Today ML smashed all the DRAM guys - they don't look at cost structure of the individual companies they smack I suppose.
All the 200mm stuff can be completely written off - it does not do any good if they can't recover variable cost.
All the TW companies and Elpida drive very hard 300mm and they might just smell blood because they invested heavily in 300mm and they keep ramping (I am convinced their goal is to drive 200mm out and gain share - of course gaining share is a never ending battle).
This could get very dynamic in case certain supply is taken out of the market (they probably won't cut back on 300mm but reduce output from 200mm);
the change in wafer size and the economics related to wafersize makes this DRAM glut (IMHO) a very unique situation which is being ignored by sell side pundits. Just imagine 5 to 10% is driven out of the market.
DRAM makers start reducing production
Printer friendly
Related stories
Comments
Email to a friend
Latest news
Hans Wu, Taipei; Rodney Chan, DIGITIMES [Wednesday 23 May 2007]
Some DRAM makers have started cutting production for DRAM memory in order to reduce their losses from fast falling prices, according to industry sources.
The makers also hope that the reduction in production will help DRAM return to a supply-demand balance in the third quarter, the sources added.
The sources indicated that although DRAM prices have shown signs of stabilizing after falling fast over the past few months, makers still face heavy losses from the memory segment.
Current prices have dropped to an average of about US$1.10, close to the DRAM production cost at 12-inch fabs on 90nm processes, the sources said. Any further declines of US$0.20-0.30 would mean that makers would rather stop production completely than lose money, the sources commented.
Splinter quotes that close to 50% is still @200mm which seems to be high compared to LRCX guess.
Michael R. Splinter
In the memory area, the number of projects is actually going up. The timing of how these get completed I think is always a question but the DRAM guys continue to invest as bit growth grows. Still a huge amount of the capacity is on 200-millimeter, someplace between 40% and 50% of the capacity on 200-millimeter.
We are expecting over the next couple of generations that we will need at least 10 to 12 new factories to take over the total 200-millimeter capacity that is currently in flash and in DRAM.
So the DRAM projects continue to grow and they continue to create new ones. Foundry right now is much more filling in capacity that’s open in buildings around the various four or five foundry companies that you would track in a group.
Robert Maire - Needham & Company
Just as a follow-on to that for a little more detail, is there an inflection point at a particular price at which 200-millimeter memory capacity will come offline or 200-millimeter fabs become less viable, so to speak? Are we at the $2 price or $1.75 price, or is there some way to look to see out in the future where the inflection would be of the memory industry shifting at a faster pace to 300?
Michael R. Splinter
We haven’t looked at it from a pricing standpoint. We look at it more from a technology standpoint, Robert. These technologies keep shifting and 200-millimeter really isn’t cost competitive in the 55-nanometer range and beyond, so it just kind of runs out of gas and especially in the flash area where the technologies have been moving so rapidly to the leading edge, really driving technology now.
We think those older 200-millimeter fabs are going to run out of economic steam fairly quickly. Our prediction is over the next two generations of technology. |
