Thanks, Gottfried.
1. I can't think of a reason why outsourcing, by itself, should lead to a lower ratio of semi-equip/chip sales. Less equip spending by chip companies that outsource, should be exactly balanced by foundries buying more equip (other variables being equal).
2. I used to follow the memory chip companies, as the "canary in the mine", with their inventories and pricing indicating where we were in the cycle. Now, foundries are probably as good an indicator. Or maybe consumer spending.
3. Is the chip cycle getting less cyclical? With chips in toys, cars, cell-phones, and everything else consumers buy, are the peaks and troughs of demand for one end-use of chips, balanced by different peaks and troughs for the many other products? Or is the chip cycle simply getting correlated with the general economic cycle, and we've just been lucky enough not to have a severe recession for a long time?
4. If the chip cycle is getting less cyclical, that would be a reason for chip companies to be able to predict future demand better, maintain higher fab utilization rates, and lower the semi-equip/chip sales ratio. Is this what's happening?
5. Under the hood of my new Prius, the box with the electronics is about as big as the engine. Green products = sensing and controlling more variables in the product's performance = more information storage and processing = more chip use. The fraction of a new car's value in chips, has been steadily increasing. This, IMO, is today's impact of $4/G gasoline, and global warming, on AMAT. The photovoltaic business (growing rapidly from a very small base) might have an important impact, at some unknown point in the future. |
