Login:
Password:

AMD to Restructure, May Outsource More to IBM, Chartered

By Scott M. Fulton, III, BetaNews

April 19, 2007, 10:23 PM

During a report of perhaps the worst quarter of performance in the company's history -- what its executives called a "perfect storm" -- AMD CEO Hector Ruiz revealed an extraordinary number of options are on the table for an unprecedented corporate restructuring.

Exploring the possibility of adopting an "asset-light" business model, AMD may go so far as to license manufacturing rights for its microprocessor designs to long-time partners IBM and Chartered Semiconductor, as well as expand its licensing relationship -- one many didn't know AMD actually had -- with Intel.

Ruiz told analysts that, as AMD's mix of customers shifts from selling mainly to the channel - to businesses that purchase in volume - to OEMs which produce designs based on AMD processors that often sell to consumers, his company needs to reconsider how best to address that customer shift.

"One of the things that has become pretty clear in our experience, we've had now for a number of years some experience in partnering with people such as, for example, IBM in joint development programs. That has taught us a lot about how you can do some asset-light strategies, since we didn't have to build an R&D laboratory to do that. We've had also an experience now for a number of years with some of our friends in the foundry business, and in particular, Chartered Semiconductor...Through the acquisition of ATI, we now have a perspective into a very asset-light model that we're pretty excited about learning more about."

In recent years, Motorola's communications division and Texas Instruments have both adopted so-called asset-light business models, that serve to reduce their retooling and maintenance costs by trading them for production costs that are more manageable. As part of the tradeoff, they rely more and more on foundries such as Chartered, whose own business models are based on producing other companies' designs in a cost-controlled fashion.

Ruiz told analysts this afternoon that AMD has already made IP licensing arrangements with a number of partners. For it to go more asset-light, if it so chose, all it would need to do is exercise the options it has already planned out.

"We have a broad array of IP licensing agreements with many people in the industry," remarked Ruiz, "and every one of the options that we're considering for an asset-light business model takes all those into account, and incorporates those into the agreements that we have. We have a lot of flexibility in those agreements. As a matter of fact, I would put it in the category of damn near infinite.

"Our plans are to start really narrowing down the choices and exploring the opportunities with the various possible partners that we could have," Ruiz continued with his characteristic transparency. "It covers the whole gamut of opportunities...We already have a joint development agreement with IBM, which we view as an asset-light strategy from the point of view that we don't have to build an R&D facility."

Ruiz will assemble a special task force, he told analysts, whose job will be to ascertain the health and status of his company, and make recommendations as to what cuts can be made while doing the least damage. Ruiz himself will chair this task force, which will also be staffed by CFO Bob Rivet, executive vice president for sales Henri Richard, and executive vice president for media and former ATI CEO Dave Orton.

"While I expect this task force to be temporary in nature," Ruiz stated, "lasting no more than a year, I expect this transformation to be bigger and [have] more dramatic an impact than the one we undertook in 2002."

Perhaps this task force's first order of business will be reconciling Ruiz' two primary goals, which some could say seem disparate: to work toward a more "asset-light" model, while at the same time accelerating, if possible, the progression to full 65nm production and later to 45nm production. AMD executives this afternoon gave the full impression that any effort to slow down this progression or turn down the volume of this product shift, could jeopardize the company even further.

But there must be changes in AMD fabrication plant development plans - at this point, they can't be avoided. Fab 30 is AMD's main production facility in Dresden, Germany, and one of its oldest buildings, producing 200 mm wafers. Fab 36 has been under construction there for the past three years, being built to produce 300 mm wafers.

Imagine a wafer as a kind of transistor farm, and you can understand why producing bigger wafers automatically means better output. AMD's plan had been to convert the 200 mm facility into an all-new, 300 mm facility to be called Fab 38. As President Dirk Meyer told a Lehman Bros. analyst, "Compared to that change going forward, we will reduce the rate at which we convert Fab 30. We'll still completely ramp Fab 36."

Pre-production samples of the first 65 nm chips using AMD's Barcelona architecture were shipped during the last quarter, Meyer continued. More samples will be shipped this quarter, and customers will be able to start shipping Barcelona-based systems in the third quarter. No delays there.

Next: Will AMD be forced to make peace with Intel?

Continued. . .
1 | 2 | Next >>

Add a Comment (27 Comments)

BetaNews reserves the right to remove any comment at any time for any reason. Please keep your responses appropriate and on topic. Foul language and personal attacks will not be tolerated.

Name (required):

E-mail (required):

Enter Your Comment:

By dahri

posted Apr 20, 2007 - 10:31 PM

i seriously think they should replace their CEO or whoever in position responsible for managing research and also marketing. they need new strategy.

Score: 0

By foxfyre

edited Apr 20, 2007 - 1:40 PM

This should not come as any great surprise if you have been aware of their economic model!

But as Intel tripped ONLY because it thought that it could dictate the evolutionary process in the form of a 'jump discontinuity' moving from 32 to 64 bit chips in the form of Itanium and were (stupidly) surprised that most preferred a compatible and smooth transition process which AMD provided, Intel is STILL the major R&D house in x86 CPU development.

IBM long ago eschewed the x86 market and went far ahead with Power (which Apple smartly used until the gulf between current Power technology and tiny desktop systems rendered the desktop Power chips several generations behind current Power offerings with no real incentive for IBM to commit more resources to deal with such a relatively 'old' niche market- thus making it a smart move for Apple to move to the commodity x86 market with a provider who has incentives to stay current in the particular market space).

So IBM has moved far ahead with R&D and basic research, although they have not been focused on x86 products. Thus it is a very smart move by AMD to license all of the technology it can from IBM (as it has in the past!) and to take advantage of IBM's production capacities and their economies of scale to produce whatever they can for AMD (as AMD has utilized in the past as well.)

In other words, only the glitch in the value pricing model due to Intel's arrogance created the temporary perception of AMD being able to maintain a toe to toe battle with Intel.

This is simply a correction for AMD and a return to the classic value pricing model where AMD can best leverage what they can provide while taking full advantage of the benefits of partnering with an R&D leader who is not directly competing in the x86 market space.

A smart move.

Score: 0

By rsx508

posted Apr 20, 2007 - 10:43 AM

God, I hope AMD doesn't go under. Intel will go back to laziness, prices will creep up again, and innovations will slow down.

Score: 0

By horsecharles

posted Apr 20, 2007 - 10:53 AM

They're supposed to re-teke the lead with their upcoming quad-cores...supposedly....debatable in some quarters....

Score: 0

By mjm01010101

posted Apr 20, 2007 - 12:58 PM

Intel's core duo's performance has been confirmed for 1.5 years now.

AMD, as a competitor, really should have had an answer, 6 months into this. But it's April 2007 and they have nothing to show for it, and they actually may have distracted their org with the ATI buy.

They still drag massively on marketing, as well.

Score: 0

By Heero

posted Apr 20, 2007 - 1:44 PM

http://www.legitreviews.com/article/490/1/

I think AMD is starting to bounce back...

But this is only one revue... we'll see how it goes.

Score: 0

By mjm01010101

posted Apr 21, 2007 - 10:48 AM

Competitive in one market segment, on the cheap end, ain't gonna help AMD very much. AMD: competitive when the stars align just so-- just ain't gonna do it.

Score: 0

By Heero

posted Apr 21, 2007 - 5:16 PM

Actualy, if AMD can regain the low-end market, it will do them a lot of good.

Having the top-dog spot is nice, but at the end of the day, the number of low end parts they sell totaly out weighs the number of high end.

The best way to start to gain some money, and momemtum is to start getting people to buy your products... low price, and performance are key. So if AMD can get that, then it will help them a lot.

Same with videocard... how many people buy super high-end videocard? Small segment... the majority look at the intro/cheap stuff.

HP and Dell sell more low end computers then they do super high-end gaming, or server rigs(Doesn't mean they don't sell a lot of them), but the majority of the revenue is in affordable price range.

If AMD can prove to have the better cheap chip, then more people will start buying those chips. And then when they want to upgrade to something more powerfull they'll know that the cheap AMD stuff is good, so they'll be more inclinded to buy another AMD chip.

Score: 0

By ce la vie

posted Apr 23, 2007 - 6:35 PM

here is an interesting discovery:
otellini gets 6.2 mill + stock options + perks
ruiz gets 16.5 mill + stock options + perks
the gluttony of upper management at amd needs to be made realistic according to time space continuum.

amd already has the low end market. semprom is the best selling chip amd has. actually semprom is just an improved version of cyrix. any geek can create another version of semprom (you don't need mit for that).
if you would like amd to bounce back, you should recommend quality over price. the reason why amd is still cutting prices is because their inventories are not moving according to their predictions. on top of that, ibm is not moving fast enough with innovation for amd.
there are so many problems with amd, it's amazing this company still behaves as a corporation.

Score: 0

By rsx508

posted Apr 20, 2007 - 1:35 PM

I don't think anyone can argue against how badly AMD has handled the ball. I seriously hope they can regroup and begin executing on target and improve their marketing. Not to "beat" Intel, but simply continue competition. We as customers need that to remain as long as possible.

Score: 0

By horsecharles

edited Apr 20, 2007 - 8:44 PM

I totally agree...we need competition...any: i was disappointed when Sony dropped out of fabbing Cells-- those processors have/had a chance, however tiny, of reaching desktops(& in the process giving us a Windows, & more importantly a viable(tons better actually)- depending on how well & fast other OS supported it) 64bit, alternative).

Score: 0

By horsecharles

posted Apr 20, 2007 - 10:04 AM

Dear AMD:

1. License the Cell Processor-- it's way better than anything you or Intel have now or on the horizon!!!
2. Yes, Virginia it does have OS...there are 3/4 Linux variants supporting it, in addition to Solaris.
3. Get together with Cell consortium(+ their various / individual preferred partners)- ibm, sony, toshiba, motorola, ti, apple, etc. + Sun(ask Sun about their technologies currently gathering dust-- where they can string together dozens of processors & ram modules with NO bus needed).
4. Bring Google on board on whatever terms they require.
5. Get Ubuntu onboard too.

Bingo!!! You will now be HUGE, putting us all on supercomputer-type systems.

Score: 0

By CarLox

posted Apr 20, 2007 - 9:08 PM

ok.... is the cell processor a multi-treat processor?? (meaning that it can process more than two task at the same time?) i dont think so for what i've read lately so hell no

Score: 0

By Heero

posted Apr 20, 2007 - 10:41 AM

If they can make a Cell x86 compliant, then I can see it working. Otherwise it's going to be the same story with Apple and the G line processors... Great chips, just no one writting code for it.

Score: 0

By horsecharles

edited Apr 20, 2007 - 10:52 AM

Some Linux have done that(however rudimentarily-- but even better: i wish they would just steer us to a better hardware platform / eliminate the bus bottleneck.

Score: 0

By Niro

posted Apr 20, 2007 - 12:16 PM

I don't understand...you actually WANT AMD to go under don't you??

Score: 0

By horsecharles

posted Apr 20, 2007 - 8:43 PM

Not at all-- they've now fallen so far behind, that IF their next processor does not compete with Intel-- they'll be in dire straits; in the face of that, why not take a chance on a revolutionary & much improved platform, that if it sticks, will allow them to RULE.

Score: 0

By Alex Stevens

posted Apr 21, 2007 - 9:27 AM

"why not take a chance on a revolutionary & much improved platform"

You mean like the Itanium? It won't work, if people can't run their 20 year old game of Reversi they freak out and won't buy. Whatever AMD does it has to be x86 compatible.

Score: 0

By horsecharles

posted Apr 21, 2007 - 10:41 AM

Folks would leave x86 if/when they saw the performance difference-- IF some current app/game/etc. were compiled for that platform-- huge IF.
I was hoping, one example, somehow the cell processor(esp. with for example 8-16 gb ram) & some Linux variant somehow made it from ps3 to desktop....pipe dream.

Score: 0

By Alex Stevens

posted Apr 21, 2007 - 1:08 PM

It would have to be supported by Windows though to be successful.

Score: 0

By horsecharles

edited Apr 21, 2007 - 7:06 PM

Yes, and that wouldn't happen...brings to mind how Windows was supposed to support PowerPC, then didn't, staying w/ 486/Pentium.
Had that happened, we'd all've been on 64bit & supercomputers for a long time already-- & likely Cell processor would also be supported by Windows.

This delay in getting to 64bit is the only chance for another OS to vie with Windows, but that window is about to shut tight. The best realistic chance was stupid Apple, but they not only chose the wrong processor, they pledged monogamous fealty to it and to one single company.

I'm still hoping someone with another OS and some multimedia app or game specially-compiled will do a demo on some advanced rig with 4 or more processors each with 8-16 cores & say 16gb ram...show the world how that runs circles around...
There's zero interest in porting Cell to desktops, not even in discussing or trying....

I'm not bashing Windows-- just trying to get us all on more advanced hardware & fostering more competition....

Score: 0

By domino360

posted Apr 20, 2007 - 2:11 AM

Hey AMD, the more you "outsource to IBM" the less original you become. Which means that it's time for IBM to make a move.

Score: 0

By Alex Stevens

posted Apr 21, 2007 - 9:29 AM

AMD started out making clone chips, like the 8088. If it weren't for licensing Intel's chips they wouldn't even exist. When they finally started making their own processors they could barely compete, until they got lucky with the Athlon.

Score: 0

By foxfyre

posted Apr 20, 2007 - 1:28 PM

The irony is that the vast majority of their past 'successes' have been a direct result of their licensing IBM technology!

Where do you get the idea that AMD is a major research based organization?

They are simply going BACK to a model that was successful in the past where IBM produced much of their product. It makes complete economic sense to piggyback upon IBM's economies of scale.

Score: 0

By Heero

posted Apr 20, 2007 - 4:12 PM

Of course it does...

AMD doesn't have the same cash pile that Intel does. AMD can't afford to sit around for three years beating a dead horse like Intel did.

Totaly 100% agree there.

Score: 0

By PostDeals

posted Apr 19, 2007 - 11:14 PM

Intel does have you beat, especially with their new offerings and architecture. I may go back to intel on my next processor purchase. It was a great run while it lasted.

Score: 0

By horsecharles

edited Apr 20, 2007 - 10:58 AM

I did already, but don't know why i jumped the gun(where's the performance improvement?) and not gave 64bit more time to go mainstream(well yeah, i actually know: it's gonna take 2/3 years for 64bit to hit mainstrem). I can already foresee this new purchase becoming outdated very soon(as far as running 64bit apps that is).

Score: 0