The Sandbox

What's happening behind the scenes in the semiconductor manufacturing industry? Read this blog by Senior Editor Ann Steffora Mutschler to find out - and chime in with your thoughts and questions.

Profile

About Me

RSS Feed

Add this blog to your RSS newsreader!

By Category

Semiconductor Manufacturing Articles

Blog

The Sandbox

« Non-US solar market may hold m... | Main | Canadian province investing in... »

Tuesday, July 15, 2008

Industry says ‘no’ on 450-mm wafer size, but no pushback on Intel

I’m starting to wonder what the real story is behind all the talk on the 450-mm wafer size. I’m here at Semicon West in San Francisco this week, where ‘450’ is the buzz everywhere you turn, with most equipment vendors saying they won’t go down that path.

SEMI recently released a report on the economics of a 450-mm wafer transition and concluded that for now, “Shrinks, new materials, and new processes will continue to advance the industry on Moore’s Law, but there are simply not enough R&D resources available to continue such advancement in nodes and processes AND to work on a 450-mm wafer size transition.”

This report is corroborated by a recent survey of semiconductor industry insiders by Wright Williams & Kelly Inc that found some survey respondents did not expect to see 450-mm wafers in production until 2014 or beyond, with more than half of the respondents indicating that 450-mm wafers would never happen in production manufacturing, marking an increase in negative sentiment over the 2007 results.

After the enormous cost to make the 300-mm wafer transition, semiconductor manufacturing equipment vendors are rightfully cautious, however that didn’t stop Intel, Samsung and TSMC from saying in May that they want a 450-mm wafer pilot line in 2012, justified by one Intel exec who said last month that equipment makers are overestimating the cost of moving to 450-mm wafers.

This seems to fly in the face of what Stanley Myers, president and CEO of SEMI said yesterday during a panel discussion when asked if there is push-back on Intel, given that most industry players do not wish to make the move to 450, “Intel is THE customer, so they are part of the analysis,” referring to SEMI’s report which leads me to think that some parties are holding back what they truly think, or are too shy to step up and say no to Intel, which frankly has called most of the shots in the equipment industry. Given the revenue-generated power and capex spending of Intel, Samsung and TSMC combined, I’m starting to think the equipment makers may have no choice but to begin working on the move to 450 whether they like it or not.

Chime in with your thoughts.

-Ann Steffora Mutschler, Senior Editor

Reader Comments

(Add your comments)

at 7/15/2008 8:00:41 PM, Samuel said:
In this industry, you could not say no to the big company, because we still got a lot of competitors. When the 300mm start up we provide the first tool to the market of 300mm and now we are the leading vendor in this area. We believe the story will happen again, so although it is difficult we still need too fulfill big customer''s request.

at 7/16/2008 1:48:42 AM, JimJ said:
I wonder if there has ever been an equipment industry wide balance sheet produced showing the cost of 300 mm deveelopment and the recovery. I believe that it would still be very much in the red! Now the pressure to spend at least double the amount to sell half the number of tools to an industry that never wants to pay a real price, always justifying bargain seeking against "future" business! In any decision process one key element that keeps getting put in as a low priority is the metrology needs, as nodes come and go the challenges only increase even for the current 300 mm piece of floppy silicon, let alone the saggy pancake that would be 450 mm! So while it may be possible to grow the Si, even process it, deciding what the result is will become a first class nightmare. Now as metrology (and I include all aspects of yield management) is ,by definition, that which you don't need if your process is perfect and that perfection is what every boss insists that they will have, then metrology becomes a dirty word and not worth any part of the decision/evaluation process in a wafer size change. Wake up you guys, tell Intel that if they want it, then they should pay for it (up front)

at 7/16/2008 10:14:50 AM, JKYWANG said:
When scaling is slowing down, increasing wafer size is the surest to drive the cost per die down. The winners will be the one who can migrate to 450mm quickly and figure out how to keep scaling it down to 16nm. Unfortunately, it appears there are only three winners and a bunch of others now.

at 7/16/2008 12:41:30 PM, Mel said:
Neither TSMC or Intel or Samsung are buying much now or have really given any confidence that they will buy at the same rate as before. Converting to 450mm will decrease the number of fabs with pressure not to increase cost of equipment too much. It is expected that at the current rate of usage for 300mm, TSMC will need only 1 450mm fab. What's in it for an equipment manufacturer. Intel, Samsung, TSMC need to share in the development cost

at 7/16/2008 6:51:01 PM, Hoochie said:
The difference is that the equipment companies have the upper hand compared to when 300mm was introduced. If AMAT, TEL, LAM, Nikon, or even Varian say no, then it is pretty much a veto for the whole industry.

at 7/21/2008 1:12:31 PM, Keith said:
I simply cannot see the potential that we can ROI the investment in 450 mm in any reasonable amount of time. The 450 customer base will be too small, the 450 ramp will be too slow an the investment will be too high. We would be better off sinking the cash into government bonds and just laying off our development staff than investing in 450 mm! The ROI would be much better. With no ROI incentive, we simply won't undertake the project. If that's what it takes to stay in the semiconductor equipment business, then maybe we'll just have to get out of the business, because it simply doesn't make any sense to make investments that won't offer a reasonable return potential. We do have a choice. We can simply choose to not do something stupid.

The Sandbox

Profile

RSS Feed

Recent Posts

Recent Comments

Most Commented On

Archives