To: Robert Douglas who wrote (8241 ) 2/16/2000 5:00:00 PM From: Katherine Derbyshire Respond to of 10921
>>I'd be interested if you could provide any of your model's assumptions behind this prediction of overcapacity beginning later this year. << The key assumption is that chip makers started ramping their capacity purchases in July 1999. >> For starters, do you happen to know how many fabs will be coming on stream and how much they will increase aggregate capacity (in percentage terms)? Also will production capacity be increased by other means, such as larger wafers, and how much will that increase capacity?<< No, I don't have good estimates for how much capacity will be coming online. But all the major players have announced significant expansions, and some, particularly foundries, are continuing to spend far more on capacity than the sustainable 15-20% of revenues. No, I don't expect 300 mm wafers to play a major role in this cycle. Shrinks are continuing, of course. >> Secondly, what are your predictions on demand growth for the next 2 years? Are they substantially below the increase in capacity that you are forecasting?<< I think we're in the process of a shift in demand from PC-driven to communications-driven. More important to equipment dynamics, though, I think we're also seeing a shift from single-supplier production (PC MPUs) to commoditized production (foundries). In the DRAM market, demand keeps climbing in unit terms, while equipment purchases ebb and flow with chip selling prices. I would expect foundries to bring this kind of dynamic to other segments, too. The question isn't really *whether* capacity will outstrip demand, but *when.* The faster chip makers build fabs--and this most recent quarter for AMAT suggests they're building very fast--the sooner they'll catch up. Katherine