Re: what's the maximum revenue AMD could take in without expanding?
Most of AMD's current parts are around 100mm2. They're going to be adding transistors, but shrinking the feature size, so the die size may not change much. AMD's theoretical capacity for 100mm2 die from the two FABs, figuring 11K wafers/week, a yield of 60% and an edge loss of 10% is about 24 million per quarter (from Austin and Dresden). If AMD brought its ASP up to $150 (Intel is presently over $200), and sold all those parts, its annual revenue from CPUs would be $14.4 Billion.
AMD has a lower cost structure than Intel, the scenario depicted above requires no new plant, about a thousand new hires at Dresden, and some more silicon ingots, marketing dollars, etc.
It's not gonna happen, but in such circumstances AMD would probably net $7 or $8 Billion. With a share float of 150 million, that's $46 per share earnings. Even with a multiple of only 25 that gives a stock price of $1,150.
Enough fantasizing, my oft repeated estimate is $9 earnings and $280 stock price for the end of this year. Paul's estimate is lower (I think he expects a loss). The reality will likely be somewhere between the two.
Regards,
Dan
PS - You might also note that AMDs yields may be better than 60%, and that Dresden was supposed to be layed out to allow for easy expansion to 4 times the current size.
On the other hand, actual capacity and theoretical capacity usually differ a great deal. Intel is cranking out up to 1 million coppermines per week from each of 5 FABs, and the parts seem to dissipate like the morning fog on a sunny day. Maybe the other 4 FABs produce a hundred thousand parts a week, each? |
