Semiconductor industry must collaborate or die

Research and production costs are rising at a time when money is in short supply

The semiconductor industry will have to collaborate in the future to keep a rein on costs, according to leading lights in the industry.

Experts at Semicon West 2009 were in broad agreement that the industry must change its business model in the face of global recession and ever-advancing research and development (R&D) costs.

"The global economic crisis is fundamentally reshaping the semiconductor industry," said George Scalise, president of the Semiconductor Industry Association. "We are approaching a tipping point where previous strategies will not work. We must work together on R&D."

Scalise explained that new and better technology is driving the costs of research and production up at a time when money is increasingly tight.

Mike Campbell, senior vice president of engineering at Qualcomm, suggested that the industry should follow the Bell Labs model, and have institutes for joint research.

"You can't have 300 tiny differences in semiconductor processes between manufacturers. If suppliers are going to survive you have to have a basis for a common point of technology," he said.

However, Mark Durcan, president of Micron Technology, warned that collaboration could only be taken so far.

"Collaborative research has a significant place, but only in pre-competitive work," he said. "We're all under the same pressures and that is a part of the solution. But if we are going to live in a world with fewer memory developers, we have to get value for R&D by getting competitive advantage from it."

Durcan pointed out that capital expenditure on Nand and DRAM had fallen 60 per cent in the past year.

Governments were cited by some as a possible saviour of the industry in that they could pump money into research, but some were sceptical.

"Asking government to do this work is like getting my teenager to drive himself," said Tien Wu, chief operating officer at Advanced Semiconductor Engineering. "We need government to put money into research, but government may not be the best candidate to do the research."