Online information source for semiconductor professionals

We all waited and waited and waited some more to find out from AMD and Hector Ruiz what exactly was its ‘asset smart’ plan. Now that it has been revealed, it is very close to what we and others thought it would mean.

That said, the bare details provided in the press release and on the conference call raise some important questions that in some respects should have been answered in the first place!

A simple one to start with was the lack of reference to the fact that in all reality, AMD has become a fabless company. Granted, it has a 44.4 percent stake in ‘The Foundry Company,’ but as the financial wizards will no doubt point out, the fabs have exited the balance sheets.

As any fabless company knows, dealing with foundries becomes an integral part of business operations and with that comes the need to have a ‘Director of Operations’ and a backroom team of engineers to handle that side very closely. In essence, AMD as a fabless company will have to have a unit dedicated to this. I am sure they have this in place as the ATI acquisition would have brought with it that set-up, but this was something the company didn’t mention. It is important, as a restructuring within AMD would or should be taking place to handle its new found freedom from manufacturing.

Going fabless is fundamentally a big gamble for AMD, something that will probably take time to unravel. A point some may not have realised is that the only other microprocessor company that was fabless is up for sale and only after it had become an IP company rather than a fabless company. There is a precedent being set here and that must make Intel very happy.

Another aspect that concerned me was the notion pumped out by AMD that demand for leading-edge foundry capacity was something that was in strong demand. Ask SMIC, Chartered or UMC how much of its capacity is allocated to 65nm-and-below production and you will find it is very small. The number of companies moving to smaller geometries also gets smaller by each technology node.

Also, if demand for foundry capacity was that strong, then why are wafer ASPs in decline and the major foundries cutting CapEx each year?

Even worse is the fact that SMIC has struggled since birth to actually turn a profit, so why should we think that a new foundry start-up in Europe will fare any better?

Importantly, why is there a rush for the foundry spin-off to equip a shuttered fab in Dresden and charge ahead with a new fab in Luther Forrest, NY? Fab 36 has obviously been able to handle all of AMD’s MPU demand and has barely used its then foundry partner Chartered for extra capacity requirements.

Part of that conundrum was answered in an email from AMD last night. Apparently, the graphics chips currently being produced at TSMC will migrate over to Dresden (Fab 38). However, the issue of why AMD/The Foundry Company see the need for the fab in NY has yet to be explained!

With over a $ billion in funding from NY State on the table, it may well be a case of "the fab can be built at zero cost, so why not go ahead?". Figuring out what they can do with it after that it is another matter, but they will have at least 12 months to figure that one out!

That brings me to another important point, something Bill McClean at IC Insights noted in an email last night. The CapEx plans for the foundry seem, on the surface, to be impressive at $3.6 billion to $6.0 billion over the next five years. However, Bill noted that:

Foundry Company’s 2009-2013 capital spending budget is forecast to be $4.8 billion (when using the mid-point of its announced $3.6-$6.0 billion spending plan).  The funds will be used to install capacity to produce devices for AMD as well as to enter the IC foundry business.  IC Insights believes that this financial commitment is uninspiring at best.  It should be noted that AMD’s capital expenditures over this and the previous four years (2004-2008) are expected to total $6.4 billion, and look at what shape the company is in today.

Intel is expected to spend $25.6 billion dollars in capital expenditures from 2004-2008, four times what AMD will spend over the same time period.  IC Insights believes that Intel’s capital expenditures over the next five years will total about $30 billion, which would be five times the Foundry Company’s most optimistic plan of investing $6.0 billion over that time! 

Uninspiring seems to say it all, but Bill had more to say on that matter.

Given that Foundry Company’s capital spending budget for the next five years is forecast to be less than AMD’s capital spending outlays over the previous five years, it is highly unlikely Foundry Company will help AMD gain additional market share.  In fact, AMD will be lucky to sustain flat market share in the MPU segment.  Moreover, a significant portion of the Foundry Company’s budget must be allocated to the new foundry business, and this, of course, will take away precious spending needed for MPU production!  

Obviously, AMD has survived to fight another day by going fabless, but side-stepping the unnecessary hype in which AMD chose to announce the news, AMD will continue to struggle and its foundry friend doesn’t look set to do any better.