|
||||||||||||||||
Once we got into our conversation in earnest, however, Madhavan was relaxed, engaging, and seemingly happy to explain how he got to where he is today - Chairman, CEO, and Co-founder of Magma Design Automation in Cupertino, CA. If he was at all sleep-deprived, it certainly wasn't apparent. Madhavan grew up in Southern India, where he went to college and earned a BS in electronics and communication from KREC (Karnataka Regional Engineering College) in Surathkal. That's not really the whole story, however. In fact, he says his parents were constantly concerned because he was far more interested in playing cricket than in studying. No matter the fortitude he exhibited in trying out for teams at various varsity and championship levels, nothing could dissuade his parents from worrying about his commitment to hard work and serious life goals. When Madhavan went on to graduate school at Queen's University in Ontario, Canada, his study habits showed few signs of improvement and he says he (and his parents) actually thought he might fail after just 2 semesters. Somehow however, he pulled it off to everyone's surprise and earned his MSEE. After graduation, he landed a job at BNR (Bell North Research, the research arm of Nortel in Ottawa), and it was there that he discovered his niche. Working at BNR, he found he needed a CAD software application to complete a chip design he was struggling with. There was no Synopsys at that time, no Design Compiler, so he wrote his own synthesis tool. Madhavan says, “Pretty soon, other guys in the company wanted my software for their designs and suddenly I became one of the EDA guys for the company. Which was lucky for me. There were 250 EDA guys there and many of them were the best I've ever worked with.” By 1991, he found himself working at Cadence Design Systems in San Jose as a BNR engineer involved in a long-term partnership between the two companies called the Analog Alliance. Madhavan says, “That's when I met Lucio Lanza. He was Vice President for Business Development at Cadence and a General Partner at USVP. Lucio gave me several start-up business plans to look over and evaluate. He would just hand them to me and say, 'Read this.' By showing me those business plans, he helped me to understand the venture capital business and how ideas are funded.” “It was particularly interesting to me to learn that you could actually earn a salary working at a start-up, that you didn't have to be self-supporting. I wasn't a rich kid and I had no idea that anyone could work for a start-up if they couldn't support themselves on family money.” Jim Solomon was also at Cadence at that time, leading the Analog Division. “Jim convinced me to join Cadence as a full-time employee in 1991, and I worked intensely on Cadence's Spectre HDL for a year and a half. But I found that it was practically impossible to get a new tool out [the door]. There was a political quagmire in the company that [constantly bogged down] efforts to sell the tool I was developing.” “Meanwhile, while I was there, Vinod Agarwal talked to me about licensing BNR BIST software and since I had worked on it earlier, it was natural for me to help found LogicVision. So Vinod Agarwal, Michael Howells, and I started LogicVision in the Bay Area in 1992, 1993. But at that time, I still had no idea of doing synthesis to compete against Synopsys.” “While I was at LogicVision, I had an opportunity to use Synopsys tools again - to integrate LogicVision BIST into Synopsys - and having worked on synthesis before and watched the failure of Cadence and Mentor in synthesis, I felt there was room for another synthesis player to compete directly against Synopsys. I had done the synthesis tool back at BNR, I looked at Design Compiler, and felt I could do better. So I left LogicVision and founded Ambit Design Systems in 1994.” It had taken quite some time to get funding for LogicVision, so Madhavan was prepared to be patient in raising the funding for Ambit, as well. But the process was far more difficult than even he had anticipated: “Ambit was a different company [from ones the venture capital companies had seen before]. We were rejected by 30 VCs on Sand Hill Road [in Menlo Park, CA]. Harvey Jones was very popular at that time, and both he and Aart de Geus seemed to have convinced the VCs that Ambit would not be successful.” So Madhavan abandoned the search for VC funding and turned to private investors instead. He and his Ambit team came up with 47 private investors. “John Sanguinetti was the first investor in Ambit. I had met John through Open Verilog International - actually before that, at Verilog conferences of Cadence. John and three Chronologic Board of Director/Management people were the first four investors. John had proved by doing a Verilog simulator against Cadence that competing against Goliath is possible, and that made me think I could take on Synopsys with Ambit.” “Then, more people looked carefully at us and decided to invest as well. Chromatic Research was our first benchmark and we beat Synopsys at the netlist level by 20%. But at the GDS II level, we were minus 3% compared to the Synopsys results. [Clearly], we had gains in the front-end, but the back-end kept drifting.” Nonetheless, Ambit turned out to be a successful business venture and Cadence bought the company in 1998. Even before that, however, Madhavan had concluded that there had to be some way to bring the logic and physical portions of the design process together - a way to “melt” logic and physical design from the front-end through to the back-end. So in 1997, he went looking for funding once again and founded Magma Design Automation. He says, “This time it was different. It was during the Internet Bubble and there were lots of people in Silicon Valley who had made a lot of money. We went into our meeting with the VCs wanting to secure $1 million [in funding], but came out 45 minutes later with $2.2 million. [Going forward], every round that we did turned out to be at the right time and right place.” Madhavan says it took 3 years to create the company's product, BlastFusion. He also says it simply wouldn't have happened in today's economy - money is too scarce and investors too nervous to wait out the technology development cycle. The Magma tools are based on what Madhavan calls the “unique idea of fixed timing - something that is needed to do complex chips. We had to re-write the RTL to GDS II [flow] to make it work. It's taken 1.2 million lines of code and is the most complex EDA project [ever undertaken].” Madhavan says Magma raised several rounds of venture funding, totaling about $115 million, and proudly went public in the fall of 2001. But he says, “More importantly - from day one, we've been the underdog in this industry. It's been David versus Goliath, where the competition all along has said that the 'fixed' thing wouldn't work. But out of our 90 R&D employees, 34 have PhD's - a lot of them are from IBM, Cadence, Synopsys, or Avanti - and they all [continue to pursue] fresh thinking to create our products.” “This team is passionate about our products, about making the products, and getting them out the door. In fact, every time Cadence or Synopsys says our system doesn't work, it makes it even easier to motivate our engineers. Our attrition has been very small because our people feel that what they do here is more than just a job.” Madhavan says there are 300 employees today at Magma and the company has been profitable for 5 quarters. He also proudly reports, “Dataquest gives us 31% market share. But at 0.13 micron, our market share is closer to 40%. Our ambition was to be in the top 4 companies in EDA. Courtesy of the Avanti acquisition [by Synopsys], we've achieved that goal. We've had $300 million in bookings and we're growing, while still keeping our start-up mentality. I know we have the maturity to be the leader. Our goal is to be the top company in IC design automation by the end of 2007. Period.” Asked about his infamous DAC 2000 panel encounter with Joe Costello in Los Angeles 3 years ago, Madhavan says Costello was “factually incorrect” when he made the accusation that Madhavan left Ambit after they were purchased by Cadence, that he simply starts companies and leaves with his winnings as soon as they're acquired. “I had come to the conclusion [at Ambit], that the back-end is where the profit is. Joe was wrong; it was painful for me when I left Ambit and I left a whole year before they were acquired by Cadence. I've been here at Magma for 6 years now and you've just heard my 5-year vision out to 2007. I'll be here to see that vision realized.” Meanwhile, Costello's skepticim is the least of Madhavan's concerns. He says that it's taken all these years to finally convince his mother that he's made a success of things. He says, chuckling, “She couldn't understand why I left BNR, which had 12,000 employees, to go to Cadence, which only had 1000. When I left Cadence to join LogicVision, which only had 4 employees, she lost it completely. She just couldn't get it. Recently, someone showed her an article about me in an Indian magazine, and she threw up her arms in joy. I think she's started to believe I might have [done something right].” “There's no entrepreneurial tradition in my family - it's almost an anti-business mentality. In India, when I was growing up, working for big companies was [the ideal goal]. But even in school, I was different. I commuted to school on the bus every day and I liked to read comic books, so I ran a comic book library for profit on the bus. When [it was discovered], I got suspended for a week. Nobody gave me credit. Instead they said, 'We must halt this criminality!'” Madhavan, however, is not afraid to give credit where credit is due. He says he's been lucky to have had three mentors in life - Ed Vopni, his manager at BNR, Jim Solomon, and Lucio Lanza. “I was just a lazy guy til I met Ed Vopni. He told me, 'If you go on at this rate, you're not going to amount to much.' I started working harder after that.” “I found Jim Solomon was really inspirational in speaking about his early experiences - his success inspired me. When I asked him what mistakes he had made, he told me he wished he had done the Analog Division [at Cadence] as a startup, which helped me with the idea of becoming an entrepreneur. Jim did not specifically urge me to do start-ups. In fact, when I decided to leave Cadence, he tried to tell me that start-ups are difficult. But his earlier comments had [already] left their mark in me.” “And Lucio Lanza has also been a mentor, in an unplanned way. He helped me by letting me evaluate those early EDA business plans. He told me that if I thought something was a good idea, he'd go with it - which gave me confidence.” How does Madhavan feel about Joe Costello and John Cooley? He says, “To be fair to Joe, he always kept people [at Cadence] focused on an entrepreneurial culture. But it's the intensity in small EDA companies [that fosters innovation].” “Of course, John Cooley [took us on] in 2001. He found some flaws [in our tools] and, although at the time it felt painful, it was actually good for us in hindsight. I think what he said about us was 5% fact and 95% not true. But Cooley made our team fight even harder. We're a company with a highly competitive spirit and we hate any bug that makes us lose. So by accident, John Cooley did us a favor. In fact, if he had ignored us, it wouldn't have been [as useful].” Meanwhile, with Roy Jewell in as President and COO at Magma since 2001, Madhavan's day-to-day responsibilities are no longer at the operational level. Madhavan says today, his job is to set the vision for the company: “And I'm not micromanaging, but I let the management know what's happening. I like to talk to everyone, to walk around and learn what the issues are. I'm accessible to all employees for that [kind of discussion]. Even with 300 employees here, I know just about everyone's name. I don't try to fix problems, for instance in Sales or Engineering, but I point out issues and ask how they can be solved. I don't want to be dictating [policy].” “And I'm happy to hear criticism. Life is all about the steps of correction that you take. Everybody here should have the freedom to talk to me - if I were to stifle that feedback, it would be a mistake. I'm working 18 hours a day here. It's part of the passion of doing a start-up and the passion for the technology.” With all of this, how does Madhavan relax? He says the answer is simple: “Roses.” “I don't exercise. In fact, I don't have time to do anything for relaxation outside of work, but I do have 500 roses. So when I'm in town, I tend to my garden. When the lot next to my house came up for sale, I bought it and started by planting 4 rose bushes there. It ended up in a gardening spree. It's just something that happened and now it's a pretty big garden. In the beginning, I had no idea how to raise roses, how to do any of this - but as you learn, you become more of an expert.” Madhavan is clearly delighted to report that he has expert gardening assistance in the form of his two young children. He says they love to help him tend the roses each Saturday morning, and that helps to ease his occasional guilt when he is away from the family so much on business. His wife also works full-time during the week, but ideally she would also be involved with the family gardening on Saturday mornings, as well. Madhavan says, “I admire my wife a great deal. She has taken great care in raising the children, while also handling her job. It's not easy when I travel so much.” He adds with a chuckle, “It turns out, however, that she has terrible allergies to roses. She can't even come out to see the garden. Every time I buy more rose bushes, she says, 'Oh no! Not more roses!'” All told, Madhavan says living and working in Silicon Valley is as good as it gets - both personally and professionally. “The beauty of this place is the cultural melding and the entrepreneurial spirit. But for this place, I would never have [done the things] I've done.” And of course, he adds, the weather is just perfect for someone with 500 rose bushes - and counting. Peggy Aycinena is Contributing Editor at EDA Weekly and can be reached at peggy@ibsystems.com. |
|
|||||||||||||||
|
||||||||||||||||
|
||||||||||||||||
