Advanced Micro Devices Inc (AMD)

[sgolds]

Elmer, I will decline your challenge. Last time I worked in the semiconductor equipment biz was 1985, as you know my expertise is in computer networks and security. Thus I am not shy about saying that I cannot name the steps in the conversion process.

I do have a good feel for what is reasonable, though. Converting from 130nm to 90nm in two quarters with only 10% of the factory down at any one time - that just is not reasonable.

Here is the rub: In your earlier post, you stated that you figure 10% for development. Let's see, your words were I and others in the field have estimated 10% for development purposes. Now, if development means pilot lines for tranferring technology from engineering to manufacturing and tuning the process ahead of mass production then I would concur that this does not exceed 10% of capacity.

I just do not buy that you can convert the whole factory in two quarters with only 10% taken offline.

Here is a comparison point: When Intel takes down a factory for their 'copy exact' method, how long before it is up and running with the new process? Since Intel did all the development in a separate pilot fab, this should give a rough idea of the magnitude of the conversion task. Remember, they take 100% of the factory offline.