Elmer,
Intel should be producing at least that many processors from each of their 300mm fabs if they are running at full capacity (what is their capacity anyway?) and have good yields. Do you know if they are running at full capacity?
Underutilized fab capacity is not preferable to lower margin production: Message 20224540
I will also tell you that processors are not the only products running on 300mm wafers.
Good point. There may be other uses for the state of the art 90nm silicon, worht its weight in gold. For example, this particular product, Sponge Bob Square Pants Bikini Bottom Boat Race amazon.com could benefit from such silicon.
So if you don't know what products are running in those fabs and you don't know the capacity of the fabs and you don't know if they are running at or near capacity, how are you to estimate anything?
Well, you said it makes no sense not to run at capacity, I told you that it makes no sense to use the highest end process technology for low margin production (non-CPU) production. Therefore Intel must be suffering disastrous yields.
Or could it be that making and selling more, at lower prices, would not increase revenues by the amount greater than the increase in production costs?
That's still well below the capacity of F30. How do you explain that? Is it demand limited?
At volume of 10M processors of average die size of say 100 to 120 mm^2, would probably be close to production limits (with good yields), IMO.
So limited demand for both K7 and K8 has been the reason that output was less than potential, in your view. You're the first person here who is willing to acknowledge that possibility.
I have been saying it for over 3 years now. I think K8 was demand limited right through Q1, and now, in Q2, we may have an unexpected surge in demand, which AMD can barely fill, given the number of wavers they started.
The new 90nm supply will need to generate new and greater demand, assuming there is 90nm supply.
Supply will need to generate new and greater demand... I guess we are all supply siders now <g>. (Both Reagan and Nixon turn in their graves, Reagan smiling, Nixon not)
The way I see it is that there are 2 entities out there, the old $70 ASP AMD and the New $150 ASP AMD. The 130nm K8 just seeded the market of the new AMD, 90nm K8 will deliver the volume. I am not too concerned about the sum of units sold by the old and the new AMD. I am mainly interested in the New AMD going up.
Back to 90nm supply generating new demand: yes it will, to a degree, but 130nm is already doing fine, at least this Q. 130nm just can't deliver enough of the supply longer term.
The areas where the new demand will be generated is mainly in notebooks, blades and servers, which will benefit from somewhat lower power consumption.
I think it's time to address the point that AMD will need to take market share from Intel if they are to make any reasonable headway.
Unitwise, it is not really necessary for AMD to take market share away from Intel for AMD to do well. The overall market is growing well (BTW, even in post 9/11 NYC, the recession was just today declared over as the city grew by 7% (annualized) in the last quarter).
But the bigger conributor is AMD moving up with ASPs. Where Intel's ASPs will go, I don't know, maybe stay the same, possibly even going up. One indicator where AMD will likely make headway is $ market share, but in growing market, Intel may not even notice.
Joe |
