I also am worried about AMD and its "delayed" 130 nm process, but looking at some financial numbers, let's look at one alternative.
Inventory of finished goods was $90 MM at 3/31. This implies something like $70-75 MM in CPUs (assuming share of non-other revenue) or roughly 2 MM chips. Now to sell 8 MM chips in Q2, AMD would sell 2 MM in April, 2 MM in May and 4 MM in June. Inventory from 3/31 would mostly have been gone by May 1 (except for mix issues). So, you say, that would have been good time to have 130 nm chips. But let's look at production. It was say 10% of total wafer outs in early April moving to 50% my end of June - based on charts shown at recent conferences and shareholder meeting (my memory could be off here, but not by much). So roughly 15% of April outs should have been 130 nm which at roughly twice the die per wafer would mean that 130 nm would have been .3/1.15 = ~25% of chips in April, or assuming chips out=chips sold in a quarter with inventory flat, this would be about 700,000 chips by the end of April. Now, we know many of these have been sold into laptop channels which means that they hardly had enough for a major product launch. In May, it should have been about 30% of wafer outs or about 50% of chips out or about 1.4 MM chips. Now assuming a few weeks to a month to out the door, only now is AMD getting sufficient quantities of chips to release desktop Thoroughbred.
I only half believe the above is the reason for delays although it is highly plausible (remember, marketing folks always win over tech guys).
Jon |
