Elmer, re:Celeron die cost of 21.25 and total cost of $42. Exactly how did you arrive at total cost? Did you divide R&D + Depreciation by the total number of chips Intel shipped? Supposedly Intel shipped 8M Celeron's last quarter, about a third of total production, so you are saying that Intel's depreciation + R&D was 3*(42-21.25)*8,000,000. That comes to a little less than $500,000,000.
Unfortunately for you, Intel's R&D alone (they don't break down depreciation) in last quarter's 10Q was $617M! You really should apportion at least 25% of "marketing, general, administrative" to the Celeron also. That was $766M.
So if we take 25% of M.G.A. and 33% of R&D, the "non-die" costs of the Celeron, assuming 8,000,000 were sold, is 0.25*766M+0.33*617M=$395M. Divide that by 8,000,000 Celerons and the "non-die" cost per chip was $50, not $20. And yes, I didn't even count depreciation!
The true cost of each Celeron chip is at least $70, not $42. The lower numbers seen in analysts reports are marginal costs to produce additional chips, and assume that MG&A, Depreciation and R&D don't rise with higher volume.
Petz |
