Sequoia Communications CEO Dave Shepard was out of town last week when the demise of the San Diego fabless semiconductor design company came to light. I was told he was furious over the shutdown, as Sequoia had finished its product, a chip for use in cellphones and wireless devices, had secured customers, and was roughly a year from breaking even. The explanation was that the eight-year-old startup had been unable to raise the additional venture capital.
But if he was upset, the semiconductor CEO had processed the news, so to speak, by the time I finally reached him yesterday. Instead of talking from Sequoia’s ground zero, Shepard spoke from a considerably higher altitude, saying, in effect, that the problems that swamped the San Diego startup chipmaker extend well beyond the company itself.
“The issue is bigger than Sequoia Communications,” Shepard says. “I think the venture-backed model for semiconductor startups is broken. The complexity of these chips has just gotten so high, it just takes so much money to fund a startup nowadays, that to the VCs, it’s just not worth it.”
In a nutshell, Shepard says semiconductor startups are getting squeezed by sharply higher costs and dramatically lower valuations. A new chip technology that would have required two to three years and $15 million to $30 million in startup funding to get to proof of concept a few years ago now requires five to eight years and something closer to $80 million. For that kind of startup capital, Shepard says VCs want to see semiconductor buyouts of $400 million to $500 million. But deal values have plunged. At a meeting organized earlier this month by CommNexus, San Diego’s non-profit wireless industry association, one presenter said the median value of 97 semiconductor M&A transactions in 2000 was $484.1 million. In 2003, there were only 47 deals and the median deal size was just $144.5 million. The trend has worsened with the economic downturn, and Shepard says valuations for semiconductor startups are now in the range of $75 million to $100 million.
Shepard says he was unable to sell Sequoia, even though the company had contracted with several customers in China and had a finished product—a sophisticated multi-mode transceiver designed to accommodate the burgeoning market for various 3G mobile phones. In addition to transmitting and receiving standard cell signals, Shepard says Sequoia’s transceivers perform multiple functions, such as translating analog RF signals into the digital code used by mobile devices.
After raising close to $75 million from VCs and other investors, Shepard estimates Sequoia was about a year away from breaking even. The company, which had about 30 employees, needed to raise an additional $10 million. Sequoia’s investors instead decided to cease operations in mid-July. (By last week, auctioneers were preparing to sell the company’s remaining assets and Shepard was taking a break in Lake Tahoe.)
“The VCs just got tired,” Shepard says. “Half of our VCs were out of money, and the other half had been in the company for a long time. So they were looking at a low return on their investment.”
He adds, “It’s not really anybody’s fault. I’m not mad at the VCs.”
He maintains that what happened to the San Diego startup “is not a Sequoia-specific problem.” The changing economics of semiconductor innovation have ramifications for the entire industry, he says, because the big chipmaking companies like Qualcomm, Intel, and AMD have traditionally counted on smaller startups as a key source of their innovation.
Worldwide, Goldman Sachs data shows semiconductor M&A transactions declining from 137 deals with a total value of $33.7 billion in 1999 to 86 deals with a total value of $2.2 billion so far in 2009. Semiconductor IPOs have been practically non-existent this year, with Goldman Sachs showing 10 deals worldwide with a total value of $43 million. The CommNexus meeting, which Shepard helped organize, was billed as “Endangered Species Alert: Fabless Semiconductor Startups Threatened With Extinction!”
Shepard says he not sure how the broader problem is going to get solved. It’s possible that China could move into the semiconductor R&D vacuum, although that could raise broader issues for U.S. economic policy. “The semiconductor industry is a sensitive area that touches everything,” he says.
Meanwhile, an online auction of Sequoia’s office furniture and electronic testing equipment is set to begin today.
The company’s most-valued assets, the design database for the chip and its intellectual property, including Sequoia’s portfolio of 20 issued U.S. patents and 15 patent applications, will be sold through a separate process. Proceeds will go to the investors. “I have no idea what the value is, but it will be under $10 million,” Shepard says. “It will be vastly less than what they put in.”
