To: Petz who wrote (3644 ) 8/4/2000 11:08:06 AM From: Charles R Respond to of 275872 Petz, <I'll concede that the CPU business is not very cyclic, but, to the manufacturer with less than 1/3 the market, it is very cyclic.> I would say it is more line "it could be cyclical" instead of "it is very cyclic". That's why I keep beating to death the importance of growing market share. A focussed company without anything else to fall back on will do this a "little" better. <Imagine if something like FDIV happened to AMD, even a year from now. There is also high technology risk for AMD. For example, we all know that if Pentium 4's turn out to yield to 2 GHz by the end of the year, AMD is totally screwed.> I think FDIV incident was a significant event in this business but the financial impact of it was insignificant measured as a percent of sales or profits - even that is a one time event. Investment community moves on. <In short, the CPU business has always been extremely cyclic for AMD, because there is no room for error when competing with the 800 LB gorilla.> I would phrase this as "there is not much room to keep making more errors than your competitor" <Finally, even a "growth recession" of less than 15% YOY growth could cause ASP's to tumble 30% and result in 15% YOY revenue declines.> If a 15% growth recession happens ASP decline will be much worse in businesses like DRAM, SRAM, FLash because there is lot more competition there. x86 is the best business to be in for ASP protection. <So, I like the fact that CPU's are "only" 70% of AMD, and I would like it even better if they were 60%.> I would prefer 100% in x86 biz with 40% in desktop; 20% in laptop; 20% in servers; 20% in embedded, chipsets and other. These market segments have different enough characteristics to give a solid base and make the company less susceptible to dramatic growth swings. Chuck