EW: (1) will Intel's profits hold up; (2) low margin on cheap chips; (3) threat of commoditization.
(1) Who knows? Too uncertain. Too many unanswerable questions. Don't worry about it. If you knew the answer you could make some real money. There is no sure way of making real money, therefore you cannot know the answer. I do wish they would stop buying stock back, and use those billions to buy up selected semiequips.
(2) Low margin on $50 chips. The costing of chips is arbitrary. Old process and plant = low cost (depreciated plant). On a marginal cost basis the new chip of a given performance is always cheaper (more chips per wafer), bigger wafer (fabbing cost tends to be proportional to number of wafers, not area of wafer). Thus if one freezes a nice design (say 400 MHz with 512k L2) it should be possible to shrink the size smaller and smaller, and reduce avoidable costs roughly proportionally. These cheap chips can only become commodities if there are other competitors. Where is AMD or NSM going to get $4 billion to build a new fab to produce $40 chips? They cannot produce such a chip at $50 with what they have. If it were merely a matter of cheap but adequate chips I saw a 486DX2 for $4, but I don't think it was selling.
(3) As long as design and architecture improvements continue, the combination of new capabilities and faster processing is going to offer advantages. What's the saying my old auntie used to cackle: "Microsoft makes Windows slower faster that Intel makes Pentiums faster.". As long as I am boss, I am going to spring for extra MB's, extra GB's, and extra MHz -- hell, it only costs a dollar more! When I recruit one of those $80,000/year grads I give her one too. Damn the expense! The encouragement and speed might increase her performance by 1%. Speed sells! I don't really need 8 cylinders, but what the hell. This is my axe.
|
