Field Notes from DAC 2008: EDA—A View From Sand Hill Road

One of the several excellent panels at the DAC Pavilion on the DAC 2008 exhibit floor was called “EDA: The View from Sand Hill Road,” referring to the collection of venture firms gathered at I-280 and San Hill Road in Palo Alto. The panel moderator was Lucio Lanza of Lanza techVentures, an early-stage VC and investment firm. Lanza is a legend in EDA and he’s quite a character as demonstrated by his opening remarks. Lanza said that EDA is not an industry. It’s a segment of a much larger industry called semiconductors (which is part of an even larger industry called electronics). He noted that the size of the EDA industry segment has been essentially flat for 15 years despite tremendous technical development efforts made at many EDA companies. “And this is really annoying,” he quipped. “It provokes thought. Why? Is it the business models? Is a wider view of EDA needed? What would EDA 2.0 be if we were to redefine it?”

Juan Antonia Carballo, a charismatic industry event gadabout and general partner with Argon Venture Partners, was asked to answer first because he spilled water all over Lanza during the pre-discussion preparation. “Venture investment is based on rewards for risks,” he said. The attractiveness of an investment hinges on how long it takes to get money back from the investment and the multiple of the monetary return from a company’s sale or through an IPO. If the time to money is too long or the return is too low, the investment community looks elsewhere. For EDA companies, there are few potential buyers for an EDA startup company; There are essentially three—Cadence, Mentor, and Synopsys—so the market for startup EDA companies is pretty limited. In addition, you need a tremendous amount of domain expertise to start and EDA company. For these reasons, said Carballo, investment money is now looking at other sorts of ventures.

Longtime industry observer and analyst Erach Desai heaped more issues onto the EDA pyre. “There’s poor correlation [between the EDA companies’ fortunes] with the semiconductor industry’s fortunes,” he said. Further, FPGAs have proven to be a viable alternative to ASIC design starts for many projects, limiting EDA’s growth. However, some EDA companies have evolved. Perhaps, suggested Desai, [Joe] Costello had the right idea—design services—but imperfect execution [or premature timing]. Companies such as eSilicon and OpenSilicon, which are value-added ASIC providers, may be expanding the market but are not counted in the EDA census.

In the end, many interesting things were said during this panel, but little was resolved. Sort of like real life.