Now Hear This!

Offering news and business analysis for the design engineer, Managing News Editor Suzanne Deffree filters the electronics industry's developments and trends to explain how what's happening in the board room today can impact the tech innovation of tomorrow.

Advertisement

Profile

About Me

RSS Feed

Add this blog to your RSS newsreader!

By Category

News Center Articles

Blog

Now Hear This!

« Moto’s mobile spinout an... | Main | This week in gEEk: Reflections... »

Thursday, September 4, 2008

AMD, speak now or forever lose your piece of the MPU market

Sep 4 2008 3:17PM | Permalink |Comments (13) |

It’s Q3 -- the tail end of the quarter, in fact -- and AMD has yet to make a firm statement on its manufacturing plans. While Hector Ruiz said he would be “the most disappointed man on Earth” if AMD hadn’t shared plans for its manufacturing strategy by the end of the year, many industry watchers and participants are expecting news on the chip maker’s plans by the end of this quarter and their wariness of the company is growing as each silent week passes.

Perhaps the headline on this blog post could have been: “AMD, speak now and be split into pieces.” More and more analysts are predicting the breakup of AMD into two entities, one design and one manufacturing, with the most recent including some sound processing benefits in his report (see, “AMD fab-lite strategy announcement in 2 weeks, analyst says”).

As this blog has stated before, if AMD plays its cards right with manufacturing, it could have the finances to return as a strong challenger to Intel. If not, then it’s just a matter of time until the game is over for the debt-ridden company.

And manufacturing is looking to be the final ingredient in AMD’s financial recovery medicine. To the company’s credit, it has made some tough choices in the past nine months to regain its financial footing and move toward the profitability it promised by the end of the September quarter. It tackled the needed management changes (pop Hector out of the power position and pop in Dirk Meyers). It cut headcount substantially to lower expenses (10% of its employees will have been handed pink slips by the end of Q3). And, instead of clutching to the idea that it can be a everything to everyone company, AMD is selling off its lackluster consumer units to companies better suited to drive those markets and is gaining cash infusions in doing so (like Broadcom’s near $193 million buy of AMD’s digital TV unit, which, by the way, also lowered AMD’s headcount by 530 employees).

Note that we’re discussing a financial recovery here. The impact of the above choices on the company’s ability to innovate is a whole different animal, one that might come back to bite AMD at a later date.

Meanwhile, all eyes will be on AMD in the coming weeks to see if real men spin off their fabs and, if they do, how TSMC, UMC, Chartered, and the rest of the foundry world react to a separate AMD-groomed manufacturing operation.

How would you react to such a split? Share your opinion below.

--Suzanne Deffree, Managing Editor, News

Reader Comments

at 9/5/2008 7:05:42 AM, Beenthere said:
As the pop-up add says, you need to "GET IN TOUCH"...

It really makes absolutely no difference what AMD's intent is on "asset lite" as long as they continue to supply customers with the products they desire. AMD has been doing this for the past 30 years and shows no signs of changing that mode of operation any time soon.

at 9/5/2008 8:59:57 AM, Joe Money said:
No money, no products. See the forest for the trees.

at 9/5/2008 3:44:11 PM, Jim McGregor - In-Stat Research Director said:
Many electronics companies and semiconductor companies have exited the manufacturing business and many other plan to over the next several years, including TI and Dell. The real question is will AMD continue to invest in manufacturing process technology through the IBM alliance? This expertise is critical in designing some of the most complex semiconductor solutions and remaining competitive with Intel.

Also, focusing on one's core competencies is always a good strategy, as long as the market also supports that direction. We have already seen the shift to mobile computing, and now the market is shifting towards low-power solutions, an area where AMD is trailing not only Intel but other vendors as well.

So, while these moves may provide a short-term benefit to AMD's financial position, AMD's ability to innovate and execute under dynamic market conditions will determine the company's fate. From my perspective, I am still awaiting a glimpse of AMD's vision to make that judgment.

at 9/5/2008 4:16:23 PM, Suzanne Deffree, EDN said:
Excellent points, Jim. Thanks for the comment.

at 9/5/2008 6:11:06 PM, Nathan Brookwood - Insight 64 said:
Companies typically move from integrated to fabless models because (1) they no longer want to make the substantial investments needed to remain competitive in process technology (e.g. TI), or (2) they lack the scale to operate a fab efficiently with high capacity utilization. Neither of these factors applies to AMD

at 9/5/2008 6:13:25 PM, Nathan Brookwood - Insight 64 said:
Companies typically move from integrated to fabless models because (1) they no longer want to make the substantial investments needed to remain competitve in process technology (e.g. TI), or (2) they lack the scale to operate a fab efficiently with high capacity utilization. Neither of these factors applies to AMD; their relationship with IBM gives them access to next generation technology at a fraction of what companies going it alone pay. And if, as Jim notes, AMD has a competitive product line, they certainly can fully absorb a big fab's output. AMD has demonstrated that they know how to run a fab efficiently, at least by the usual Sematech benchmarks. Thus it's not at all clear that outsourcing its production could help AMD be more competitive, and it might really harm its ability to compete with Intel.

Conversely, it's not at all clear how AMD can pay for the tools it needs to finish fitting out its 45nm lines in Fab 36, or the conversion of Fab 30 to 38. The last time I looked, equipment vendors like ASML and Applied expect to get paid soon after their tools have been installed and qualified.

What AMD really needs is a way to finance the purchases of the tools it needs to stay in business. It maxed out its credit cards with its cash purchase of ATI and a disasterous 2007; even Countrywide no longer writes mortgages without adequate documentation. Last year, AMD sold stock to a sovereign wealth fund to add to its equity, but the fall in AMD's share price since then would make those funds reluctant to provide more cash on an unsecured basis. These sovreign funds have huge cash inflows that result from the huge cash outflows we provide every time we fill up our SUVs, and they need to find places to invest those cash flows. Investing in high-tech fabs could provide the income these oil-rich groups will need when their wells run dry.

When I arrange all the above tea leaves in my cup, they say "sale/leaseback" to me. AMD sells its Dresden site (and all the equipment in it) to a fund with deep pockets, and then leases it back. This gives the company the cash it needs to fund current operations. (Think of it like buying a new car and getting "cash back.") AMD pays rent instead of depreciation, but it gets to stay in business and gradually repair its balance sheet. Putting a deal like this together is non-trivial, and getting the approval of government agencies in the US, Europe and the middle-East would be even harder. No wonder AMD's had problems predicting when it will be able to put the final pieces in place. Let's just hope they can pull it off.

at 9/5/2008 6:15:06 PM, Nathan Brookwood - Insight 64 said:
Sorry- my first try didn't work out quite right.

at 9/6/2008 8:13:18 AM, Train_wreck said:
Hector, Wrecked AMD with his purchase of ATI. Just when AMD had a little cash he overpays for a graphics company! Hector was never a good manager when he was at Motorola and obviously failed at AMD. Only thing Hector did well was line his own pockets! That he did very well. Congratulations Hector you've ruined another company.

If AMD survives this it will be a miracle. I am saddened at the impact that Hector the wrecker has caused. A good company making good progress run into the ground.

Intel will finish catching up to AMD by Q4 of this year. With better fabs, as good or better design and more money AMD will be forced to the fringes again.

Sad, sad, sad.

at 9/6/2008 10:03:19 AM, Vattila said:
McGregor wrote: "The real question is will AMD continue to invest in manufacturing process technology through the IBM alliance?"

A recent 8-K filing from AMD extends and broadens the IBM co-development deal through 2015.

Brookwood wrote: "Putting a deal like this together is non-trivial, and getting the approval of government agencies in the US, Europe and the middle-East would be even harder."

This scenario ignores the cross-patent licence with Intel

at 9/6/2008 11:38:05 AM, Nathan Brookwood - Insight 64 said:
I forgot to add - Hector - I still haven't received my paychecks for the last 9 months for offering my good opinions to the media about AMD.

Nathan

at 9/8/2008 1:23:13 AM, Ian Proffitt said:
AMD has never been able to shake off the "also ran" tab.
As Apple is to Macintosh so AMD is to Intel, always seen as "me too" copycats in a market where real innovation is rewarded handsomely.
AMD doesn't need a manufactury and it will probably help them to develop if they divest themselves of what has become a cash sink as the market drives fabs up the technology curve.
The big question to me is if AMD do redefine their business model - will the revamp cost AMD in terms of hard cash and reputation or can they turn the split up into an opportunity?

at 9/8/2008 1:24:36 AM, Ian Proffitt said:
Typo - "As Apple is to Microsoft"

at 9/9/2008 3:55:26 PM, Raven said:
I don't think that TSMC or UMC
have to worry about an AMD Fab
gaining any market share within the first five years.
AMD has massive problems the biggest is Hector. I don't know him he's probably a nice guy and a great engineer. But
he lost control of the comapny.
Of course Intel did not help, nor should have. May I ask why you have not address a possible Duba second round investment ? What about the opportunity of new fab in Duba?

Now Hear This!

Profile

RSS Feed

Recent Posts

Recent Comments

Most Commented On

Archives