Chip Profits Crisis
By Richard Ball, Electronics Weekly -- 10/16/2002
Electronic News
NICE, FRANCE -- The semiconductor business is facing a crisis with chip companies finding it harder to make profits as costs increase, according to delegates at last week’s SAME conference.
The shift to smaller processes, such as 0.13-micron and 90nm, is cutting margins to the point where no real profit can be made from chip making.
"What we’re facing is not an engineering problem, it’s an economics problem," said Jacques Benkoski, president and CEO of EDA tool supplier Monterey Design Systems Inc. "The profitability of the semiconductor industry is at stake here."
Very few designs reach the volumes needed to make them profitable, while programmable logic is still too expensive, he said.
"The cost incurred by not being right first time is tremendous," said Guillaume d'Eyssautier, European VP at Cadence Design Systems Inc.
Robert Payne, CTO at Philips Semiconductors, added, "The chip industry won't recover when the economy recovers unless we can stimulate demand."
But rather than being an economic problem, Payne said a collapse in the chip industry would be the fault of designers failing to innovate.
"There’s a crisis of confidence in the industry," said Jim Tully, chief analyst at Gartner Dataquest. Tully said the number of ASIC design starts peaked at 11,000 a year in 1997, but by last year had dropped to about 5,000.
This trend is caused mainly by the success of integration and SOC design, but leaves fabs running at lower capacity, especially if they move to 300mm wafers.
The big winners have been the IP sector and programmable logic. IP grew about 25 percent last year, according to Tully, while FPGAs have been mopping up the low end of the ASIC market.
Electronics Weekly is an e-inSITE affiliate.
e-insite.net
[Harry: Interesting perspective. It puts a very different light on the migration to more dense technologies. The fact that the start up cost may not be justified by many designs throws a different slant on the issue. It becomes irrelevant once the more dense technologies become main stream, but in the mean time the migration to newer technologies may be further off than many thought.] |
