>>Fab throughput (supply not as high as feared)
It is true that the next 300mm wafers with 0.13µm gates can contain approximately 5x the chips that the present 200mm 0.18µm wafers carry. Yet
it is also true that because of the single wafer processing, and more process steps, that same 300mm wafer will take longer to process. <<
Not much longer. The fabs have set equivalent wafer throughput as the benchmark they want equipment companies to meet. That is, if a tool processed 100 200-mm wafers per hour, the equivalent 300-mm tool should also process 100 wafers per hour. If a 300 mm fab *can't* make substantially more chips than a 200mm fab, then there's no reason to build it in the first place. You may not get 5x, but you'll get a lot of chips. This whole argument is a red herring, though. Companies will build enough fabs to meet a given capacity goal. If their goal is too high, overcapacity will result no matter what wafer size the fab runs.
>>Sematech does not have everybody in step anymore. <<
They never did, because only a handful of companies were Sematech members. Sematech actually has *more* members now than they used to, and the offloading of technology expertise to the equipment companies has leveled the playing field to a certain extent.
>>Not all the chip companies have embarked on 300mm, and those that have are on different time schedules, and can halt
300mm build out should over capacity raise its head. <<
Leaving them with a multi-billion dollar hole in the ground and no way to recoup the investment? I doubt it. They're more likely to try to build out faster in order to beat the other guys.
>>Additionally, it has
been pointed out by Sue Billat at Robertson Stephens, some semi equipment is in tight supply and will naturally slow down capacity increases.<<
Yes, I read that interview, too. I don't recall any implication that cycles were a thing of the past.
>>An increasing diversity of market segments and chips (different supply and demand)<<
This is the most solid argument you make. However, letting different types of chips share a single capacity pool is not necessarily a good thing. Too many DRAMs and not enough DSPs? Okay, we'll make DSPs instead and create a glut there, too!
>>Cheaper processing power means demand for more chips (trend for increasing demand)<<
DRAM sales reached a record high in bit terms in 1996. A down cycle can easily coexist with strong demand.
>>Perhaps some might argue, the time to build
capacity is so long, that getting an overproduction signal early will not help.<<
Yes, exactly.
>> But here too, the 300mm fabs are special. These fabs will be able to
shift production to other chips. Future 300mm fabs will be nimble because 300mm fabs will be completely automated.<<
This makes no sense. First off, fab automation doesn't do much for process flexibility. Either your process can run other chips or it can't. Very specialized equipment does not become more flexible by simply connecting robots to it, and buying new tools is not a simple matter of going down to your local hardware store. Second, ability to shift to other chips actually increases the risk of global overcapacity. When one segment gets flooded, the overflow takes down other segments instead of just sitting idle.
>>Such systems can make the introduction of a new process for new chips almost painless.<<
Ever worked in a fab? I've never met a single process engineer who would agree with this statement.
Katherine |
