Re: That's an interesting fact. 4.4B$. Let's see, AMD is building one 65nm plant only to keep up with Intel's manufacturing prowess to be on line in 2007.
Not quite. They've started mass production at a new FAB expected to take their market share from 15% to 30%. It's a FAB designed to shift to 45nm production, too.
They'd be uncompetitive without it, and they're not sure they can pay for it, b.c they're not sure they can sell the volume it will generate.
Uncompetitive how? AMD is presently gaining market share steadily. Intel has been unable to put an SOI process into production and because of it Intel is no longer competitive with AMD in several key markets. The result has been a steady and accelerating loss of market share for Intel, particularly in several very profitable and visible segments - servers, workstations, and gaming machines. While there is always some uncertainty about the future, I expect AMD is more confident of good sell through than Intel is.
Given the growth of the uP semi industry to 2X in 2009 what it is now (AMD's estimate), AMD will certainly need to start another plant at 45nm in 2008 for operation in 2010. That's only two years away.
Actually, AMD announced at its recent conference call that it planned to start building an additional new FAB early next year to be in operation in 2008 (likely late 2008). AMD has also been working to shift some CPU production to foundries, further increasing its capacity. AMD expects to have significant foundry production of CPUs late next year.
And, that plant can't be co-located with the other plant so there are no synergies.
Actually, it could, but I agree that it would be better for AMD to locate their next FAB, that they plan to break ground for next year, somewhere else - maybe next to the Austin flash plant and design center?
AMD will have to beg, borrow and steal (read build in Germany with subsidies, borrow B$'s, win the lawsuit) in order to do this. Or, they will have to be *exceptionally* successful in winning market share from Intel in 2006 (read 50%). I find this to be unlikely.
I'll be surprised if they don't continue to take market share from Intel. And their current share is actually plenty to fund another FAB as long as ASPs keep increasing. Just how much share AMD will gain next year is something we'll all know for sure in the coming months.
Move up 50,000'. There aren't many companies around that can build the kind of manufacturing capability needed to manufacture semis at these levels. That's why Samsung is pouring 10B$ into infrastructure. There will only be a few large manufacturers and others will have to pay the price for fab outsourcing.
I agree. Even if they continue to cede market share and profits to AMD, moving from a near total monopoly position to more of a duopoly, Intel should continue to make quite a bit of money.
It's a lot like the GM of the 40's, no?
More like GM of the early 70's, with quick, flexible, and efficient AMD acting as Toyota/Honda/Nissan to tear share away from the once near monopoly producer. |
