What Would Joe Do?
Peggy Aycinena is a freelance journalist and Editor of EDA Confidential at www.aycinena.com. She can be reached at peggy at aycinena dot com.
Camposano & Zelnik: Sage-DA promotes EDA Evangelism
August 27th, 2015 by Peggy Aycinena
Blogs are a dime-a-dozen, but you’re going to want to read this one if you want to know why distinguished veterans of EDA continue to evangelize for the viability and vitality of the industry.
On a phone call this week with Raul Camposano, newly-minted CEO of Sage Design Automation, and Coby Zelnik, President and Co-founder of the company, the point was driven home repeatedly: There’s as much of a future in EDA as there is a past, no matter what the current demographics may imply. Evolving demand in the CAD-tool marketplace means EDA companies will continue to emerge to meet that demand.
Coby started the conversation, “Sage-DA’s mission is to really change the way design rules are handled, and how information is exchanged between the foundry and the designers who continues to work off a paradigm that was established 30 years ago, or more.
“It’s true, things may have changed a bit [in the interim], but design practices and flows have not. Meanwhile, we are moving to advanced nodes that make the rules very difficult, so entire flows and techniques for enabling design have become even more difficult and lengthy.
“What we want to do at Sage-DA is to change the way designs are handled, all the way from the beginning to the end. How designs are defined and validated in the fabs, how they’re represented to designers, and how designers find those representations in the design rules and flows.”
“Is there possibly overlap between Sage-DA’s offerings and PDF’s,” I asked.
“No,” Coby said. “PDF is focused on services and on developing their own software to the extent they can deliver their services more efficiently. Also, they’re focused more on yield and less on enabling EDA tools for designers. There’s no real overlap between our two companies.”
“When was Sage founded and who are your investors?” I asked.
“At the end of 2012,” Coby said, “so less than 3 years ago. I founded it with Maarten Berkens, who is CTO. As far as investors are concerned, I would rather not make them public.”
“It takes a lot of guts to found a company today in EDA,” I said, “given their chances of survival continue to hover around nil. They don’t survive, there’s no viability.”
To counter that heresy, Raul weighed in: “The EDA business model is now well understood and mature. I don’t know the statistics specifically, but if you compare EDA to other industries [our investment returns are] not terrible because we get valuation based on channel issues.
“Investors don’t value your channel, however, they value your revenue. So although there’s no huge upside, there’s still great potential for creating value. If the investment is not too rich, there is room for a very decent return.”
Chuckling, he added, “Yes, in the 1990s there were several huge IPOs in EDA, but those were small revenue/big IPO [situations] and we all know that will not happen again.”
“No more IPOs in EDA?” I asked.
Raul replied, “To be a little inflammatory, the average age in EDA companies these days is approaching 50 years old, and in US-based EDA companies it’s even higher. So the potential for innovation and for attracting young people who want to create new things is very limited.
“That’s why EDA companies are regularly acquired. Most EDA companies, those who have customers and create [solid] technology, are acquired. So yes, it’s difficult to foresee if there will ever be another IPO, even though Apache almost did it [before being acquired by ANSYS].
“If an EDA company can get into the $50 million range, however, there’s still a chance.”
“Is Sage-DA headed into the $50 million range?” I asked.
Raul chuckled again, “We would certainly like to. Although clearly we’re not there yet, we are providing the innovation that’s needed today.
“There is a lack of software and tools in the fab to help you manage information, the design rules that result in specifications, tech files and DRC decks. If you have tens of thousands of rules in those decks, you’re crying out for more information. Things need to change to make that happen and that’s where Sage-DA is focused.”
“If you look at the Synopsys story, which I was part of,” Raul added, “the designers used to love to use schematic capture. That was the way people understood their designs. But when the gate count [grew too large], writing in VHDL or Verilog was the only answer.
“Similarly today, everyone knows how to use DRC decks. But there are now tens of thousands of rules, so it is a space that is shouting for, clamoring for innovation. At Sage-DA we will do it.”
“Isn’t the solution set in EDA already fully defined?” I asked.
Raul replied, “The solution set in EDA is not fully defined. Yes, there are a lot of solutions, but there’s still a lot of room for innovation. And there are many early-stage companies [with solutions to contribute], but they are often short on investment dollars.
“Clearly it’s not the same as when we were all first inventing the industry in the 1960s, 70s and 80s, a time when we had just learned how to design ICs. But now today, we are quickly approaching a trillion transistors on-chip and there is definitely room for innovation in that!”
“If there’s that much opportunity in EDA, why is it so difficult to attract young people?” I asked. “Is it because, as I’ve heard so often, young people today are too soft for the challenges of EDA?”
Raul was animated: “That is not true! Young people today are not soft and they know there are an incredible number of opportunities in EDA, but young people are attracted to younger industries.
“Every industry matures and it’s hard to attract top talent to EDA when those software companies [we are competing with for talent] are growing faster. There are a lot of alternatives for students coming out of school today, it’s just a fact of life. If somebody sees that the problems in EDA are hard or harder than other domains, and EDA companies are not growing as fast, they’re probably going to go elsewhere.”
“So surely there must be some financial reward to attract young people to EDA,” I noted.
Raul responded, “You’re not going to become a billionaire in EDA, but if you work hard and are particularly [diligent], you will succeed. Of course, we’re all in it for the money, and the returns are certainly decent in EDA, but it’s not easy. It’s totally doable, but not easy.
“When you do succeed, however, you can sell the company for $50 million. An excellent return, particularly if you haven’t taken a lot of investment money. Software, EDA included, is always written by teams that are not huge, so the returns for any one individual are very decent.”
“That brings up another question,” I said. “What is the optimal size for an EDA company?”
Raul had an answer: “The company can be bigger, or smaller, but at most it involves tens of people.
“If you remember The Mythical Man Month, essentially the theory said that all software is done by no more than 4 people. Of course, it depends on what you’re doing.
“The problems in EDA are highly algorithmic and take a long time to solve. With social networking software, on the other hand, things are based on statistics and what others are doing, the trending [thing]. The more people you have contributing to the product the better. But EDA is deep stuff and not easily understood by [the average coder].”
I offered an irreverence: “So EDA is possibly too difficult for most people. The industry is actually composed of a small priesthood of the enlightened?”
Raul was not offended: “The typical problem in EDA is very difficult. Of course, once you get [your solution] into volume, once you’ve deployed it in the field, then it requires a lot of AE support to help the customers. That effort does requires a huge team.
“But what I’m talking about is the small team [for initial development] of the core algorithms.”
“How do you feel about EDAC?” I asked. “Does it still have relevance as the EDA industry consolidates down to just three real players?”
Raul responded instantly: “I love EDAC! I was a director at one point and seriously consider the work of EDAC to be very useful and very worthwhile for three good reasons.
“First, the Market Statistics Service, though not very precise because the big companies may or may not be entering their data, is very useful for the small companies who cannot afford their own market intelligence.
“Second is the lobbying aspect. I was part of the effort to get the government to allow us to export SPICE to China and certain other communities. It wasn’t easy to get that done, but EDAC did it.
“Third, and to some extent most importantly, there is the EDAC work on standards. As an industry, we all agree it’s better to do things [in concert here]. We need to not be supporting, for example, 28 different versions of Linux, some old and some new. The EDAC work on standards is really important for all of the companies in the industry.
“Of course, there’s also the networking aspect of EDAC. Overall, every industry needs an association that speaks for it and EDAC does that for EDA.”
Ending on possibly the thorniest issue facing EDA, I asked, “What about re-aggregation in the semiconductor industry? Will consolidation among chip companies precipitate widespread in-house development of CAD tools? There really aren’t enough customers left for even as few as three EDA companies to service.”
Raul was unmoved by the prospect: “You’ve probably read the Synopsys announcement that Intel is more than 10 percent of their revenue. That’s more than $220 million in revenue from Intel alone. Intel is spending a lot of money on third-party tools, because they also buy from Cadence.
“Clearly there is no danger to the EDA industry from what you call ‘re-aggregation’. In fact, there’s only more efficiency for the customers who rely on independent EDA tool companies, the people who actually develop tools for a living.
“Tools will never be a core technology for Intel, or [any large semiconductor company]. EDA software is just too different from selling computers or chips, and Intel and the other large companies know it.
“Also, if you were to work in a support role at one of the big semiconductor companies doing tools for design, you would quickly see that these are not problems that naturally attract great engineering talent in the semis.
“People want good careers, so people will always gravitate to the core technology of any business, and that’s not tools for the semis. People who are truly interested in EDA will work instead directly for third-party tool vendors where tools are the core technology.”
Raul added more justification for his hyper-confident prognosis: “Also, far from just having a few IDMs or foundries in the world today, the fabless model has been growing very strongly over the last few years.
“Looking at those fabless customers, there are 26 today bigger than $500 million per year. That’s 26 companies doing chip designs. If those customers each spend 1-to-2 percent on their manufacturing and design tools, that’s several million dollars per tool customer. A very reasonably sized market for the EDA industry.
“And that’s not counting the IDMs like Intel that are spending $400 million to $500 million each year on tools. There’s Samsung, Toshiba and other IDMs that are growing. Again, a very reasonably sized market for EDA.”
“Also, statistics from Gartner shown at Semicon West last month say the number of fabs worldwide is now at around 300, including the analog fabs, and will stay that way for at least 5 years. This also is a very rich ecosystem of potential customers for EDA.
“Yes, there’s been consolidation, but concluding that this trend spells the end for third-party tool vendors is a total exaggeration!”
“Okay, the future is good,” I said. “So what’s the exit strategy for Sage-DA?”
Raul laughed, “We don’t have an exit strategy. We’ll think about that when it’s time, and yes, that may be when we get into the $50 million range. But things can change, so I won’t say [that metric] is set in stone. Of course most companies will be acquired, but there’s always a chance for another type of exit.”
“By the way,” Raul concluded, “it’s not just three big companies in EDA. ANSYS must also be included on the list. It’s larger than Synopsys, participates in EDA, and acquired Apache.
“In fact, there’s nothing to prevent even more companies becoming big EDA companies in the future. Things can always change!”
Santa Clara, California – May 6, 2013 – Sage Design Automation (Sage-DA) has been founded to develop technology and products that automate the rule-based design and verification paradigm. Sage-DA was founded with initial investment from venture capital and angel investors including Alex Shubat, PhD, former President and CEO of Virage Logic (acquired by Synopsys in 2010) and Michael Burstein, PhD, EDA veteran and co-founder of multiple EDA companies. Coby Zelnik, former CEO of Sagantec leads the company as President and CEO.
Santa Clara, California – August 17, 2015 – Sage Design Automation Inc., the company that is redefining physical verification, announced today the appointment of Dr. Raul Camposano as its CEO. Dr. Camposano has had a distinguished career both in industry and academia. He was CTO, senior vice president and general manager at Synopsys for over 10 years. He also served as CEO of two startups, Xoomsys and Nimbic (acquired by Mentor Graphics). Dr. Camposano is active in the EDA and semiconductor professional community, serving on various boards and technical committees. He is a fellow of the IEEE and holds a Ph.D. in computer science from the University of Karlsruhe.