle with the process up to the present. 45nm won't be like that. AMD's defect density should be considerably better at launch than Intel's was in early 2007. Again, the lead is not 12 months.
That is what you do not understand; AMD won't need a year to "get the kinks out of its process". 45nm for AMD will not be like 65nm. You are still incorrectly trying to project 45nm based on 65nm when things have shifted dramatically. With phase shift masks and double masking 45nm is more difficult for Intel than 65nm. In contrast, with DFM and immersion 45nm is actually easier for AMD than 65nm.
I haven't seen any recent yield data published by Intel, but the last thing I saw showed 65nm yields (at least the defect portion of the equation) looking better than 90nm. 45nm was just getting ready to ramp and looked comparable to 65nm. So I'm not sure where this assumption comes from.
I have never seen AMD give defect density data that is as easy to interpret as what Intel provides. If it is out there, I'd welcome a link. So claims that AMD will have better yields at start up than Intel (or vice versa) are pure speculation.
Arguments that yields for one company a year into a process node are better than yields at the other company at start up are also speculation. But it is speculation based on a well documented fact. Process yield improves over time. This is true for both AMD and Intel.
The one statement that is correct is that Intel had some new challenges in patterning. But what everyone keeps overlooking when they bring up double patterning, is that you don't do it on all the layers. There are only a hand full of steps that are "critical" to device performance. And you only double pattern on those. So this only affects a few steps on the process. As such, the impact to the overall process is limited.
Sparks said...
God forbid they should admit that Pheromones is a drop dead dog, let alone 45nM, Deneb, Shanghai, Triple Cripple, whatever.
We should know more in a little over 2 weeks. I'd be curious to see how well Puma is doing, since Scientia used to tout that as the next huge AMD win over Intel. Of course, he's been touting (or tooting :) for quite some time, and so far the AMD train hasn't left the station..
Speaking of idiotic AMDZonerz predictions :), this from DailyTech:
Netbooks capture 9 of the top 10 notebook sales spots on Amazon.com
Intel's Atom processor has become the darling of the netbook market. Intel's 45nm processor runs at 1.6GHz and provides adequate performance for everyday computing while keeping power consumption very low.
Intel's Atom is also attractively priced so it has become a must-have processor for manufacturers looking to jump in on the $300 to $500 netbook market. Nowhere is this more evident than on Amazon.com's laptop sales leader board according to Mac Rumors.
Netbooks currently occupy nine of the top ten spots on the sales chart. The only non-netbook on the top 10 list is the 2.4GHz Apple MacBook which rings in at $1,203.98 after a $75 rebate. On the other hand, the netbooks on the list range in price from $329.99 for the Acer Aspire One and ASUS Eee PC 900 16G to $479.99 for the ASUS Eee PC 1000H.
All of the netbooks on the list -- with the exception of the Eee PC 900 -- use Intel's 1.6GHz Atom processor. Of the nine netbooks on the top ten list, seven of them are running Windows XP Home.
If we expand the standings to include the top 20 notebooks, netbooks occupy 16 of the top 20 spots.
The brisk sales of Atom-based netbooks seem to indicate that Intel is making inroads on increasing production of the popular processor. DailyTech reported in late August that some manufacturers have been forced to use Intel's older Celeron M 353 due to Atom shortages.
I'd guess Robo is about 99.9 times smarter than Ah-Ben-Stoopid who keeps insisting Atom is a failure. Just goes to show Abinstein = EPIC FAILURE. :)
What process would the [NY] plant be built on, on what node, 32 perhaps?
Intheknow had it.. a few things/detail to add if interested. Speculative timeline:
- AMD exercises option sometime in '09 (this will likely have to be after Asset Removal is implemented), but has to be before the option expires mid-09 (unless NY extends it)
- Assuming they break ground in late '09, figure about a year to get the shell (foundation, walls, major structural, facility elements). This puts the shell "done" in H2'2010
- The equipment move in is the schedule swinger. If AMD is really driving the schedule hard they can actually start moving in equipment while the shell is being finished up - though I would find this unlikely. Figure ~1 year to install and qualify pilot line equipment. Then figure another 6 months to certify the process. This would put production ready likely early 2012 best case.
Given 2012, I would say AMD would follow the Dresden approach - startup on 32nm in low volume (this gets equipment qual'd and matched and verifies the fab) and then likely transition to 22nm - this obviously is dependent on 22nm being on schedule (and if it isn't they could just do more 32nm volume).
I really don't think the NY fab really has much to do with being close to development at IBM - that may be a small bonus, but frankly this is all about the Benjamins.
And Sparks what I suggested about capital raising is not based upon sound business strategy, but survival. At some point AMD will need to retool to stay on the technology wheel, and it's either pay for all of it (retool Dresden) or some of it (NY subsidy).
It is painful to retool a single fab for every node - as much as you plan for no impact you have equipment move-ins/move-outs, facility work and production does get impacted. Intel abandoned this approach and generally skips a node the 90nm fabs go to 45nm the 65nm fabs to 32nm, etc. Intel can do this because they have a need for the lagging capacity via chipsets and old/low end CPU's. This is why if AMD is going to stick with fabs they need to bring the graphics and chipset production in house.
It is painful to retool a single fab for every node - .... Intel can do this because they have a need for the lagging capacity via chipsets and old/low end CPU's.
That has been the model up to now, but Intel is beginning to move away from that model. With repartitioning of the CPU and chipset and the reduction of the chipset from two chips to one, the waterfall approach will no longer work as well.
Going forward, I believe Intel intends to focus on a greater degree of tool reuse for each node. I believe it was Bob Baker who articulated the new approach. I'll have to see if I can dig something up on it.
BTW, I hadn't considered having NY effectively eat part of the retooling cost. You would think that Saxony would have been willing to work some sort of a deal to keep AMD's operation going there. Maybe the deal just wasn't as juicy as what NY put on the table.
You would think that Saxony would have been willing to work some sort of a deal to keep AMD's operation going there.
AMD got a subsidy for F30 (200mm) when they moved there, they got a subsidy for F36 - while I'm sure there may be tax incentives offered, it is difficult to justify subsidies just to retool an existing plant. This is not like a new plant where you are creating new jobs... this would amount to pretty much just keeping the existing job base there.
My guess is this would be a hard sell to the German people for 1.2Bil. "Give us 1.2Bil and we will basically just continue to employ people and probably not create any new jobs?"
When AMD got the F36 subsidy the unemployment rate was 15.5% in Dresden. Also the job benefit of adding a new plant/capacity is nowhere near the benefit of the first fab built as a lot of the people infrastructure can be shared and the job creation is mainly just the technician/engineering support for the fab. F36 at the time was projected to add a mere 420jobs (would you pay 1.2Billion for that?)
The initial F30 subsidy was ~0.5Bil and F36 was somewhere around 0.7Bil (and I think they may have also provided some favorable loan terms as well)
ITK-
I deliberately left out the 450mm variable from the wretched NYS/AMD equation simply because AMD isn’t part of three member consortium, TSMC, Samsung, and INTC who was behind the 450mm move a few months back.
AMD was conspicuously absent. Further, if memory serves, one of the members is getting cold feet, and it’s not INTC. Additionally, AMD cannot afford the price of entry to 450mm club. (Perhaps NYS can?)
More importantly, all you guys didn’t offer a guess on what process.
PDSOI ain’t cutting it, and it ain’t cheap.
Wanna offer a guess?
Moose- The Doc is brilliant. Ya gotta give ITK credit, too. He’s been pounding the pavement, and beating the drum with this ATOM thing since day one. As for me, I’m not that smart. Being on the lunatic fringe, if it ain’t got a 1kw PS, I just look to the left and cough.
Seriously, when I get my hands on one of these things and give it a whirl, I may just buy one for my kids. If it’s that good, maybe I’ll keep one in the Gang Box, merely to keep track of hours, materials, change orders, etc. At these prices, they’re cheaper than most of the hammer drills my company gives me to drill holes into concrete.
As far as AMD’s 3Q numbers are concerned, I can’t wait.
“(would you pay 1.2Billion for that?)”
I’ll answer that! (You love to needle me) Peh, hey, if you’ve got a “revenue stream” that includes NYC and its’ five burrows, Nassau, Suffolk and Westchester counties, Cops, Teachers, Fire Fighters, Powerful Trade Unions (including mine), and an 8.5% sales tax, of course you would, “A mere bag of shells”, I say.
Besides, Hillary probably thought that this was a good idea, ‘let the boys up in the sticks have their little pet project.’
Really, what the hell else do they have up there, IBM and a bunch of prisons? As you know, with that genius clock work ticker, WE have been carrying the state since the British named us.
Ha! AMD’s employees are in for a rude awaking! It ain’t no Sunnyvale or Austin up there, especially in the winter! Dress warm folks, and buy a snow blower.
Oh yeah, Wrecktor apparently likes the ponies, and we all know he loves to gamble.
SPARKS
Besides, Hillary probably thought that this was a good idea, ‘let the boys up in the sticks have their little pet project.’
When she ran for Senate she made a very big deal about job creation in upstate NY (which I believe at the time was really hurting and the huge swing vote in the NY Senate democratic primary) and promised improvements. When questioned on this when she ran for the democratic presidential nomination she said well she couldn't do it because there was a Republican president (funny she didn't mention that excpetion/qualification at the time she promised upstate job growth).
So now you have the AMD subsidy - which is an attempt to throw a bone to another NY politician failed promise. It's just a shame there is this much money being thrown at a, how should I put this, 'questionable' job creation idea. As I mentioned above, AMD capitalized on Germany's economic and unemployment problems for the F36 subsidy - you can't really blame them for moving to where the money is.
As for 22nm process... it's difficult to say - if I had to guess I would lean toward FDSOI especially if IBM employs a double gate/FINFET approach. 450mm is not even a consideration for AMD at this point - best case in general for 450mm is the 22nm node and this would be limited to the 'big boys' (Intel, Samsung, maybe TSMC) and is highly unlikely.
“you can't really blame them for moving to where the money is.”
Agreed, further, let’s not forget AMD will be getting nearly free legal counsel in the shape of the State Attorney General’s office, at the expense of the Taxpayers, of course. Call it a fringe benefit.
Naturally, Andy Boy had a good teacher, Big Daddy Mario. Hence--------
“which is an attempt to throw a bone to another NY politician failed promise. It's just a shame there is this much money being thrown at a, how should I put this, 'questionable' job creation idea.”
Question: How do you spell Shoreham?
T-W-E-L-V-E B-I-L-L-I-O-N
That’s how I spell it, in 1980’s English.
I’m afraid, very afraid; this 1.2B is only the beginning. In fact, my original premise asked, “Why aren’t those scheming lechers in Albany asking these questions?” regarding AMD’s solvency and technological position. As you so succinctly pointed out, they don’t know, don’t care, or both. They see a fancy FAB, a busted 300mm wafer, a couple of clean rooms, and they’re awe struck. They slither back to the Limo and say, “Man we’ve got to get one of these; Think of the revenue”
Hey, to you process magicians I’m an educated idiot, those guys are just plain idiots. Christ, they wouldn’t know a Tri-Gate from a Hell Gate or FDSOI from the FDR Drive. That’s a fact.
SPARKS
nonny: "I'd guess Robo is about 99.9 times smarter than Ah-Ben-Stoopid who keeps insisting Atom is a failure. Just goes to show Abinstein = EPIC FAILURE. :)"
I think that his claim was based on Atom being too power-hungry for the small devices it will eventually be aimed at, and too performance-poor for general computing. I think that Intel itself realized that this was not a world-beater, just the first iteration of a device that they have big plans for. The fact that it's become such a hot seller so far is just a bonus, IMO.
Abinstein also implied that the high sales are a product of pressure from Intel. While I am willing to believe that companies like Intel take advantage of their position when negotiating with customers (to put it diplomatically), it's hard to believe that they'd even think of leaning on companies to use Atom. With so many countries investigating them for anti-competitive practices, does it really make sense for them to risk further sanctions in order to sell a few more Atoms? That doesn't make any sense to me. None at all.
I still think many of the comments there are more "this is what I wish was happening" as opposed to balanced and realistic analysis. I suspect that the talk about how close AMD is to Intel in process development falls into this category as well.
I still think many of the comments there are more "this is what I wish was happening" as opposed to balanced and realistic analysis. I suspect that the talk about how close AMD is to Intel in process development falls into this category as well.
The thing that sticks in my craw is the way that process claims are made in such an authoritative manner. For a poster to definitively claim that another person is wrong when that poster has no real knowledge of the topic is asinine. I find the general level of posts on process related issues to be of very low caliber on that blog.
The whole reason I occasionally import junk from there to this blog is to correct the most egregious errors. Since there are individuals who read both blogs, my hope is that they will be able to at least find accurate data here to offset the misinformation from the other blog.
450mm is not even a consideration for AMD at this point
I don't disagree, but if I were going to invest billions in a fab knowing that 450mm was in the works, I would want to future proof my investment. All I'm suggesting is they should size the fab to allow installation of 450mm tools at some future point. Retrofitting an old fab for a larger wafer size seems to be only marginally cheaper than building a new one.
Sparks said As far as AMD’s 3Q numbers are concerned, I can’t wait.
Thanks to the bailout bill fizzling in the House, AMD's stock is now blow what Sci refers to as it's all-time low - $4.31. It's dropped over 16% today. In contrast, INTC is down less than 6%. Funny, Apple and Google are also way down today...
Tonus said think that his claim was based on Atom being too power-hungry for the small devices it will eventually be aimed at, and too performance-poor for general computing.
Ah-Ben-Stoopid rarely passes up an opportunity to slam Intel, so he turned it from being unsuitable for his purposes to being a total failure, which obviously it is not since it is selling so well. And I agree that it wouldn't make sense for Intel to push the Atom, when they can't even make enough to satisfy current demand. If Intel ever releases the low-power chipset then I think they'd have a real winner esp. with the dual-core Atom. I seem to recall a review on Tom's or maybe DailyTech where the Atom compared very favorably to the competition, unexpectedly so.
All I'm suggesting is they should size the fab to allow installation of 450mm tools at some future point. Retrofitting an old fab for a larger wafer size seems to be only marginally cheaper than building a new one
It's a bit hard to do this at this point in time as there are no real standards (SEMI/Sematch/consortia) yet. The longer people drag their feet and continue to push the 300mm prime garbage, the more likely it is that the early adopters (Intel, Samsung) will work directly with equipment manufacturers and start establishing it themselves.
One of the biggest challenges on 450mm (other than initial costs) will probably be the handling and queueing. There certainly won't be 25 wafer lot sizes (it will be 12-13 or fewer) which will mean queueing will be an issue - that is getting enough wafers staged near a tool. This may involve a departure from the stocker models and a need for local "small stocker"/queue stations near especially high throughput tools.
Additionally the delivery systems will likely change significantly (there are both floor-based and overhead approaches being explored) and that may be one of the biggest difficulties in converting a 300mm fab to 450mm without essentially gutting it.
At this point, I don't see that there is a lot you can plan for if you are close to building a 300mm fab. In AMD's case (assuming they are successful) - they are probably just best off planning to build out NY fab at 300mm normally and then completely gutting F30 or F36 if/when the time for 450mm for them comes (or look for some more money from Sparks and friends and build another fab!).
This kind of conversion is not like a tech node conversion and I'm not sure how effectively it can be planned for or done on the fly (meaning still running 300mm production while you are converting some of the fab to 450mm). 450mm is so far off it really is probably not worth any calories of thought for AMD. You can alwas 'plan', but what would the plan be based on when there are no even marginally definitive specs for 450mm tooling/wafer delivery
Moose-
Regarding AMD’s price drop----
“It's dropped over 16% today. In contrast, INTC is down less than 6%.”
Couple AMD’s long term debt, presently 5.3 billion, with a VERY tight credit market I’m sure those who dumped AMD’s shares were very concerned about AMD’s sustainability. Lehman sold the 1.6B note to AMD in August 2007. Given the current credit crunch, I believe future such opportunities will be very few and very far between.
http://www.physorg.com/news141053978.html
In fact, AMD is holding in to its present 1.4B in cash tooth and nail. More importantly, I’m sure some thought that the 700B government bailout money would be a factor in maintaining AMD’s solvency. Obviously the market reacted to the contrary. The closer we get to 2012 the uglier the debt looks, considering it’s achieved junk status (CCC-) already.
From the same article------
“AMD used the cash infusion to pay down debt from its $5.6 billion acquisition of graphics chip maker ATI Technologies and for other corporate expenses.”
That pay down was only 300 million, peanuts.
Add this to AMD’s “perfect storm” rationalization/explanation on 1Q 2007.
http://www.crn.com/hardware/199200068
http://www.theregister.co.uk/2007/04/23/amd_ruiz_head/
AMD’s claim to achieving “operating profitability” at the end of this year is a stretch. Any way you cut it, it is AMDese speak for still loosing money overall. It’s still lipstick on a pig as far as I can see.
INTC’s Xeon 5xxx series has dominated the large and lucrative 2P market.
http://www.theregister.co.uk/2008/09/08/intel_new_faster_and_greener_xeons/
In the final analysis, there a few more clouds looming on AMD’s Perfect Storm horizon.
INTC’s scheduled Nov 19 price cuts.
A large 3Q loss.
Bloomfield
Failure to reach “operating profitability” at the end of 4Q.
Basically, they are going to need more money and it might not be there. And, keep an eye on the cash on 3Q #’s.
SPARKS
Recent finance mess means AMD goose is totally cooked. Nobody not even the Arabs are lending money any more. AMD is in need of lots of financing to continue, let alone how they will get more to build a fab.
I also can't believe in this political enviroment that NY politicians would float a loan to a company going BK and ask the taxpayers to foot another bill.
I also can't believe in this political enviroment that NY politicians would float a loan to a company going BK and ask the taxpayers to foot another bill.
Folks seem to be missing the boat on this one... this is not a loan it is a grant with conditions. There is no payback of the money - the theory for NY is it brings in direct and indirect jobs and tax revenue and NY eventually recovers the money over time... regardless of whether this theory is right or wrong, it has nothing to do with AMD paying anything back.
People also seem to think that this was done recently and can be pulled back given current conditions. The offer was made well over 18 months ago (back in '06) and the ball is in AMD's court until July'09. Perhaps the reason 'it is hard to believe this offer was made in this environment' is due to the fact that it wasn't made in this environment, it just happens to be approaching expiration soon and thus people are confusing the expiration timing with the timing of the offer.
"I’m sure some thought that the 700B government bailout money would be a factor in maintaining AMD’s solvency."
The bailout is a fund that allows the government to buy financial instruments such as CDO's and the like. It would have very little direct impact on non-financial firms.
It would have potentially an indirect impact - at least that is the questionable (in my view) theory. This theory is that once the financial institutions can clear some of this nearly untradeable paper (who wants to buy financial instruments built on risky loan mortgages with a rising forfeiture rate and declining house values?), they will suddenly start making "good" loans again.
The problem is two-fold. One - 700Billion is still a drop in the bucket and while significant, it's significance is based more on the fact that this will prevent these assets from going to zero worth, and if the US gov't starts buying it, maybe some other folks will as well. And in the long run if/when housing recovers, the gov't may even be able to re-sell the paper at a profit (they just need to sit on it long enough to let things unwind)
The second problem is the assumption is that financial firms will just start lending again and credit will magically loosen up. This is unlikely after seeing several banks and firms go under, you will see much tighter standards and worse terms (rates). Companies like AMD will have a lot harder time borrowing regardless of the bailout and if they manage to get the credit it will come at a much higher cost.
The other issue with the belief that credit will loosen up once the paper is off the books is leverage and regulation. Some of these firms were levered at 30:1 or higher, which is unlikely to be allowed anymore - especially as firms like Goldman who have announced they are becoming a bank which means they will no longer be allowed to operate at those higher leverage ratios. WIth lower leverage there will be much less lending available (instead of $1 in assets being able to support $30-40 in loans you now will be able to loan much less)
As such I'm glad the bailout failed (though I fear the impact to the stock market). If you think the Congress is better able to set the prices on assets that are completely illiquid, when they can't even count the # of votes that they have for a bill, then perhaps we should let Congress run firms like Intel as well! Congress is no better equipped to spend my money then a cashier at the gas station near me (which is why that person collects it and give it to someone who knows what they are doing).
“Companies like AMD will have a lot harder time borrowing regardless of the bailout and if they manage to get the credit it will come at a much higher cost.”
This is/was the point I was driving at, and I said “some thought”, not me. However, I’ve said it before and I’ll say it again, as most contributors on this would agree, AMD’s days of hand outs and free passes are over.
The press has stopped blowing smoke up their readers asses (even Fuddie is a bit more objective). Embattled financial intuitions will be looking at the fundamentals with a more objective/cautious eye than they’ve done during the past few years. Buy and hold recommendations on AMD during 4Q 2006/1H 2007 were absolutely criminal. They had nothing. I knew it, and I’m a nobody from Podunk. How could I be so right and the pretty boy/power brokers be so wrong?
After we get over this catastrophe, let’s hope performance will be king and the weak and the reckless will die, the way they’re supposed to. Sound recommendations and credibility will be paramount to the re-emerging/re-regulated financial institutions. Remember the Ad, “We earn our money the old fashion way, we earn it”?*
I can remember when banks and big money signified restrained conservatism, not wild eyed speculation. I will be very happy when they dump the loose cannons overboard with the rest of the bilge.
Smoke and mirror tactics, marketing ploys, and “3 card Monties” simply won’t cut it in this new environment, now and in the future.
Hell, they don’t even refer to AMD as ‘The Scrappy Little Company’ anymore.
That’s a start.
SPARKS
*The way Intel Corporation does.
Cash and Short Term Investments
11,961,000,000.
Total Cash Dividends Paid -1,539,000,000.
AMD’s days of hand outs and free passes are over
I have 1.2Billion reasons to think otherwise! :)
Ok... that was a low blow - my apologies.
sparks: "How could I be so right and the pretty boy/power brokers be so wrong?"
Would these be the same guys who were charging headlong into the current financial crisis? I can't shake the feeling that there are some sugar daddies out there waiting to see if AMD's next chip is a "real winner" so that they can dump more money into their laps.
After all, this current financial mess comes barely a decade after the dot-com bubble. Did we learn our lessons then? Doesn't seem like it.
Tonus-
“Would these be the same guys who were charging headlong into the current financial crisis?”
To my comment-------------
“"How could I be so right and the pretty boy/power brokers be so wrong?"
As you know by now, I work in the Big Apple. My interest in the how’s and why’s of what make things work has taken me to some of the best, most exclusive places on the island. Hotels are my company’s specialty, the big ones. The buildings themselves represent a microcosm of every facet of the industry, from network hubs to 1600 ton chillers. Really, if it’s got a wire, I can do it. If it’s got glass fibers, I can do that too, AND test/qualify them for the engineers.
Occasionally, when another Forman/buddy needs some ‘special help’ they’ll pull me, and send me to a place like, say, “Golden Sacks”. No, that is not misspelled, that’s what we in my industry call the place.
Nothing, absolutely nothing is EVER refused from these guys. “What ever they want” is the mantra, no P.O., no ticket, “Just do it. We’ll figure it out later.” “If they tell you to stop working, STOP! Don’t worry about it, you’ll get paid.” And brother can they SPEND!!!! I dare not go into specifics. It’s like one of the INTC boys on this site divulging proprietary process secrets.
What’s more relevant however is the sight of DOZENS of limousines parked around the building for blocks waiting in cue for the next ‘Prince’ to be ushered to his respective vehicle. Further, that’s OUTSIDE the building, never mind the “catered lunches” we’ve gotten for a job well done. Shall I say extravagant?
“Hey, (SPARKS), you jackass, I can’t believe you’re eating those disgusting, f**king snails. Yo, Pete, can you believe the shit he’s eat'n!”
Consequently, believe with every essence of your being, when I said “pretty boy/power brokers” that this was an understatement to the EXTREME. I will not, however, criticize these boys (too much), because I know all to well which side my bread is buttered on.
They’ll be back in full swing, and so will I. I saw it ten years ago.
M-O-N-E-Y
SPARKS
"They’ll be back in full swing, and so will I. I saw it ten years ago."
Actually, I think things will be fundamentally different for Golden Slacks (the other name they go by). They have gone the banking incorporation route and have already gotten an expedited approval from the gov't (as I think either Lehman or some other investment house did).
Having gone this route they can no longer gamble (speculate) while levered to the hilt... Goldman made money with the run-up in the financials and then made money on the way down shorting those firms and did all this at huge leverage ratios which only multiplied the gains. However with their new status they will have to deleverage over time (not sure the specifics on how fast this has to be done). What this means is that no matter how smart they are, they will earn money at a much slower rate as they cannot use $1 in assets to lever $30-40. This means lower EPS, a lower EPS growth rate which will translate into a slowdown in stock price growth and behavior of their stock to be more like a traditional bank. So I'm sure there will still be plenty of fat cats and an overabundance of meaningless VP titles, but things will be fundamentally different. On top of that Goldman made money by being smarter than the other guys as an investment house, as a bank, it will be a lot more out-hustling the other banks - something that remains to be seen if they will excel at.
The more I think about it, the more it looks like Intel has a crystal ball operating on, well, all crystals :).
If the economy does slip into a recession, I'd bet that even the upper middle class will forego the high-tech toys this Christmas, and instead of that $1500 laptop under the tree, some kids may find a $500 Netbook instead. So Intel's Atom investment looks to be a genius move, whereas AMD's giving up on Bobcat (if indeed true - AMD officially denied it I believe), looks to be an Apple Genius Bar move (i.e., a tad moronic :).
It seems that AMD’s releases/roadmaps have become pretty much an also ran nonevent.
http://en.expreview.com/2008/09/25/amd-energy-efficient-processors-going-45nm.html
There are dual cores (?) mentioned.
So take your pick, Double Troubles or Triple Cripples.
Man this is getting uglier.
SPARKS
Yawn, what no noise about AMD and the upcoming 45nm Shanghai, where are all the AMD fanbois.
AMD is only a year late and a few billion short.
I can't wait to see the power and performance numbers. Don't expect anything grand on 45nm with traditional leakage oxide and crappy SOI
Ars Technica had an interesting piece on Moore's Law the other day. The assertion that you can't change the process node the fab is running on is not completely accurate as you can reuse most of the tools from one node on the next node as long as you don't change wafer size. Other than that, I thought it was a pretty good read, if a bit over simplified.
And here was an interesting piece on mobile computing. Combine this with real, viable speech recognition and most of the barriers for small devices (other than battery life) virtually disappear.
"I can't wait to see the power and performance numbers."
Actually the few leaked power #'s seem pretty decent - it looks like AMD was able to push the cliff that was about 2.5GHz on 65nm to probably about 2.8GHz (maybe 3). Not huge but considering there is no gate oxide scaling, if they can get 10-15% speed and hold power constant to slightly down that's not too bad (especially as they are not ready to go the high K route yet).
There is more to power consumption than gate leakage - if AMD was able to keep subthreshold leakage under control, couple that with the typical Vcore reductions for a new node, and maybe some design tweaks from 65nm I would not be surprise to see some decent improvements. Again nothing earth shattering but it won't be as poor as the gains (if any) from 90nm to 65nm for AMD.
And SOI doesn't HURT leakage - you can argue over the benefits of using it (and whether it justifies the added cost for the substrates) but it is not hurting the power #'s. It's not the SOI that makes AMD's process 'crappy'.
I'm not an AMD fan, but don't sleep on AMD's 45nm. It is behind Intel on both schedule and performance, but then again this has always been the case for AMD.
"I can't wait to see the power and performance numbers."
This is interesting. There are some preliminary pre-release numbers and they don’t look good. What “G’ said being on the cliff is gospel. Their will be no 3 GHz part.
According to this:
http://www.xtremesystems.org/Forums/showthread.php?t=195780
The thing burns 176 @ (total system load) 2.3 GHz
It is an approximate 12% reduction from its 65nM Pheromone counter part.
As for the performance of both INTC and AMD top bin components, this THE ONLY comparison I could find.
http://www.digit-life.com/articles3/cpu/amd-athlon-64-x2-6400-p1.html
Even with a top bin, cherry picked CPU there is absolutely no way Deneb will come close to a 2.8 to 3 GHz Yorkie. It ain’t gonna happen. Not against MY Yorkie it won’t!
20 percent is what AMD is claiming according to this:
http://www.tomshardware.com/news/amd-shanghai-cpu,6440.html
If we take the Digi-life Pheromone numbers, add 20%, Denube may get close to an E8500 (on a good day).
Then there’s Core i7 ready to piss all over AMD’s parade. Therefore, we are back to the same old rant and mantra where AMD’s chips will be compared to INTC’s 1 year old + designs; just to be fair of course.
Of course there are the monkeys who will overclock a Denube and cry we did it, we beat a Q6600! Why don’t we just take a Denube and a QX9770 and overclock them both, just to be fair, of course. ANYTIME BABY, ANYTIME!
Horseshit, I say. De Nube will be another $235 dog.
Woof.
SPARKS
What “G’ said being on the cliff is gospel. Their will be no 3 GHz part.
the cliff part.. yes... the 3 GHz I think they will have one (with an outside shot at 3.2GHz on future stepping late in the 45nm lifespan). The 'cliff' means the high clock will be low volume, high TDP part and may be limited to desktop only.
Of course, you will not see actual TDP’s (no mater how AMD cares to measure it) until it is actually released, specially the Black Edition part.
That’s the dirty little secret is it not?
During a test (I can’t find the link, and I’m still looking) it was shown TDP rose exponentially for every 100MHz increase in speed to the point where it made Prescott look like an energy efficient part. (At the same clocks)
The article also stated that De Nube was basically an 80% total revamp of the Barcelona design. I don’t know how much of that to believe. However, AMD is claiming (pimping) TDP’s around the 60 watt area at 2.3 GHz (nearly half of Pheromones). I’m sure you’ve already have done the math at 3 to 3.2 Ghz by the time you’ve reached the end of his sentence. I’d like to know your estimates. (Since, you’ve never been wrong thus far.)
Actually, given the pre-release information, or lack of, plus AMD’s penchant for crowing when they’ve got some thing good, I suspect the thermals are still the biggest issue, no mater how they serve this concoction.
No Hi-K, no milk and cookies.
Because of the lack of data available on actual Pheromone/De Nube TDP,s , graphs charts, etc., I’d like to get a few Amprobes and do the measurements myself, just to satisfy my own curiosity. (AM2 boards and Pheromones are so cheap)
On second thought this would be a cataclysmic event. Being a Times Square Sized poster boy for an INTC fan, weird things would happen as soon as I touched the chip. Sort of like what happen to Professor Quill at the end of “Harry Potter, The Sorcerers Stone”.
Plus, my cadre of INTC chips would never forgive me, and probably all go on strike, Laptops inclusive, and QX9770 would never treat me to a 4.27 Gig run again.
God forbid
SPARKS
By the way, if AMD does have a target, this should be the one to shoot for.
We'll see.
http://www.newegg.com/Product/Product.aspx?Item=N82E16819115041
SPARKS
AMD 45nm is nothing but a dumb shrink from a dumb ass company
With no materially oxide scaling they won't be able to effectively scale the transistors and improve transistor source-drain leakage. You need thinner electrical oxide thickness and with the horrendous leakage of the SiON oxide at these dimensions IBM had limited options.
So all you get with 45nm is more transistor density and reduced gate capacitance. That will help dynamic power a bit, but standby power will not get much benifit. Sure width of transistors will go down but they have added more transistors.
They do get the benifit of learning from their royal fuck up between design and silicon models so a Barcelona like disaster is unlikely but overall it will be nothing spectacular.
Nehalem versus Shanghai will be one step closer to AMD going bankrupt.
They will never get that fab in NY going as they simply can't afford to, yet then then can't afford not.
Hector reall fucked up the priorites between 2002 and 2005
"They will never get that fab in NY going as they simply can't afford to, yet then then can't afford not."
From your mouth to Gods ears.
SPARKS
Afford? you guys just don't get it. One of major objectives of Asset Removal is to raise cash for the fab.
Is it a good move? Is it a smart move? Doesn't really matter - AMD will be the majority owner in a pseudo-foundry. The minority owners will be ponying up the cash/capital for the NY fab to get off the ground. If Asset Smart gets done, the NY fab will likely be built. If it doesn't get done Sparks you'll save a few bucks (Actually you probably won't as that money will just get spent elsewhere)
Sure this is speculation, but keep in mind the person telling you AMD can't afford to do this also thought (thinks?) the NY grant was a loan.
You can take the tick tock out of lex, but unfortunately you can't take out the irrational thinking.
You know how much denial I’m in. You always punish me with clear thinking logic and reality.
A mere Angels breathe of this deal falling though, is enough to fill my accountant with joy, and I’ve never seen him smile from January thru mid April.
I realize I was being frivolous and pathetic, but with my NYS Income Tax, I, more than most, can afford the privilege. Trust me, I have paid for it. I was just enjoying the moment.
Besides, I’d take news like this from the devil himself.
SPARKS
Well, this is New York. The city pitched in $220 million for the new Yankee Stadium and the city/state pitched in $165 million for the new Mets stadium. Both of those are considerably smaller sums than were being bandied about several years ago.
New York, where paying rich baseball teams $385 million to build economic sink holes is considered a bargain. Maybe we can get AMD to build a football stadium instead of a fab!
Yo, Toe, easy wrangler, these stadiums are money makers BIG TIME! 400 million, peanuts (!), they’ll get it all back in spades, and for the next 50 years!
First, both bad boys were Union built, ahead of schedule and under budget. Wanna talk about other jobs, permanent ones? The infrastructure of both stadiums needs people, and lots of ‘em. Someone needs to sell the beer and clean the toilets, fagettabout routine maintenance. Tax returns, PA-LEESE, 8.5 percent to every dollar spent. You can’t make that kind of money in ANY hedge fund these days. TV, Cable, Satellite, Radio, endorsements, advertising, souvenirs, memorabilia; the list is endless.
Let’s talk seats, the cheap seats, forget the corporate boxes. You give the franchise $22k plus a $1600 year maintenance fee, bingo, you’ve got a season ticket. (Don’t forget the sales tax.) Multiply those figures by the number of seats sold, and the FANS are building the stadiums, for both teams.
I could go on, but here’s the thing and there’s no getting around this. The Yankee’s and the Mets are the game in town. AMD will never generate this kind of sustainable revenue and they are definitely NOT the only game in town, not with INTC, the big league hitters out-swinging and out hitting AMD on quarter by quarter, basis. These ball teams HAVE MONEY! AMD is broke dick.
Like “G” says this deal was cut years ago, and it’s turning out to be a bad one, after all. Not for 420 three or four year jobs, most of which will be imported anyway. You simply can’t train upstate rednecks to be process engineers, not with HALF the caliber of brains of some of the guys on this site, ya don’t.
I wanna puke.
SPARKS
Yeah, one more thing. Lets not forget NYS W-2 forms of, say,
A-ROD!
.....and thats just one player.
SPARKS
In The Know,
They don’t call you that for nothing. I’ve got to hand it to ya. ASUS is shipping 700,000 Eee’s a month.
So much for Abstencha’s grand theories. ATOM is a smash hit!
Is it any wonder why INTC backed out of OLPC?
http://www.theinquirer.net/gb/inquirer/news/2008/10/06/asus-manages-ship-700-eees
SPARKS
Looks like we may finally get to learn, in less than nine hours, what Asset Smart is all about.
Anyone still have their head in the sand and think Asset Smart = the SMART manufacturing efficiency project? This blog had it nailed from Day 1.
A.M.D. to Split Into Two Operations
http://www.nytimes.com/2008/10/07/technology/07chip.html
In addition, the company said two Abu Dhabi investment firms would inject at least $6 billion into the two firms, mostly to finance a new chip factory that A.M.D. planned to build near Albany, N.Y., and to upgrade one of the company’s existing plants in Dresden, Germany
Hmmm...what a surprise!
6 Billion... there is no way this SHOULD get thru regulatory but as there is talk of the plant in NY, expect the "Stop shipping jobs overseas" gimmeckry from the NY Dems.
6Billion into a company with a current market cap of 2.6 Bil!
Advanced Technology, which was formed by the Abu Dhabi government, has promised to put up $2.1 billion immediately and contribute $3.6 billion to $6 billion more to build or upgrade chip fabrication plants, also known as fabs. A.M.D. said the two companies would share voting control equally.
It [AMD] will own 44.4 percent of the new entity, which has been temporarily named the Foundry Company, a reference to the technical term for a chip factory. The Advanced Technology Investment Company will own the rest.
Not sure how this meets the x86 license of AMD outsourcing no more than 20% of CPU manufacturing. With a 44% stake I don't think they technically own the foundry company (though the 50/50 voting may be a sleazy way around it). Of course there's always the NY pols to put pressure on Intel if need be.
With the cash infusion, A.M.D. said it was now committed to moving forward with plans, first announced in June 2006, to build a huge $3.2 billion chip fab in Malta, N.Y., north of Albany.
Sorry Sparks, I really did not want to be right on this one. And the sad thing is the majority of the NY state taxpayer money is now going to Dubai, and not AMD. I wonder how the technology export controls have been gotten around... my goodness this is really not good. America for sale.
Oh and the foundry now assumes all of AMD's Debt! Hey Sparks find the King, find the King... $3 Bil will get you $6 Bil... just a really big 3 card monty game!
I must hand it to AMD though, the timing on this is pretty much perfect. Their earnings are set for release next Thursday and now they can pretty much state ANY number and the earnings call will be probably be 95% about the spinoff.
If the numbers are good they can say this will help them get even better. If they are bad, they can say "look, we are making hard decisions to fundamentally improve"
The beauty of the timing is that AMD will likely pull out the "we're still ironing out the details" card when they get asked the tough questions like impact to margins, long term competitiveness and how is a simple infusion of cash going to suddenly make the fabs suddenly profitable (or do they simply expect the foundry to operate at a loss, eat AMD's debt, and hang on long enough for AMD to divest themselves from it fully)
Peh, nice. The “new company” will absorb1.6B of the parent company’s $5.3B crushing debt, and the White Knight Arab Micro Devices will secure the rest.
“G” your logic is inscrutably accurate, the New York deal will go through, and the 30% coupon will indeed be cashed.
Does anyone have a clue on how this will affect INTC’s licensing issues? I’ll bet their lawyers have briefs already drawn up. Hmmm, a settlement, in conjunction with ‘modified’ x 86 agreements might look very attractive right about now.
“G”, where do you hide that crystal ball?
“I'd still say slightly better than 50/50... Asset removal/smart/light will pawn off the debt and 1.2Bil in subsidies is still a 30% off coupon. Keep in mind if AMD splits (and of course if they survuve) and has a majority ownership of the foundry they may be able to use the cash they get from the minority stake they sell to get the ball rolling in NY (it's probably cheaper intermediate to long term then ramping F30)”
They will hold a 44% stake, or so it is reported.
In any case; Bravo, spot on, and well done!
However, allow me to pat myself on the back. (I’ve been a good student)
http://chicagrafo.blogspot.com/2008/05/about-amds-acquisition-by-dubai-or.html
SPARKS
Does anyone have a clue on how this will affect INTC’s licensing issues?
It's a tough call (and one probably above the pay grade of most/all on this board). While it appears as though they will have a 44% FINANCIAL stake in the company, AMD will have 50% of the seats on the board. This may be considered to have a controlling interest and hence not technically "outsourcing production"?(I don't know). If not, why get 50% of the board with only a 44% stake?
Either way, i truly hope the US gov't looks at the technology aspect of this. I also wonder if NY will look at this (effectively giving a ~700Mil grant to foreign governments).... ah, who am I kidding, these are the idiots who did this deal in the first place.
1.2Bil for 1200-1400 jobs... that'd be about $1 MILLLLIIIIOOON dollars for every job created...your tax dollars hard at work! (I guess on the bright side we can be glad these politicians aren't running any companies!)
“And the sad thing is the majority of the NY state taxpayer money is now going to Dubai, and not AMD”
“ Hey Sparks find the King, find the King... $3 Bil will get you $6 Bil... just a really big 3 card monty game!”
I can’t find the King!
I can’t find the King!
I keep coming up with Jokers!
http://www.hillaryclinton.com/
http://www.govtrack.us/congress/person.xpd?id=300087
http://www.oag.state.ny.us/
http://www.ny.gov/governor/firstfamily/spitzerbio.html
SPARKS
I dunno, I read an article a couple of years back, looking at various studies of the economic impact of new sports stadiums, and in general they have a negative return over their lifetimes. The jobs they generate are mostly low-wage, and the team gets most of the revenue generated by the stadium. But the threat of leaving the state can make a local politician open up the wallet and grease those palms.
You can be sure that they'll be talking up the "extra jobs" angle of the NY plant. And that would be really great, if we hadn't spent 1.2 billion dollars for that 'privilege'. Then again, NYC spends something like $18,000 per child per year on public schools and we can't get half of them to read above a 5th grade level. Throwing money away is in our DNA.
Are Roborat and Scientia the same person? Both blogs have been down (in terms of author posts) for about the same time now ...
Anonymous said...
Are Roborat and Scientia the same person? Both blogs have been down (in terms of author posts) for about the same time now ...
Aww, c'mon - they have completely different writing styles. Besides, Scientia is busy mudderating - er, moderating AMDZone, while maintaining that he is unbiased :).
I dunno, I read an article a couple of years back, looking at various studies of the economic impact of new sports stadiums, and in general they have a negative return over their lifetimes.
I've read (though can't find links) similar studies - but a lot of that has to do with georaphy and the specifics of the situation. For example expansion teams generally get a new stadium, people attend for a year or 2 and then the euphoria evaporates and attendance drops or the same thing happens when a team relocates to a new area.
The NY stadiums are going to draw people regardless, so you won't see the eventual drop off. Also the stadiums, I believe, are ~1/2 financed by the owners which gives a much better chance for long term return for the state.
Or put it this way... if you hd 1Bil to invest would you give it to AMD to build a plant that will be obsolete in under 10 years or build a couple of stadiums (and have a little change left over) which will probably be around longer than that.
Some of the quotes out there are just ridiculous:
“ATIC and AMD are the ideal partners with which to create The Foundry Company,” said Hector Ruiz, chairman of AMD’s board of directors, who will become chairman of The Foundry Company. “Working together allows us to combine ATIC’s long-term vision and patient capital with our manufacturing leadership, innovation and highly-skilled workforce.
So basically ATIC is ideal because they have money and will continue to sink money into it long term. Long term vision? I think he is referring to vision to keep sinking in money - I don't think ATIC has a long term vision of the industry or where tech is heading. What exactly do they bring to the table other than money? That is the ONLY reason why they are ideal.
“We are as enthusiastic about AMD’s potential today as we were when we made our initial investment last year,” said Khaldoon Al Mubarak, chief executive officer and managing director of Mubadala. “This increased investment is a strong vote of confidence in AMD’s Asset Smart business strategy, evolved leadership team, and best-in-class technology.”
Yeah, that investment last year is working out really well - if the stock manages to just triple, they will be about break even! And best-in class technology? You can stretch a bit and say competitive but 'best-in-class', really?
'evolved' leadership team... how's that for an interesting, polite spin on the "old" leadership team?
The transaction is expected to close at the beginning of 2009 following satisfaction of conditions such as approvals from regulators, transfer of previously-confirmed New York incentives to The Foundry Company, and the approval of AMD stockholders for the issuance of common stock and warrants to Mubadala.
I can just see Sparks... "so you're saying there's a chance!?!?" (Don't count on it, buddy!)
One of the better summaries on the web (and a pretty good site for things like this):
http://www.fabtech.org/news/_a/official_amd_goes_fabless/
“We are as enthusiastic about AMD’s potential today as we were when we made our initial investment last year,” said Khaldoon Al Mubarak, chief executive officer and managing director of Mubadala.
I don't think there has ever been a better example of the "lipstick on a pig" analogy than that. Yee-ha Khaldoon, you go right on polishing that turd!
“Both blogs have been down (in terms of author posts) for about the same time now ..”
This is true. However, the Doc cautioned us few months back, due to personal endeavors, he wouldn’t be devoting as much time to the blog as he did in the past.
That doesn’t mean the quality of this blogs content or interest has suffered. On the contrary, we’ve had a few moments of intrigue and drama. In fact, at one point, it got downright ugly. Personally, I must confess, when process debates flair, although heated, there is a wealth of knowledge contained within, and these guys are the Socrates of the process set. Knowledge is the key, which ultimately underscores the difference betweens this blog and the other one you mentioned.
I’m the biggest Intel fan this side of the Hudson River, but it doesn’t mean I have lost my objectivity. That would be stupid. Frankly, posting something stupid here and YOU WILL get your ass handed to you, irregardless of your personal preferences. No one gets a pass, and that includes me an INTC, ever loving, overclocking, nut case. That’s a fact.
ROBO need not micro manage, edit, or moderate this site. Due do his OBJECTIVE and insightful analysis he has attracted the same type of contributors. And, if I may be so bold as to speak for the others here, that is just the way we like it.
SPARKS
“ approvals from regulators, transfer of previously-confirmed New York incentives to The Foundry Company, and the approval of AMD stockholders for the issuance of common stock and warrants to Mubadala”
This is rich, Senator Charles Schumer backing an Arab held concern to the tune of a billion plus, while chief rival Intel spends tens of billions building facilities in Israel. AND, Schumer, directly or indirectly, (I don’t care) is sanctioning an INTC investigation through the NYS Attorney Generals office.
Who knows perhaps he can get a Dubai World port built near/in, say, Williamsburg, Brooklyn.
http://gruntig.blogspot.com/2006/10/hasidic-protest.html
You simply can’t make shit like this up.
SPARKS
So you're saying there's a chance? (somehow that Dumb and Dumber quote just seems to ring true!)
However, during a conference call with press and analysts, Hector Ruiz, who will relinquish his current role as AMD's executive chairman and chairman of the board to become chairman of The Foundry Company, said that negotiations are just now beginning in earnest with the State of New York about their potential investment in the proposed 300-mm fab. "They have so far been very enthusiastic and I am confident we will get an agreement and the incentives. But we have no signature as of today."
http://www.eetimes.com/news/semi/showArticle.jhtml;jsessionid=QTF1L4F10PP4EQSNDLPSKH0CJUNN2JVN?articleID=210800047
Here ya go "G", some more Jokers for ya.
http://www.cbs6albany.com/news/paterson_1257684___article.html/state_announcement.html
SPARKS
http://www.reuters.com/article/marketsNews/idUSN0748080720081007
And the whip comes down.
And the whip comes down.
SPARKS
"Intel has serious questions about this transaction as it relates to the license and will vigorously protect Intel's intellectual property rights," Mulloy said of AMD's announcement.
Mulloy said Intel has asked AMD to make the agreement public but he said it would not, so he was not at liberty to discuss the matter in detail"
SPARKS
Yeah - I think the Intel stuff is basically just spin and a desire to get as many of the details out in public as possible. Intel still has the anti-trust to deal with and if they can start painting this for what it is - they may score a point or two in the public perception.
While they technically could potentially fight it, there is no way they would get anywhere with Congress out to get the big bad evil (read: successful and profitable) companies. Any guesses on how fast Cuomo, et al would push the antitrust investigation if Intel tried to shut this deal down?
No offense Sparks, but I'm glad I don't live in NY anymore those politicians would drive me absolutely nuts. With the state as Democratic as it is, once the primary is over, the race is over for most offices; and most primaries seem to be a race to see who can be the most liberal.
I love the quotes/backhanded compliments:
Sen. Schumer declared, "Upstate New York is back on the map.
Apparently before this it was just hick dairy farmers and upstate didn't matter? IBM anyone? Albany tech center anyone? RPI and various state universities? Somehow a single manufacturing site puts a huge and diverse geographic area on the map? Does the Intel Rio Rancho facility put all of New Mexico back on the map? How about the Chandler plant for Arizona?
"Cutting spending alone will not bring us back -- it's going to be wise investments. We can emerge from this crisis if we look beyond and invest in our future."
Apparently 1Mil/1job created is a wise spend of money? There are so many industries which would have more bang for the buck - a semiconductor fab is fairly light on the labor front in terms of $ spent. Welcome to the constant drip of money Sparks - if NY wants another plant (which if on the same site, a 2nd fab would provide even fewer new jobs)they'll need to pony up plant another 1.2Bil...
And Patterson is an idiot if he thinks a single MANUFACTURING plant is going to significantly keep local graduates in the area. Any guesses on the breakdown of those 1400 jobs and how many of those are engineering? (nfar fewer than you think) Any guesses on how many of those jobs will be TRANSFERS from Germany, Austin, etc..?
WTF
All this talk for quarters about A$$ lite and all Dirk and Hector could find was some Arabs to pony up a few extra billion. The Arabs are they really that stupid to believe they can get into manufacturing silicon CPUs for a profit? They got billions where they could buy a distressed bank for cents on the dollar and they instead buy a disfunctional CPU company that hasn't made money 9/10 quarters in their life.
Look at the track record of the IBMs, Motorola, HP, DEC, Toshiba, Hitachi, Samsung. Really small companies with no history in silicon. They failed to effectively compete in the profitable x86 / CPU market, and the fucking Arabs think they can do it?
AMD now gets a bigger fab and a sucker to finance it to pump out inferior 32nm chips in NY. The good for us it keeps intel running mad and fast so we'll for certain get even better parts from them and laugh as Shanghai dissapoints and the Arabs look to fill their white elephant with other 32nm chips.
The fact of the matter the market for 32nm among anyone but the highest profit and volume CPUs is small. AMD continues to lose money and now we got the Arabs helping them.
By AMD support the Arabs, buy INTEL well you know where they have most of their factories.
The funniest thing is to hear the pomp from the silly new consortium, that takes the cake
Question- how does this affect AMD's CPU and GPU development? Are they still 'close enough' to their fabs to maintain their pace in development of 32nm and 22nm technology? Is there a danger that they will fall farther behind Intel in process tech if they are disengaged from the fabs?
I was under the impression that not having your own fabs could hurt process development. Is this more a case of AMD having its cake and eating it, too? (ie, spinning off the fab for financial reasons, but maintaining a relationship that allows them to avoid the hazards of being 'fabless'.)
Are they still 'close enough' to their fabs to maintain their pace in development of 32nm and 22nm technology? Is there a danger that they will fall farther behind Intel in process tech if they are disengaged from the fabs?
These are the strategic questions noone seems to want to analyze.
Start with the opposite - do things get better/faster by separating process & design into 2 companies? Obviously not, so the absolute best case is no change.
On 45nm, AMD seems to be making a big deal about DFM (design for manufacturing) - something every IC manufacturer has been doing, except they give it a name (like APM) and make it sound innovative. These design rules come about from close interaction between the process and integration folks and the design folks with each of them having a reasonable understanding of the impact of what they own has on the other guy. Things like - How much can I give in on a design parameter for a better process or if I tweak this specific process (with impact to process stability) how much gain does it enable on the design?
The real problem with the foundry approach is that they can tune the process to work optimally with AMD's design, but what about other customer needs? Are they just stuck with what the foundry has tweaked for AMD? This would potentially put the foundry at a disadvantage to other foundries for non-AMD designs.
The other real conflict is the capacity planning. AMD will always want to move to the newer node as fast as possible, but for a foundry there will be a lot of demand for the lagging nodes - many customers will not want to update to new processes (mask cost, design cost) if they are running medium to low volumes. So when upgrading fabs there will be pull from both sides on the speed of the ramp and how much capacity gets dedicated where and where process engineering is prioritized (remember AMD does the CTI approach so they are constantly do incremental steps - if something can't be done on two nodes - which node do you prioritize). This is why AMD needed the 50% voting share (wll that and likely the x86 license).
As for process development it appears this responsibility simply switches to the foundry so it should be failry similar to what things currently are for AMD as IBM has said the new foundry will be a partner in the 'fab club'
Here's my question - what happens if the foundry fails? Unless AMD renegotiates the x86 license they have to produce 80% of their CPU's in house. If the foundry fails are they essentially done with CPU's? And then devolve into graphics only (ATI)? While I don't wish this, that would be a rather ironic - they acquire ATI, setting off a chain of events which leaves the company as....ATI.
“Congress out to get the big bad evil (read: successful and profitable) companies.”
Oh how true. Oh how ironic, those IDIOTS in Federal and State government are deliberately trying to cripple our most successful and profitable enterprises, Microsoft and Intel.
Incidentally, the same group of elitists are supporting and bankrolling the prime mortgage lenders who leveraged their capital 30 to 1, in the interest of greedy commissions obtained from Nuevo ‘real estate magnets’ who couldn’t afford the long term purchase price of homes they couldn’t afford. This is 700B plus 300B, respectively. This will come one trillion overall.
And…………… the home buyers get to renegotiate their mortgages so they can KEEP THE HOME THEY KNEW THEY COULDN’T AFFORD INITIALLY, marvelous!
Now we’ve got the sellout of American companies to foreign interest, partially subsidized by America taxpayers, in order to obtain “fresh capital”, and make the market “more competitive” for THE MOST successful American company WORLDWIDE, INTEL. I give you the unprecedented and historic AMD financial debacle.
Exacerbating this issue is foreign investment buying INTC’s IP on the cheap! A back door sale of x86, to foreign interests, astounding! Further, those absolute imbeciles in NYS are photo taking and patting themselves on the back for a deal that is good for the State and good for the Nation!
The CEO’s of these respective Corporations make monumental blunders, and we the American taxpayers subsidize the sellout, pay the freight, and deliberately hurt prosperous American companies by default. This is free market, Government intervention, at its absolute finest; all at the expense of the American taxpayer.
They’ve got the whole damned thing all wrong and ass backwards.
I hope to God Intel sues the lot, and pulls their license, at least in the interest of national security.
My contempt and distain for Hector Ruiz is unfathomable. He is absolutely despicable.
SPARKS
My contempt and distain for Hector Ruiz is unfathomable. He is absolutely despicable.
Despicable? Clearly he is incompetent and his history prior to AMD, suggests it's not just a one time event. But I'm not sure I would say despicable.
He's made very bad tactical and strategic blunders and in typical fashion of a CEO, he thinks he is best equipped to make things better. I have no doubt he is trying to put AMD into a better place but unfortunately he has let ego, and perhaps an overestimation of his skills/under-estimation of Intel's skills get in the way.
The real criminals (figuratively) remains the board - there sole purpose in life is to look at the strategic direction of the company and provide an impartial(?) look at where the company is headed and at the folks that are running the company.
In my view, how these board members allowed Ruiz to remain in charge as long as he did, and allow AMD to follow a flawed strategy (market share at all cost) for as long as they did is unconscionable. For them to collect money while doing this is an amazing dis-service to the shareholders, complete negligence on their part and borderline criminally/civilly negligent.
If there was an activist shareholder like Icahn involved - I have no doubt the board and Ruiz would have been thrown out quickly (though Icahn is too smart to get involved with AMD's mess).
You hear politicans bash all these CEO's and there excessive pay, but there is absolutely no mention that the system in place to govern this (company board of directors) is broken. Fix the board and you will fix a large part of the problem - the issue with the board is, that while it is theoretically elected - it is a good ole boy network, and there generally is 5-10 major stockholders who have a chunk of the shares who pretty much decide it.
Now what would happen if you allowed stockholders to sue the board members specifically for gross incompetence/negligence and not just the CEO's? Might they be a little more vigilant?
I despise the man for what he did and didn’t do for the company, the share holders, his many employees, customers, and ultimately the nation in regards to technological leadership.
What he didn’t do was to capitalize on his lead by focusing the companies recourses on it’s “core” business, take stock in the power and might of his adversary, and produce one innovative, successful product during his tenure.
Opteron and its success was certainly not his baby.
What he did was run the company straight into the ground with a ridiculously expensive purchase of a graphics design company. Dilute shareholder value, twice. Run slash and burn tactics to maintain market share while operating under monumental, consecutive, quarterly losses. Squander valuable resources (people’s jobs inclusive) and time on a pipe dream that never came to fruition, and ultimately, sold his failed strategy to foreign concerns out of dire, desperate, necessity.
To you the boards’ actions, or lack of, may be unconscionable. To me the individual at the helm has the final decision to set the course of the company’s long term statigic goals. Blinded by never before seen success, I believe, at the time, they would have followed Wrector straight into hell.
They did.
You know I don’t relish disagreeing with you, but that's how I feel, right or wrong. Every individual in his own time will seek his own level of incompetence.
He did.
Wrector Ruinz is a classic case of the ‘Peter Principle’ gone obscenely wrong, despicable.
SPARKS
Dow down 2%, AMD down 11%... who said you can't make money in the market these day (bought some AMD puts after the artificial bounce yesterday!)
Looks like the market has had a little time to digest this and may have realized, OK AMD got some cash - but did anything fundamentally change splitting one company into 2?
Apparently Wall St realized margins will be the same or worse with a middle man foundry margin stuffed in. Production costs and capabilities are going to be the same for the intermediate future (it's not like the ARAB partners are bringing chip making expertise) and that just gets passed on to AMD now. The only thing that really has changed is the cash flow dynamics (as AMD now outlays on a consistent basis as opposed to the bursts associated with ramping factories)
Funny thing looking at AMDZone - you have Ab-in-stupid spouting off and showing his complete lack of knowledge of manufacturing. He has told "the people" that AMD will have F38 (the F30 conversion) fully tooled by end of '09.
Now there are many, many reasons why this is completely wrong but I'll go for the most simple to understand one:
Anyone know the leadtime on buying a 193nm immersion litho tool? Put it this way if AMD placed an order TODAY for all of the scanners they needed (which is impossible but bear with me); at best they would dock ~Q4'09/Q1'10 and then you have to, I don't know, install and qualify them! Even Sparks couldn't hook up the electrical that fast never mine the plumbing, exhaust, chems, water, etc!
Now factor in that there is probably no way that an equipment supplier could build that many tools at once as they also tend to have other customers who spend a bit more money than AMD and forecast their tools YEARS in advance to secure the build slots at the supplier. But Intel isn't using immersion on 45nm, surely I must be wrong? Well, by now many of the 32nm tools have been forecasted and I'd bet some of the PO's have already bee cut for the early 32nm ramp tools (beyond development)
It's not like the supplier can turn on a dime and say, 'oh hey we need to build 5 more of these things, let's hire some more people and buy more parts'! Oh and as the deal has not officially gone thru, AMD probably couldn't order the tools today anyway; though I imagine they are forecasting them.
Of course I could be wrong and Abinstein, with his world class reputation for understanding manufacturing could be right. (did I mention his amazing proof that AMD had >50% better yield than AMD based on counting the # of factories?)
For some humor - you should take a read over at AMDzone. (He's also thinking AMD could get Silicon out of NY fab by end of 2010, but don't get me started on that!). I'm almost tempted to open an account but I know I'll be branded a troll as soon as I show the myriad of holes in some of the speculations being thrown around over there.
Tonus said...
I don't think there has ever been a better example of the "lipstick on a pig" analogy than that. Yee-ha Khaldoon, you go right on polishing that turd!
I think these countries are so awash in our petro-dollars, that deals like this are pocket change to them. They don't really care that much how profitable AMD will be - they just want to get their foot in the door with a hi-tech company. They know that eventually the oil will run out and so they are spending $$ to diversify and attract talent at all levels, from R&D to top-rated university profs.
If ever there is an economical solution found to alternative energy (hydrogen fuel cells, solar, tabletop fusion, etc), there's a growing chance it'll be discovered in one of the petro-rich countries...
"Even Sparks couldn't hook up the electrical that fast"
Maybe, maybe not, but I’d give my eye teeth to build a FAB.
Of course, this would be a dream come true. I get goosey thinking of the possibilities.
Here’s how it would work.
They’d get not one, but several electrical contractors who specialized in their respective areas of construction/infrastructure. (They must be qualified and approved Contractors to bid, invitation only)
Each company’s General Forman would report directly to each supervising project manager in charge of that particular area. Mains and power distribution would be one facet. Emergency power generation would be another. HVAC/environmental, building management, fire control, Networking/Fiber, Server Infrastructure (power only), lighting, and communications, etc., would have their respective contractors working in unison responsible for their respective benchmark timetables/schedules. (Oh, the joy, I’m drooling!)
Each General (Forman) would have a maximum 5 Forman there would be 5 Sub-foremen (Straw Bosses) under each Straw, 5 mechanics with a defined number of apprentices mandated by the Local Unions’ agreement with the Owners, usually two. If you’re not keeping score, that’s close to 200 guys, no sweat. Factor in Locker Pups, too. (Guys who expedite Tools and Materials) And, nimbers that could be from each company. (That's how IBEW LU#3 built NYC.)
If you are one of the Generals, Forman’s, or Straws you don’t touch your tools, you supervise. (General’s and Forman NEVER touch tools, the straws will, occasionally, but they could be reprimanded. I have) The Generals report to the companies’ Supervisor(s). Supervisors are god.
Never f**k with the Super, I’d wire hell for mine. (I racked a company VP against a piece of switch gear one night, my Super covered my ass. The VP was reprimanded for breaking my balls during a critical shut down. I was elevated to legendary status by the other Forman)
To most folks a construction site looks like a goddamned hodgepodge of guys running around in a huge mess looking for something to do.
In reality, everyone has a job, their being watched, the quality and time required to complete each task is carefully scrutinized by all.
There’s no laying down on this one, trust me, especially if it’s Fast Track. (Multiple shifts)
You give the IBEW a job like this; we’ve got the talented people and companies to blow this thing out.
Yeah, I know, that includes Arab Micro Devices.
SPARKS
Well, straight from the horses mouth - a place where I go for ALL my Silicon Process technology information (Fudzilla):
http://www.fudzilla.com/index.php?option=com_content&task=view&id=9869&Itemid=1
"AMD to abandon SOI after 45nm"
Well I'm not sure he's got it completely right, but apparently he thinks SOI starts to lose it's benefit on smaller nodes and the maximum benefit was on the 90nm node (hmmm, what blog did I hear that on 6-12 months ago?). I'm just spitballing here, but perhaps Intel knew this and was a significant reason why they did not switch to such an "innovative" technology like AMD did? (I thnk I may have read a paper on this)
Actually I think he's got this bungled too - I suspect AMD will do both SOI and bare Si processes. Depending on IBM's FINFET strategy and future memory (on chip) technologies SOI may serve purposes beyond the leakage (junction) benefit.
And on a completely different tangent - the ultimate in real world technology application ('email goggles'):
http://www.extremetech.com/article2/0,2845,2331965,00.asp
Google - where were you 5 years ago when I needed you!?!?
Sorry fella's my wife is up to her tricks again.
SPARKS
Louis Vuitton SacLouis Vuitton sac à main Louis Vuitton sacs Sacs Louis Vuitton
Post a Comment