Santa Cruz, Calif. — No one can say Ivan Pesic doesn't think big. The outspoken founder and CEO of Silvaco International Inc. expects to be running a new company that will become the No. 1 full-line electronic design automation provider within a few years — with no venture capital money and few if any acquisitions.

Silvaco last week announced the spin-off of Simucad Design Automation Inc., which will sell the analog/mixed-signal IC design tools developed by Silvaco. Simucad's stated intent is to go for an initial public offering in the fourth quarter, to add digital synthesis and placement and routing in 2007, and to become the No. 1 provider of simulation and IC physical-design tools.

"In four or five years there will be only two [full-line] CAD guys left," said Pesic, who is also the CEO of Simucad. "I tell you one thing. It's going to be Simucad and whoever is the second guy."

Bold claims indeed, but Pesic's past suggests a remarkable resolve. In 1984 Pesic started Silvaco on his own, and financed it by running an auto body shop on the side for five years. Over the years, Silvaco became a significant provider of technology-CAD (TCAD) and Spice simulation tools. Pesic owns it, and Simucad, lock, stock and barrel — no debts, no bank loans, no venture capitalists.

A fierce competitor in the courtroom as well as the marketplace, Pesic won legal battles with Technology Modeling Associates (TMA) and MetaSoftware, ultimately collecting $20 million from Avanti Corp. following its purchase of MetaSoftware. More recently, he won a lawsuit against Circuit Semantics Inc., which, according to a court judgment, used Silvaco trade secrets in its DynaSpice product.

Today, Pesic is suing some of Silicon Valley's largest companies — including Intel, Agilent, Cypress, Cirrus, AMD, Specular Networks, NetLogic Microsystems, AMCC and Tvia — alleging they continue to use DynaSpice, which he calls a stolen version of Silvaco's SmartSpice.

What if these companies decline to purchase Simucad products? "I really don't care," Pesic said. "There are 900 companies in this Valley. People buy my software because it's the best." Pesic, in fact, looks forward to bringing one or more of the defendant companies to court, "so we can send a very clear message that stealing software is not cool."

Simucad is starting life with a revenue run rate of around $35 million a year. While financial analysts say that's on the low end for an IPO, Pesic says that's no problem; he expects the run rate to be up to $60 million by September. Unless he reaches that target, he said, he won't do the IPO.

That's a big leap in a single calendar year, but it will be "very easy," Pesic said. "In the last two years we worked night and day, and now we have all the new products in place," he said.

While Silvaco will keep the TCAD tools, Simucad will get everything else, including the SmartSpice, Harmony and SmartSpice RF simulation tools; Silos Verilog simulator; AccuCell and AccuCore library characterization tools; Expert layout editor; and Hipex parasitic extractor.

SmartSpice and Silvaco received several accolades from users in an E-mail Synopsys Users Group (ESNUG) mailing last year. Several users lauded SmartSpice's performance, accuracy and considerably lower price than competing products from Synopsys Inc. and Cadence Design Systems Inc. "Ivan [Pesic] sells good tools cheap, and he provides good support too," concluded ESNUG moderator John Cooley. "Maybe not so crazy, this Ivan."

A native of Yugoslavia, Pesic came to Silicon Valley 30 years ago, where he earned MS and PhD degrees from Santa Clara University. When he was a CAD manager at Monolithic Memories, he decided to start an EDA company and figured he could do better than Daisy Systems, Valid Logic or Mentor Graphics. But he had also taken MBA classes at Stanford University, and was so shocked by the "arrogance" of venture capitalists that he decided to avoid them completely.

Pesic launched Silvaco in 1984 with Utmost, a still-existing tool that provides parameter extraction, device characterization and modeling. To raise money, he ran an automobile body shop, and found that fixing 50 to 60 cars a week generated enough to run a five-person engineering company. "That was hard," he said. "I was eight hours a day in the body shop and 10 hours in Silvaco, every day, including Saturdays, because that's when we delivered the cars." In 1988, Pesic sold the body shop. A year later Silvaco entered the TCAD business because it was a new market, Pesic said. But there was a TCAD provider, Technology Modeling Associates. In 1991 TMA sued Silvaco, accusing it of going through TMA's Dumpster to find trade secrets.

Silvaco also tried to enter the Spice simulation market during the same period, only to be sued by MetaSoftware. Silvaco successfully fended off both the TMA and MetaSoftware lawsuits, but not before withdrawing from the Spice marketplace for several years. Silvaco then came out with its best-selling product, SmartSpice, and subsequently added analog schematic capture, layout and physical verification.

Pesic said he has intended to go public with Silvaco all along, and was ready to do so in 1995, but found that the company's accounting was not set up for an IPO. It would have taken months of work and millions of dollars to file, he said. So the idea emerged to start a new company that was set up from the start with the right accounting practices.

Enter Simucad Inc., one of the oldest names in EDA. By 2003 Simucad had become a small vendor with a relatively obscure Verilog simulator. Silvaco bought Simucad, partly because it needed a digital simulation capability and partly because, as Pesic said, "I liked the name."

With the spin-off, Simucad has 220 of Silvaco's 250 employees and all products except the TCAD line. Pesic said he'll find another CEO for Silvaco and focus his attentions on Simucad, where he'll retain majority ownership. He is hoping to raise $150 million to $200 million in an IPO and use this to fuel R&D. In early 2007, Pesic expects Simucad will offer digital synthesis, placement and routing tools.

"When I enter synthesis, place and route, I will bring the prices down," Pesic said. The big EDA vendors, he said, are "vulnerable" because "pricing is very high and the product line is not really that good. The product lines grew by M&A [mergers and acquisitions] and that is not the solution for new tools." Pesic said Simucad may "acquire carefully" but will focus heavily on homegrown R&D.

Enough for an IPO?
Simucad will compete head-on against Cadence, Synopsys, Mentor Graphics and Magma Design Automation, all of which declined to comment on Simucad or Silvaco. But Aki Fujimura, CTO for new-business incubation at Cadence Design Systems Inc., noted that it's a lot harder to pull off an EDA IPO these days because of the demands of Sarbanes-Oxley regulations and because "customers are increasingly demanding total solutions."

Financial analysts say it probably won't be difficult to find an investment banker to run an IPO, but there's still skepticism about Simucad's size. "I tend to believe that companies need to have around $50 million in sales to really have critical mass with the Street," said Matt Petkun, research analyst at D.A. Davidson. Sarbanes-Oxley compliance alone, he

noted, can cost $1 million a year. Petkun doubted that in-house R&D can fuel a major EDA company. "To get to critical mass, you need mergers and acquisitions," he said. "One reason you do an IPO is to get capital to go out and make acquisitions."

Erach Desai, principal at Desaisive Technology Research, said that the EDA industry's flat revenue growth doesn't bode well for IPOs. "If in the macro environment the market is not growing, it's very hard to get investors excited," he said. Becoming a broad-line supplier is very hard, he said, and requires acquisitions.

Raj Seth, director of institutional research at S.G. Cowan, said there may be another cycle of EDA IPOs. He added, however, that "the bar relative to the last cycle has clearly gone up in terms of the requirements in scale and profitability, and the cost to be a public company has gone up."

Gary Smith, chief EDA analyst at Gartner Dataquest, believes Simucad may have a good analog play, but said that "if they attempt to get into the digital flow, they don't have much chance." He also said it's early for an IPO, because Simucad needs a track record as an independent company.

What about building a major design automation company with no venture capital, and mostly or entirely with in-house R&D? "No one has [done so] yet, which is why this is a historic announcement," Smith said. "We'll have to see how history unfolds."