R&D Gap Presents Dilemma for Industry

Aaron Hand, Managing Editor -- Semiconductor International, 2/15/2006

The semiconductor industry is at a crossroads. Market drivers have shifted to consumer products, the industry growth rate has slowed, and Moore’s Law is expected to stay intact for another three or more generations. It’s this crossroads that is really challenging the industry, with the escalating cost of the R&D required to maintain the progress creating an annual gap estimated to reach $9.3B by 2010. Ron Leckie, president of Infrastructure Advisors, presented the situation this week at ISS Europe in Amsterdam, in a talk titled, “Funding the Future.”

The industry is producing about 12 billion chips per month – that’s about two chips per month for every person on Earth. This is creating average selling price (ASP) challenges, Leckie said. The volatility for semiconductor manufacturers remains, and gets even wilder for the equipment industry, although the materials market proves less volatile.

Device makers are now spending 15-17% on R&D, according to Leckie. Combine that with increased revenues, and R&D spending has grown quite a bit, he added. But only the device industry has managed to substantially increase R&D spending, he said, with the equipment and materials industry holding R&D money flat over the past five years (although percentages have increased).

On the equipment side, it’s getting more expensive to develop the tools. And with so many new materials being developed, it takes a long time for them to be integrated into production, therefore driving up the cost of R&D spending for the materials suppliers, Leckie explained. “Look at all the money that was spent on 157 nm lithography before we pulled the plug,” Leckie said. “We go down paths because they’re a neat technology. But we should go down the ROI path first.”

So ultimately, there’s an expected $9B R&D gap by 2010 between supply and demand. Of course, you won’t ever see it, Leckie explained, because the industry just won’t spend it if it doesn’t have the money to spend, “but things will have to be sacrificed.”

There are a few – albeit not easy – alternatives for closing the R&D gap, according to Leckie. One is to strengthen business models. There’s a battle on pricing, which only serves to create an inelastic market. “If you make cheaper tools, customers aren’t going to buy more tools,” Leckie said.

Increased partnering and collaboration is another alternative. The general feeling about consortia is that they’re mainly clubs for device makers, Leckie said, but that’s changing. IMEC, for example, deals well with equipment and materials suppliers as well.

External funding would also help – from customers, for example. Leckie gave as an example Intel’s funding of Cymer for EUV sources. “In the 1990s, we didn’t see much of that,” he said. “But where there’s a situation where the customer’s going to benefit, there should be an arrangement where the customer helps fund the research.” Other funds and subsidies could come from government, particularly in fundamental research and integration labs.

Finally, slowing the pace of development may help. In particular, the industry needs to avoid blind alleys and stimulate more creativity.