Hi Ohkami; Re: my comment that the accounting periods of Micron and Infineon don't correspond. You wrote in response: "Yes, but since DRAM price went down even further in September, this would mean that IFX' write-down would have to be larger, not smaller than Micron's. It works in my argument's favour."
My figures for SDRAM pricing for the time period covering the two quarters (i.e. Micron's ended on August 30, while Infineon's ended on September 31) in two quarters are as follows (with dates rounded to previous Fridays):
SDRAM pricing (See #reply-16864878 for numbers.)
05/28/01 $33.20 (Start of Micron's quarter)
06/29/01 $18.80 (Start of Infineon's quarter)
08/31/01 $16.60 (End of Micron's quarter)
09/28/01 $14.80 (End of Infineon's quarter)
In other words, during Micron's quarter, DRAM dropped from $33.20 to $16.60, a reduction of $16.60 per unit. (In this case a "unit" is the average of the 5 cheapest 256MB PC133 DIMMs available on PriceWatch. I've been keeping these numbers since March last.) During Infineon's quarter, DRAM dropped from $18.80 to $14.80, a loss of $4 per unit.
Thus Micron's quarter included a DRAM loss in sales price, per DRAM, which is about 4x as large as the quarter for Infineon. So much for that argument, I think we can leave it alone.
Re: "Carl, as an engineer I think you really shouldn't use loaded terms like "delusion" in an argument." I'm getting tired of this. I've shot down your argument about write-down prices fairly completely, agreed? That was easy because accounting is a fairly simple thing. The problem with shooting down the argument on feature size is that it requires me to extract total costs for DRAM production. That's work, and I'm not going to do it for you. Suffice it to say that I've had plenty of chips priced by various chip houses with alternative processes, and the older processes are generally considerably cheaper, on a per transistor basis. Everyone in the industry knows that the reason the older lines are cheaper is because of the following: (1) The older equipment has already been amortized. (2) The older equipment was cheaper when it was new. (I.e. the cost of outfitting a fab keeps going up.) (3) The older processes have been running longer and consequently run with less bugs. On the other hand, it is true that the the smaller feature sizes allow more efficient use of silicon, but that is not the only contribution to chip cost. This is a very complicated thing, and you still haven't given a cogent argument why Micron's lines cost more to produce DRAM than Infineon's. If you want to do a comparison, you should find out the equipment that Micron uses, figure out how much Micron paid for it, and look up how many DRAM chips it makes per month. Then do the same thing with Infineon. I know you're not going to go through that effort, but do you really think that the industry doesn't? For you it would be a lot of trouble to do an accurate estimate of Micron's COGS, but for the people in the industry it's fairly easy. The information I've referenced above is publicly available, but you'll have to dig pretty deep to get it. The people in the industry who did do the analysis always say the same thing, and that is that Micron is cheaper.
When I have a chip in production, it's true that when they migrate the line to a smaller feature size the price (and the cost) goes down. But that's using the same line. If I instead migrate the chip to a different line, whether the cost will go up or down is a complicated thing and is not easy to analyze. But I've priced enough full custom chips to damn well know that I can't compare the cost (or price) of two dissimilar lines on the basis of the feature size. I'm sure I can dig up quotes on this, but I will only do it if you cry "uncle" on the argument over the Micron and Infineon write-down figures. If you refuse to give up on that, I will conclude that you're hopeless.
In reference to my suggestion that a better metric would be comparing the ratio of DRAM sold to the cost of capital equipment, you wrote: Re: "That's a good metric. Why don't *you* do that comparison for a change? I'm willing to accept the imperfect comparability and am looking forward to the results :-)"
There's problems with that metric as well. Companies amortize their equipment over different periods, etc.
In reference to my suggestion that you compare various costs &c., you wrote: "Because all of these costs are minor portions of total costs." Micron has 18,000 employees. In their most recent quarter they had $424 million in revenue. That works out to be $23,556 per employee per quarter or $94,222 per year. In other words, labor costs for Micron are very significant. This is a common problem I have with your arguments, they don't include a lot of facts.
Re: "why everyone, including Micron, is always staying at the cutting edge concerning line-width. Those that fall behind ultimately have to exit." Engineers have another word for the "cutting edge". We call it the "bleeding edge", as a play on "leading edge". Micron does not now, nor have they every stayed at the leading edge. They've always been a bit behind, and they've done just fine. Samsung, on the other hand, does play on the bleeding edge. That gives Samsung an advantage on extreme technology DRAM, like Rambus stuff, but it gives them a pricing disadvantage on commodity DRAM. The whole industry knows this, and it's referenced every now and then in articles.
Re: "Next, I would kindly request that you come up with more convincing arguments as to *why Micron is the lowest-cost producer* besides the statement that "it is acknowledged in the industry" to be so, which is the only constructive argument I have heard from you so far - and that doesn't exactly fulfill this engineer's requirements for a good argument :-)" Forget about it, I'll pick your arguments apart, but I'm damned if I'm going to spend a week looking through Applied Material's (and everybody else's) website for press releases of equipment sales, figures for DRAM production per hour, etc. It's just too much work. Instead, I'll take the industry acknowledgement at face value. You're the one, after all, who's rebelling against the accepted "wisdom", not me.
-- Carl |
