To: economaniack who wrote (219931 ) 12/10/2006 5:19:40 PM From: kpf Read Replies (2) | Respond to e I am sure I have made some painfully stupid comments over the several years I have posted here. I don't find this one even a bit embarrassing. AMD lists R&D expenses of roughly $1.1 billion for the last year, so I was off by about 30%. I wish you would explain how you got a breakdown of spending between product specific expenditures and process development. It would be something useful to know. If you are trying to justify the 10 figure number, I would happily accept that AMD's spending on process development is likely over $800 million per node and could be $1 billion. If that is what you meant I have no great quarrel with it, just would note that Intel's R&D is roughly 5 times greater than AMD and could also be described as 10 figure, although that greatly obscures the difference between them. How much do you figure Intel is spending on process R&D? Off hand I don't see a particular reason to conclude that Intel's spending per die has been meaningfully different than AMD's whether total cost, total R&D or some subset is the relevant measure. If you have some reason to believe otherwise I would appreciate your comments. I definitely wanted total cost for this comparison. The whole point of the discussion was to evaluate the proportional impact of fixed costs related to process development. In the end, both AMD and Intel must cover all the costs of developing and selling processors, and this is the measure of average cost that is relevant to evaluating economies of scale. Perhaps you could quibble that Intel uses Marketing and GA in ways that are better viewed as reducing revenue rather than increasing cost, and I would be happy to entertain an argument that that changes the general conclusion. In any event COGS omits even the direct costs of product development.
