Slashdot Log In
AMD To Spin Off Fabrication From Design Work
Posted by
timothy
on Tue Oct 07, 2008 08:00 AM
from the splitarama dept.
from the splitarama dept.
I.M.O.G. was one of many readers to write with the news that "Advanced Micro Devices plans to announce Tuesday that it will split into two companies — one focused on designing microprocessors and the other on the costly business of manufacturing them — in a drastic effort to maintain its position as the only real rival to Intel. 'This is the biggest announcement in our history,' said AMD's chief executive, Dirk Meyer. 'This will make us a financially stronger company, both in the near term and in the long term, as a result of being out from the capital expense burden we have had to bear.'"
Related Stories
This discussion has been archived.
No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
Full
Abbreviated
Hidden
Loading... please wait.
How is this supposed to make things better? (Score:3, Interesting)
Can someone give me some insight into why splitting the company into two is supposed to help AMD?
Re:How is this supposed to make things better? (Score:5, Informative)
There seems to be some good information here [wikipedia.org]:
IC production facilities are expensive to build and maintain. Unless they can be kept at nearly full utilization, they will become a drain on the finances of the company that owns them. The foundry model uses two methods to avoid these costs: Fabless companies avoid costs by not owning such facilities. Merchant foundries, on the other hand, find work from the worldwide pool of fabless companies, and by careful scheduling, pricing, and contracting keep their plants at full utilization.
Parent
Re:How is this supposed to make things better? (Score:5, Interesting)
I think this will turn out well for AMD, if they can maintain a good relationship with their foundry spin off and if the foundry spin off can keep up with the competition in terms of quality and technology. Although, I guess this would also free AMD to find other partners if they need to either expand production or find better production facilities in the future without neglecting parts of their own business.
Parent
Re: (Score:3, Interesting)
On the flip side the foundry is considerably more advanced than AMD is using right now. AMD has not been selling enough chips to fill up their fab. By freeing the fab, it will allow other companies, particularly after the Nvidia chip fiasco, to manufacturer without fear AMD might poach their chip designs. I think that's the motivation. In order to meet Intel and IBM on the fab front, they have to invest in billion dollar factories.
Re:How is this supposed to make things better? (Score:4, Insightful)
IC production facilities are expensive to build and maintain. Unless they can be kept at nearly full utilization, they will become a drain on the finances of the company that owns them. The foundry model uses two methods to avoid these costs: Fabless companies avoid costs by not owning such facilities. Merchant foundries, on the other hand, find work from the worldwide pool of fabless companies, and by careful scheduling, pricing, and contracting keep their plants at full utilization.
I don't see anything in here that requires two separate companies. AMD can stay a single company and still build chips for other companies to fully utilize their facilities.
It looks more like a decision appealing towards someone's fuzzy feelings: "look the fab is independent now, it's got nothing to do with AMD chips, you can hire it" and "look we're doing bad but we have a big plan to bail out out of the crysis". Ops how come I worded it this exact way :P?
Parent
Difficult to pull off that dicotomy... (Score:5, Insightful)
If they are a single company, then, internally, the two groups almost have to use each other or else seem bizarre. I.e., if the designers contracted out fabrication of a model even though their own fabrication division was not fully utilized, that would seem unhealthy. By the same token, if the fabrication division pre-empted production in-house designs for a third-party, that would similarly look bad.
With that view, it would be a tad harder for the fabrication portion of the business to attract design companies, with prospective companies knowing they are putting their manufacturing capabilities in the hands of a company that would be both partner and competitor. The conflict of interest is far from appealing.
Few large corporations under typical circumstances preserve in-house at-scale manufacturing. I.e., most x86 system vendors now at most design the system and then feed to another company for fulfillment, potentially even a company spun off from themselves when they reached a similar conclusion.
As consumers, we don't stand to lose, only to gain. For example, if nVidia has been held back in any quality/performance way by inferior fabrication companies, they may now approach AMD fabrication. Same goes for AMD v. Intel, if another fab company can deliver more aggressive process size/yield improvements, then AMD design can go to that company and produce a valid competitor to Intel.
Or it shows that both halves of the company were completely average nowadays even in only the context of their similar competitors, and still doesn't do well, but that isn't different from today.
Parent
Re:Difficult to pull off that dicotomy... (Score:5, Insightful)
"With that view, it would be a tad harder for the fabrication portion of the business to attract design companies, with prospective companies knowing they are putting their manufacturing capabilities in the hands of a company that would be both partner and competitor."
Well that's the idea here. By spinning off fabrication into its own company, other chip designers wouldn't be putting their ideas in the hands of a competitor.
Parent
ATI, nVidia (Score:3, Interesting)
ATI contracted out their fabrication in the past, correct? Since AMD acquired them, perhaps they now realize this might work for their x86 stuff. Disclaimer: I have absolutely no expertise in this area.
Re: (Score:3, Insightful)
Re: (Score:3, Insightful)
And yet, the GP may have a point. Because, it works for IBM - they are both fabbing ICs for others and for themselves.
Re:How is this supposed to make things better? (Score:5, Insightful)
I don't see anything in here that requires two separate companies.
From an organization or technology perspective, no, there isnt any reason to split into two companies. From a financial perspective, this makes a huge amount of sense. You'll note that this new company is receiving an enormous amount of new funds from investors, and taking a lot of the AMD debt. They're effectively splitting off R&D from Manufacturing, and people are free to invest in just the R&D component or just the manufacturing component.
Some folk out there thought the manufacturing side was worth a huge investment of cash. Lots of assets there that are worth a fair bit - but not if they're tied to work purely on AMD products. If the R&D side of AMD failed, then there are a crapload of perfectly good assets that would be lost, in effect. This allowed investors separate AMDs chips from AMDs fabs when investing. There is no inherent value in splitting the fabs off - except when someone is willing to spend $8 billion to fix them up, separate from the R&D side.
This doesn't mean they think AMD R&D is going to fail - its just about risk. Why tie your $8 billion investment to the ADD of the consumer chip market instead of to physical assets that will be worth something regardless of the mood of the x86 CPU market?
Parent
Re: (Score:2, Interesting)
ARM is fabless (Score:5, Informative)
Because Design without Fab worked so well for Transmeta?
I see your sarcasm, but it works for ARM and MIPS.
Parent
Re: (Score:3, Informative)
Re:ARM is fabless (Score:5, Insightful)
> Because Design without Fab worked so well for Transmeta?
I see your sarcasm, but it works for ARM and MIPS.
What ARM and MIPS have in common is they are RISC architectures with their own specification and market.
What Transmeta and AMD have in common is that they produce x86 compatible chips and thus compete directly with Intel.
Intel as a company owning their fabs has become famous for their well synchronised "tick tock" process where they successively introduce new design, then introduce better fab for the same design, then a new design etc. Such accuracy and consistency is hard to expect from a design-only company that needs to contract a third party to produce their own product, and both parties are constantly looking for a way to skim some pennies in the process.
Parent
Re:ARM is fabless (Score:5, Insightful)
Intel is in a unique position in that its R&D and fab budgets are, relatively speaking, limitless. With a lion's share of the market, Intel can afford to upgrade some of their fabs to 32nm at enormous expense with the comfort of knowing that the volumes will almost assuredly be there to make back the investment.
AMD had to play a much nimbler and dangerous game of trying to crank out volume while simultaneously playing catch-up on the fab side. It was wise of them to recognize this as a losing game. My assumption is that the AMD/IBM, et al alliance of gate research and the like will be fed into these fabs with some sort of preferential production rights to the contributors of the R&D budget.
Also, let's not forget that Intel is the subject of numerous international anti-trust suits.
IMHO, AMD has had some costly slipups but they have otherwise done an outstanding job of keeping pace with NVIDIA and Intel while on a fractional R&D allowance. In NVIDIA's case, they are pulling ahead.
Parent
Re:ARM is fabless (Score:5, Insightful)
To be fair, ARM and MIPS don't need cutting-edge performance. They are fabbed on whatever slightly older, absolutely dirt-cheap process is available. They're so small and low power already that a process shrink or two doesn't noticeably affect the overall performance of the embedded device.
Part of the reason it works so well is because companies that need to be on the cutting edge of chip tech (like Intel and AMD) pay the huge expense of building high tech fabs, then, when the technology moves on, they've got to do SOMETHING with the obsolete fabs, so they might as well contract out and make dirt cheap chips at minimal profits. After all, little profit is better than no-profits, on a fab you've long since paid for and (hopefully) profited from.
And I believe AMD was already trying to better utilize their old fabs, making (low-power) Geode chips for embedded apps and the like with spare capacity.
And this really shouldn't have surprised anyone... They've contracted out other fabs to produce AMD cores in addition to their own, but only as contingencies when they couldn't immediately meet demand... I suppose they don't have that problem anymore.
Parent
Re:How is this supposed to make things better? (Score:5, Informative)
Transmeta failed because its product sucked.
Parent
Re:How is this supposed to make things better? (Score:5, Informative)
Transmeta failed because its product sucked.
Absolutely. For those who don't know, the Crusoe uses a VLIW architecture with 128-bit words, and x86 instructions have to be decoded and RE-ORDERED in real-time into those 128-bit words. This is the same brick wall Intel ran into with optimizing compilers for the Itanium, but unlike Intel the Crusoe has to do it in REAL TIME.
Sure, the software translation layer meant that they could run Crusoe any architecture, but in the end it cost them precious performance. The chip itself wasn't much to sneeze at (two integer units and an anemic FPU), so it really didn't have the performance to spare. Then they hobbled the chip by integrating a nortbridge; this meant that ALL Crusoe-based systems would have the same video and I/O performance limitations, all in exchange for saving a buck or two on parts.
It didn't help that they hyped the successor, the Efficion, and then it didn't deliver in clock speeds or promised performance increases [vanshardware.com].
Parent
Re: (Score:3, Informative)
Re: (Score:3, Informative)
Re: (Score:3, Insightful)
Re: (Score:3, Interesting)
AMD is maintaining a minority interest - the Foundry Company has a 55% majority on the spunoff part.
My understanding is that while it's only 45% of the stock, not all stock is equal and AMD is keeping a majority of the voting rights.
Re:How is this supposed to make things better? (Score:5, Informative)
It is NOT supposed to help AMD. It is supposed to help AMD's stockholders.
This often happens with troubled stocks that have a number of different business functions that can be split off. Some of those business functions may represent a great deal of capital investment, but not return much cash. You don't want that capital tied up in idle buildings and equipment, but you probably can't sell those things to your rival who's happy to see you shrivel up and blow away.
So you split the company up. The more profitable divisions can start to appreciate in value or even pay dividends. The less profitable business can stay afloat on business from its former sibling divisions while the stockholders unload their stock in it. It's possible that new management can turn thing around.
Parent
Re: (Score:2, Informative)
FUCK! I held on too long. My AMD stock has done well over the last 8 or so years. I knew that ATI acquisition smelled funny...
This happens when the profit margins of the company are not inline with what shareholders expect from a company in the particular field in question. The idea for the split is that the one company will be in the field in question, and have much higher margins (without the other division bringing them down). The other company, while still profitable, is now in a different field wi
Re:How is this supposed to make things better? (Score:5, Insightful)
Right now, AMD fabs are ONLY used to produce AMD processors. They don't handle GPU manufacturing at this point. As a result of this, and because of the bad economy, the fab side of things would drag AMD down more than keeping the two companies in a good position. On top of this, from a pure bookkeeping/accounting perspective, it becomes easier for investors and potential investors to see a profit from one side of the business or the other.
The Athlon 64 X2 and Phenom sales numbers really are not bad, but the profits from the sales are never seen for investors if the fab side is losing money. The split will make it very clear how well the company is doing in each area. It will also open the doors for other companies to buy fab capacity from AMD, so AMD could make money by making chips for other companies. We may never see Intel use AMD for this, but other companies are out there.
The downside to this is that as two smaller companies, one side or the other might be purchased by another company, which would hurt in the long run. It's a dangerous time...
Parent
Re: (Score:3, Interesting)
I foresee a different future for AMD. I have a feeling that they're restructuring themselves to be a big ATi, which was a fabless "silicon design" company. There are plenty of competent and cheap foundries for silicon. The costs of duplicating the engineering work of others is weighing heavily on AMD, since their primary rival can outspend them by such a huge margin. It seems to me that they're betting on the survival of AMD intellectual property by having their competent engineers design stuff that's fab
Re: (Score:3, Interesting)
Or maybe, since AMD have been using IBM's process technology for a while now, but there's been a delay in integrating that process into their own fabs, they think they can come to an agreement with IBM to use IBM's leading edge fabs.
It would make a lot of sense. IBM gets to run their high end fabs at higher capacity with a product that doesn't really compete with IBM's Power-based products, but which competes with one of IBM's biggest potential competitors (Intel). They also won't have to risk more anti-
Re: (Score:3, Informative)
Re: (Score:2)
Pixie dust. Basically, by making AMD's fab capacity fend for itself, AMD's design work will, through the magic of pixie dust, suddenly not have to worry about lack of fab capacity.
If you don't see it, it's not real!
Re: (Score:3, Informative)
This quick synopsis gives you some good info on why it helps AMD: http://www.pcper.com/comments.php?nid=6262 [pcper.com]
Re: (Score:2)
Short-sell the fab spin-off. The likely split will be the parent company will keep all of the assets other than the expensive fabs, and all the liabilities will go to the spin-off. Instant profitability for one, and instant debt for the other.
Hmm... (Score:4, Insightful)
You have to wonder if this was actually a good long term idea that Intel would be doing it as well. I'm guessing this is more of an accounting trick to help their numbers look better and/or some how lower taxes. I don't own any AMD stock so this doesn't effect me too much... I just hope that they don't go under as Intel does need some one to compete against.
Backwards (Score:5, Insightful)
A) Design would rarely factor in the manufacturability of it's designs, driving up costs.
B) Manufacturing had a tendency to sacrifice quality to reduce costs.
This new corporate structure has only been in place for a few months, but so far has worked quite well. Entire product lines have been eliminated (design didn't know manufacturing was still making the old stuff). Entire processes have been eliminated (manufacturing thought they were needed to meet the final spec, but weren't).
Most of these issues could have been resolved with better management and communication, but when design and manufacturing are a single unit, these issues resolve themselves naturally.
Stick a fork in 'em... (Score:4, Interesting)
Stick a fork in AMD, they're done.
A design firm plus a foundry does not equal an integrated semiconductor powerhouse.
Who is left to compete with Intel now? At least we will have Nehalem. Get used to Nehalem, embrace it, love it. Because it's going to be around for a long, long time. At least we have the x86-64 ISA, on-board memory controller, and point-to-point processor communications as an AMD legacy. And thank $DEITY that AMD was able to put a stake through the heart of Itanium.
There won't be much future innovation from Intel without the spur of aggressive competition from AMD.
Cheers!
Re: (Score:2, Redundant)
You're right, but people aren't going to want to recognize it.
AMDs irrelevance will become clear in about a generation of chips, though.
Re: (Score:3, Insightful)
A design firm plus a foundry does not equal an integrated semiconductor powerhouse.
There is no need for an "integrated semiconductor powerhouse".
Who is left to compete with Intel now?
AMD. Because AMD doesn't have to waste money spending it on an idle fab plant or one that's not running at capacity they can spend more money on research. Meanwhile the fab business can make more money by contracting the fabrication of chips for other design businesses. Both businesses can benefit.
Falcon
Nothing to do with AMD? (Score:3, Interesting)
Um. You do realize that Intel had no intention of ever extending x86 arch to 64-bits, right?
And that the plan was to force anyone who needed more that 4GB of address space or eight 32-bit registers to migrate to Itanium?
And that Intel had strong-armed or bluffed all competing RISC vendors (except Sun) into abandoning their 64-bit CPUs based on Intel's plans for an entire Itanium 'ecosystem'?
And that they had the entire IT press eating out of their hands, blathering on about the bright inescapable future of
Abu Dhabi Investment Group (Score:5, Interesting)
It's a smoke screen for mismanagement (Score:4, Interesting)
A hot fab is useless unless you can get product to market, and sell into the markets you need to achieve sales goals. AMD hasn't done this.
While they have very good engineers, they're weak in so many places. An infusion of foreign capital makes no sense if you can't get the basics right.
Yes, Intel, IMHO, used illegal tactics to kill AMD at many turns. AMD needs to recruit the best and brightest and get a regime change in motion to diffuse their preyed-upon attitude. They could lead again, but not with the current regime.
Chopping the company into bits will be a distraction, not a savior.
Slow sucking sound (Score:4, Funny)
Once again, another American company selling off to foreign investors.
Another powerhouse US industry falling to the wayside.
Sucks that I foolishly spent many years studying to work in said industry.
At least both presidential candidates are planning to retrain me to sell cars or something.
Re: (Score:3, Insightful)
No. That "sound" is the sound of technology moving forward. At one steel mills were "high tech" but now it's a low tech "smoke stack industry" that has move over seas. It one point electronic assemby was a high tech industry and now it's moved to China where un-educated one time farmers can build iPhones. My point is that the cutting edge moves fast. The nest wave will be biology. Kids today who want to be on top and work in high tech should be taking Chemistry and Microbiology in school. Electrical
A lot of it is simply accounting.... (Score:5, Insightful)
They got greedy when they were on top, and charged too much for processors which allowed Intel to do to them exactly what they did to Intel(swoop in with cheaper parts).
They've also got some problems with maintaining any presence in the top end of the CPU market. This isn't a huge deal for fabrication as almost no one buys those thousand dollar CPUs anyway, but those thousand dollar CPUs are your next generation main stream CPUs so you've got to have them.
They've also had some issues because they aren't big enough to take what's been happening in the market lately as easily as Intel has. AMD is now worth less than they paid for ATI, they're not alone in being worth a lot less, but it's not as visible for other companies.
Essentially the biggest thing this does is allow AMD the design company to ditch its debts into AMD the fab company. Investors will be much more willing to accept debt in the fab company because at the very least the assets are worth cash and it won't be dependent on whether AMD can come up with something halfway decent design wise. If the design company goes under, they can always just go fab Intel CPUs.
The design company on the other hand, after offloading a whole lot of its debt, is much more likely to stay alive long enough to fix things. They've got to get designs out into the market, they've got to be cheaper, and they've got to be at least almost as good as the Intel parts, but they have to survive long enough to do that.
Realistically, AMD will probably buy the company back if they do survive because having your own fabrication facilities is probably key to being in the top of this market, but in the meanwhile they get to stay alive in a failing economy, a credit crunch, and a time of total lack of vision for the future.
This split, silly as it sounds, may allow them to survive long enough to do this, and at the very least might keep Intel worried enough that they don't go back to the old days for a few more years.
Beginning of the end of exponential growth (Score:3, Interesting)
I read that Gordon Moore once explained his "Moore's law" as being economic, not technical. He said that when Intel builds a new plant, each new plant costs about twice as much as the last one. so he said at some point a plant will cost more money then there is on earth so they will have to stop buiding new plants at some point and then Moore's law will end. I think what we are seeing is the front end of this. A few smaller companies are finding they can build new fab plants. Maybe in 20 year even Intel will have to do what AMD is doing and then we will see the end of exponential growth.
The key observation here was by Gordon Moore that growth in the number of transisters is due to growth in capital spending on fab plants, not technology.
Re: (Score:3, Insightful)
Not really. By decoupling it makes it somewhat easier for the fabrication company to fill it's production lines by making chips for other fabless companies. It's not quite a move of desperation or some kind of accounting trick as you seem to imply.
Re: (Score:3, Insightful)
I am not an MBA but I would hazard to guess that AMD Fabless and AMD Fab will have some sort of contract in place guaranteeing a certain level of capacity.
Re: (Score:3, Interesting)
If you have followed AMD for some time (and I have - I lived in Dresden for two years), you would know all the troubles associated with building a fan now.
For its first fab, AMD could pull it mostly on its own. Still there were some other parties in the deal. For seconds fab in Dresden it was much much more complicated: further improvements in manufacturing processes made fabs more expensive. $2Bln is quite number for smallish company like AMD to pull. And finding partners is quite hard, because many wo
Re: (Score:2)
Why would they help with the bail out? They aren't in the business of charity, they are making an investment. They're buying in to AMD because apparently they see it as a wise investment. I would assume they aren't giving money to Fanny Mae because they don't see that as a wise investment.
Re: (Score:2)
.
$6 billion spent on a factory in upstate New York is good news for upstate New York.
It takes guts to put that much money into new industrial development during a financial crisis. The money could have been invested in AAA rated bonds from Microsoft.
Re: (Score:3, Interesting)
How better to help with financial bailout than to make a big company stay profitable instead of laying off highly skilled workers?