Seven days after the Presidential Election failed to decisively settle on a new Chief Executive, I had a chance to enjoy an extended chat with Aart de Geus, Chairman and CEO of Synopsys, Inc. We sat in his office in Mountain View and discussed presidential politics, the Electoral College, and his long career in EDA. Now, two months later, a new American President has been inaugurated and I finally finish this article about Aart. Projects like this don't normally take two months, but on that fateful day last November moments after I left Synopsys I managed to drive my car under a truck headed East on 237 in heavy, late-afternoon traffic. For some reason, the drama and aftermath of the accident became an impediment to writing this and now that the article's done, it must mean that the accident is finally over.
Aart de Geus is an interesting man. Many people have heard him speak, know his story and company well, and some have even had the good fortune to hear him play the guitar. I, however, had never had a long visit with Aart, so it was an hour (and a half) well-spent talking with him.
The last two months have offered living proof that the American political system can survive a convoluted and unprecedented post-election process. Aart de Geus offers a different kind of living proof the proof that a man well-steeped in technology can successfully found, fund, and follow-into-hugeness a high-tech start-up despite folk wisdom to the contrary that start-ups require one kind of leader for the $0 to $50 million annual revenue range, a different kind of leader for $50 million to $500 million, and yet a third kind of leader for the $500 million range and beyond. Aart has, authentically, done it all. He's grown Synopsys all the way from its infancy up to its current size today of over $600 million in annual revenue.
In fact, the infancy-to-adulthood analogy is exactly how de Geus describes the history of his company, which prompted a question from me. I asked Aart if the technology folks that hail from Northern Europe seem to be particularly gifted in the art of humor and whimsy. I reminded him that his DAC'99 keynote, "System on a Ship," comparing the rigors of Columbus' voyage with today's adventures into "System on a Chip" was matched in whimsy by Hugo de Mann's keynote at DAC'00 de Mann being from Belgium, de Geus from The Netherlands. Aart declined the compliment and said he didn't know much about the art of humor, but he relies heavily on the art of analogy to get a point across. In the case of a company's development, Aart believes that just as in human life there is a beginning, a childhood, an adolescence, and an adulthood.
Europe and The West
Aart's own life began in The Netherlands. At the age of 4, his family moved to Lake Geneva in Switzerland. After 10 years living and attending school in the French-speaking region of Switzerland, the family moved to Switzerland's Basal-en-Main where Aart spent his adolescence studying in German. Aart's distinctly multi-lingual particularly impressive when you consider that after university in Europe, he came to the United States to continue his graduate work in Electrical Engineering at Southern Methodist University in Texas and studied in yet a third language, English.
Aart says Europeans that come to North America are on a quest: "Everybody from Europe who comes to the U.S. or Canada is looking to discover something." In his case, he discovered a "stroke of luck" in being assigned as a graduate student to Ron Rohrer at SMU. "He took a liking to me," Aart says, which was fortuitous considering that not only was Rohrer his graduate advisor, but also dean, chairman of the department, and provost of the university. Aart gives Rohrer attribution as the father of Spice in that he developed the Cancer program for circuit analysis for networks. (projects.openresources.com/libresoft-notes/libresoft-notes-en/node88.html )
"Rohrer essentially gave me the freedom to do whatever I wanted" within the graduate school research facilities, according to Aart. At that time, Texas Instruments had just donated 10 microprocessors to SMU, each with 16 kbytes of memory. Aart managed to put together a course in EE that had 10 graduate students developing "fun" teaching programs on the microprocessors. The programs displayed logic on the screen and allowed the students to discover what the logic would do. The project developed into one of the core classes in EE at SMU.
Aart took several lessons away from this experience. He developed an ability to manage a set of students, skills that he later capitalized on to develop his management style. To describe that style, Aart uses a farming analogy: Everybody counts on the [farming] team and there's always a role for everybody, which produces an ecosystem that manages itself. A self-sufficient farm suggests a self-sufficient company.
Meanwhile Rohrer was back at his home in Virginia. He suggested that Aart come and stay with him and his family a time-honored tradition, per Aart, that hails back to the time of Socrates when a student spent time living with his teacher. Aart made the trip to Virginia and ended up working nearby at General Electric in North Carolina developing a group committed to tool development an effort that ended up relegating Aart's graduate studies to "the night hours."
At GE, Aart worked on tools that designed logic with multiplexors that included much more efficient logic. He decided to develop gate arrays with mixed logic and multiplexors, an effort that coordinated with GE's move into the gate-array business. "Nobody in his right mind wanted to design with multiplexors," Aart says, so it was decided to write a program to solve the problem. "It was only much later that we started to look at papers [in the area] being presented at the time." Aart and his group eventually realized that there was already 20 years of research that had been done on synthesis "all of which I knew nothing about," he says.
Aart began to assemble a team at GE, starting with one open requisition for an entry-level manager. In the first summer he was permitted to hire one student. In the second summer, he was allotted three students. By the third summer at GE, his group had grown to 23 students. In that third summer, GE discovered that all of these students needed cubicles, and so forth, even though they were effectively earning minimum wage.
Aart explained that in 1985, "GE decided to get out of the semiconductor business. It had never quite worked for them. So, I started interviewing in Silicon Valley. But, the team we [already] had was as qualified in tools as we could find here." So Aart asked his team, "Why don't we take the synthesis [problem] and see what we can do with it."
How do you do that? First, you write a business plan. Then, you ask, is it ethical? Is it okay to do that? How do you go about planning the business so as not to go up against GE? Well, according to Aart, there's a simple answer. You write a business plan and propose it to GE. After all, it was clearly their IP. Period.
So, "a couple of us put together a business plan to present to the board of the general manager." The GM liked business plan and Aart & Company were successful. Based on that meeting, GE decided they would put a significant equity stake into the enterprise. "If we did well, they would benefit. I had a chance to present to Ed Hood, who was second in command at GE at the time. I had one hour to present the plan. Support for the business was predicated on raising another $5 million on our own." Hood was obviously swayed in their favor. The group received $1 million from GE and returned to Silicon Valley to raise the rest, convinced that raising the rest of the capital was "clearly going to be a piece of cake."
They first received $1 million from Harris Semiconductor which represented a magic conjunction between people who knew the technology and those who knew how to work with GE. Then, they got a $250,000 loan form GE, above and beyond the initial investment. Aart says that, in hindsight, he "shudders to think that I didn't realize that we would have to pay it back!"
The Map as Dart Board
Soon thereafter, the business was up and running. They knew they were underway when they managed to sell one copy of their CAD tool to Harris Semiconductor. Aart gives credit where credit is due: "The fact is that Harris was incredibly supportive." By September of 1987, Aart & Company had indeed finished raising the requisite $5 million, which he says wryly was "a great deal of money in those days."
Now they had to decide whether to house the company in North Carolina's Research Triangle Park, on Route 128 outside of Boston, or in Silicon Valley in California. Most of his team was fresh out of graduate school, so everything that they owned could fit into the trunk of a car. Subsequently, he only lost one member of the team when he decided to settle the company in California. "I alone had a mortgage or a loan," Aart adds, which explains why the move was probably hardest for him.
Here in Silicon Valley, Aart and his team had access to a wealth of intellect from U.C. Berkeley and Stanford, an array of relevant legal talent, venture capital input, and entrepreneur camaraderie. They took occupancy of their Silicon Valley offices in October of 1987 and promptly started trolling for additional staff with which to build the company. Harvey Jones was recruited as president, having previously served as CEO of Daisy. Harvey brought much needed marketing and sales expertise to the young enterprise. Aart & Company made their first appearance at DAC (the Design Automation Conference) in June of 1988, where they formally introduced their product for the first time.
I asked Aart if he felt lucky about this string of events that lead up to the start of Synopsys. He responded, "Luck favors the well prepared," and added that a fortuitous combination of management, graduate students, geographic location, viable business plan, and marketing expertise were augmented by having the "right technology at the right time. It was an era where schematic entry was beginning to taper off, and the age of RTL simulation beginning. We teamed up with Gateway and their simulator. All of these things worked to our advantage."
Aart de Geus is a meticulous technologist, so this fortuitous combination like luck may have been nothing more than the apparent reward for the well prepared. It's hard to dispute his telling of the history when you look at the fact that, by all reports, Synopsys has owned the lion's share of the synthesis market for a decade now. In many ways, you could argue their product has been synonymous with the term synthesis in the design flow.
But, Aart says that a company like Synopsys can suffer from being so closely identified with one part of the design flow. Important as synthesis may be, there's always the concern that "You're just a one-trick pony." I asked him what that means. He said, you ask the pony if he can do a new trick. The pony replies, "Nope. I want to do the old trick." So you have to convince the pony to expand its horizons, to expand its growth. He feels Synopsys has managed to do that; now [only] one-third of the company is involved with the synthesis tools, while the remainder of the company has been given more bandwidth.
Returning to the tutorial value of analogy, Aart explains that in the years since its inception, Synopsys has progressed through the stages of baby, toddler, adolescent earning it's driver's license replete with the "nervous mom" standing by and on into adulthood. In fact, Aart says that the company has had two main phases in its growth, during which it was effectively two different companies the pre-IPO Synopsys and the post-IPO Synopsys.
He reports with pride that, today, Synopsys has over 3000 employees across multiple sites, and includes among their two largest acquisitions, the purchases of EPIC and Viewlogic. He qualifies the statement: 40% of Viewlogic came to Synopsys, while the Board Systems Division ended up being at Innoveda. In essence, according to Aart, Synopsys has had a very steady, albeit fortuitous, growth over its history.
Today, there are multiple business units in the company some of which have endured merger and acquisition dynamics and all of which have faced the challenge of meeting real-world customer problems. Recently, an additional challenge has emerged the transition to the new licensing model that is sweeping through EDA.
"We're not lacking in ideas in what we can do. As this became clearer and clearer, it became more obvious that the business model was flawed. Selling something once [to the customer] and then [selling them] trickle-purchase updates [was not working]," according to Aart. Instead, he says, "The 'rental model' is where we need to be with a recurring revenue stream."
The analogies continue: "If the landlord doesn't attend to the plumbing in the apartment, we will move. Therefore, making that design [work for engineers] is where the challenge lies [in EDA]."
Aart acknowledges that, financially, it's been difficult for the company to make the transition to the new licensing model. Synopsys made the transition "cold turkey" and the accountants "loved" the transition, he says with irony. Investors saw $229 million in revenue last quarter (Q3 2000) move to a projected quarterly revenue of $125 million for Q4. But "once you do it," according to Aart, within 18 months you can expect to see the revenue numbers stablize and then reflect the accurate growth of the company. Like many in EDA, he is convinced that, overall, it will help the industry to make this transition.
"I do believe that an industry, like a company, goes through a maturation process. In particular, adolescence to adulthood. You start to behave in a more constructive way. The whole industry depends on each company doing well. [We're seeing] a natural growth for this industry. In some industries, the companies [compete] each other to death. But, my sense it that for EDA, everyone in EDA will do better [with the new business models]."
As far as market share is concerned, Aart addressed the oft-stated conundrum in EDA: a part of a limited pie or grow the pie. He argues that it's the pricing scheme in the industry that does significant damage to EDA. "You'll do a lot of discounting to make your quarterly numbers. [But it shouldn't be] just willy-nilly. In fact, competitive pricing is better for all of the other companies as well. Therefore, removing that [fixed pricing model] makes the whole industry healthier."
I asked: "Does that mean the EDA industry is not healthy?"
Aart replied, "It does not"
How about the perception that Venture Capital wants too much control of a young company?
Aart says, "Strong teams with good products don't need to give their souls away to 'evil' VCs. [The team] can establish the ratio of who owns what. And, if they don't like it, [the VCs] can move on. 30% to 50% of capital should be devoted [to further development] of the company. After all, having a large share of a loser is not the way Venture Capitalists make money. They make their money by having their share of a great winner. VCs have no time. Their ideal situation is a company that does well and one where they don't even have to attend the board meetings." Aart added: "But, my experience may not be the norm."
Is Synopsys assisting small companies these days? "Synopsys has a small investment fund and we are involved in helping start-ups which are involved in technologies peripheral to where we are."
Aart again insists that he's not trying to be whimsical when he uses an analogy to clarify what someone is trying to say, and using fairly few characterizations.
For instance: "The DNA of a company is the actual behavior that reflects how they feel about themselves. They can be bullish or defensive. I would argue that [for the] long term survival [of a company], childhood diseases and other ups and downs are [important for] companies that are very successful. If one is unprepared if you've never seen an earthquake then you never secured the water heater in the basement. GE, HP are companies that are in a different league [because they knew how to be prepared]."
I asked him to explain, however, how companies can stay innovative as they grow to the size of GE and HP.
"If you make it [as a company], then you need a portion that is new and fresh and a portion that is battle-hardened that knows why you're there. [It's the same] with your own kids; you let them take some risk. In terms of a company that [translates to] the creative management of chaos."
How does Synopsys continue to grow?
"We're already very interactive in the U.S., and in Europe we have three or four centers as well as operations in Bangalore, India. The company has a network of management, local-engineering managers, and product groups [there]. We're developing Design Sphere for the use of tools over the Internet. The thing is still in its infancy, but geography can be partly removed electronically. One can then work where the business is not and certain tasks can be done [remotely]. This thinking is [also] in its infancy. [It will lead to] an increased [design] bandwidth on a worldwide basis. The infrastructure [of the Internet] is incredibly powerful."
Getting Away from It All
In talking to Aart, the impression is strong: His life is, in many ways, synonymous with the life of the company. I asked Aart if he's ever able to get away from it all. He does play the blues guitar and, on rare occasion, is able to pursue his oil painting.
But, not surprisingly, he says the company requires virtually all of his "mental attention. Not a day [goes by], when I don't think about a challenge, an issue, or an opportunity [related to Synopsys]. The issue in not going home at 5 PM. The more difficult challenge [is to leave it behind].
Subsequently, Aart has just been subjected to a Daddy Performance Review. His 8-year-old gave him a low score somewhere around a 5 on a scale of 1 to 10. The 5-year-old, not wanting to be overshadowed by the first reviewer, gave Aart a 1000 tentatively hoping that this score also falls somewhere between a 5 and a 10. Aart's pretty convinced that performance reviews depend on the guy giving the review.
Reviews, or no reviews, Aart's sure that "you always need a common vision for what you want to achieve. It's important along with the work environment, the salary, the upside potential for everyone, [as well as] the people you work with and what they do." He's clearly proud that "Synopsys has had a large impact there's virtually no sophisticated chip today that's not been designed with Synopsys [tools]." And, he's clearly even more proud that "Our people are excellent."
Aart appears to simultaneously celebrate and lament the fact that, today, designers are fully aware that "what can be fabricated is more than what can be designed. Clearly, the level of investment needed to sustain design at Moore's rate is enormous." With Aart leading the call for additional funding for EDA, the industry may very well continue to hope for the best.