"I'm not sure our current business models are the best business models to go after that [DFM market]," said Ted Vucurevich, CTO of Cadence Design Systems Inc., at an "Ask the CTOs" panel discussion here as part of the Design Automation Conference Monday (July 24). Vucurevich suggested that the winning business model would be more "risk/reward," rather than simply supplying a tool that vendors cannot guarantee will increase yield by a quantifiable percentage.

Vucurevich, as well as Synopsys Inc. CTO Raul Camposano and Applied Materials Inc. CTO Mark Pinto were unable to produce a concrete number for the size of the total DFM market, defined by moderator Kurt Keutzer of the University of California-Berkeley as the amount of money newly available to EDA through DFM solutions. But Pinto noted that the amount of money that chip makers spent on yield, including test structures, resolution enhancement technologies and other areas, was roughly $2 billion.

"A lot of people would love to go after that billion dollar TAM, but aren't sure they have the right model," Pinto said. He added that the most appropriate DFM model would be "some type of hybrid, royalty-based model."

Pinto added, "If you can get yield up even 5 percent, that's a huge amount."

The dilemma is not a new one for EDA vendors, which have seen the average selling prices of tools stagnate and even decline in recent years, partly because of the difficulty in quantifying tools' productivity improvement in design.