Dr. John Sanguinetti is CTO of Forte Design Systems, a high-level synthesis company he co-founded. Previously, he founded Chronologic Simulation to develop VCS (he was principal architect), and co-founded CynApps. Sanguinetti has a Ph.D. in computer science from the University of Michigan. –
In the following interview, Sanguinetti discusses the ethos of the Entrepreneur in EDA. If you want to know more about Dr. Sanguinetti, an extended profile is available on EDA Confidential.
* Q – Do you have to be rich first to be an entrepreneur?
* John Sanguinetti – No, you don’t have to be rich to start. It probably helps to be in a position where you don’t have to support your family, but even that’s not a requirement.
If you’ve got a spouse that can support you while you’re in the low income phase, it helps, but these days it’s gotten harder to do a startup on a shoestring. It’s gotten to be much more difficult without external funding. I know several people today who are trying to get companies going, and living without any funding. Eventually, they’ll have to get funding.
Fifteen-to-twenty years ago, people did it fairly commonly in EDA, but now the bar has been raised. Where 3 guys working in a garage for 18 months used to be able to get a product out, now that’s rare.
Now it’s more like 6 or 7 guys working for 3-to-4 years, because the products have to live in a more complicated environment. Still, it’s uncommon to find 6 people who can all work for nothing for 3 years or more, which means you have to raise money. Whether you’re rich or not doesn’t matter.
* Q – Is it complicated because of the technology, or because you’re going up against major players?
* John Sanguinetti – It’s not so much going up against the major players, it’s that you have to be compatible with them. If you’re going to make a product that’s going to fit into the flow, you’ve got to be compatible with what’s on either side of you.
Today the design flow is a richer environment, the interfaces are more complicated, the languages are more complicated. You’ve got to do so much more work to deal with all of it. Here at Forte, for instance, we absolutely have to be compatible with the tools downstream.
* Q – Does increased complexity in the hardware also have an impact?
* John Sanguinetti – If you’re up in the front end of the design flow it’s not that much of an issue, but at the backend it probably does. Logic design, for instance, is not that much more complicated. There’s just more of it.
Again, the major players have put a lot of effort into making their design environments as rich as possible, and somewhat different from each other. That’s how they lock in their customers.
For instance, Synopsys has put a lot of features like Vera, additional language features, and utilities into VCS, but those features are not standard. If you, as a customer, want to switch to someone else’s simulator, you have to give up all of those features. If there’s enough of those features, it becomes a real hindrance in trying to switch.
It’s not just Synopsys doing it. Cadence and Mentor are doing it, as well.
* Q – So the big players are creating a higher barrier of entry for startups?
* John Sanguinetti – Although they would certainly like to do that, I’m not sure anybody has a real strategy: “Gee, if we do this we can lock out the entrepreneur, the little competitor.”
Their real intent has been to lock out their big competitors. The little guys are just collateral damage.
* Q – With that reality, how do I go about becoming an entrepreneur?
* John Sanguinetti – As always, you have to have an idea. The best one is where you’re solving a problem that nobody else’s tools have solved. Just displacing a single tool in the design chain is really hard, and also of questionable worth. But if you’re adding a tool, you can still do that as an entrepreneur. It’s just a little more expensive than it used to be.
Most importantly, you have to have an idea!
* Q – If you get an idea while working at a big EDA company, don’t you risk being sued by your employer after you leave to start your small company?
* John Sanguinetti – Traditionally, it was engineers in the user community, not EDA, who started tool companies. They’d say, “I’d really like a tool that does this particular thing, and I have a problem with the Synopsys or Cadence or Mentor tool. I think I could do it better.”
Occasionally you’ll get somebody who’s been in EDA, who thought they could do something better in a particular part of the design flow, solve a problem that’s not being addressed. But, I don’t know of many cases where somebody has left one of the big companies, and been threatened with a law suit. There have been a couple of high-profile ones, but not that many.
* Q – Is anyone even contemplating a startup in EDA today, or has it all been done?
* John Sanguinetti – No, people are still doing it. I’ve been working with a number of people recently who have been trying to get EDA startups going. The biggest problem they face is raising money. It’s just really hard to bootstrap one of these companies.
That being said, I actually know one that has bootstrapped itself and is definitely going to make it. And there are other startups that companies like Synopsys and Magma are putting money into, as well as Andy Bechtolsheim.
* Q – Speaking of Bechtolsheim: Is there a cookbook on how to become a successful entrepreneur?
* John Sanguinetti – The easiest way is to hit a home run the first time, then you’ll have lots of latitude to try again, and fail, and fail again, several times. But if you fail the very first time, you’ve got to have the gumption to stick it out. It’s pretty tough.
I do know a guy who’s done that, but not in EDA. You’ve got to take your hat off to someone like that. They really want to create something and when it doesn’t work, they try again, and again, and again.
* Q – Best if you have someone in your household who can support you?
* John Sanguinetti – Either that, or you’ve got to have the resources to survive without a steady income.
* Q – How do you see the influence of VC money in EDA?
* John Sanguinetti – The VCs have run hot and cold on EDA, but I’ve been amazed over the last 10 years at the number of VC deals that have gone into EDA startups.
Particularly because, if you’re a VC, I have a hard time understanding why you would invest in any EDA company. Our industry just doesn’t support the VC model.
* Q – What is that model?
* John Sanguinetti – You need a 10x return on your money in 5 years. Maybe you can get that return if you put $100,000 in, but you can’t get it on $5 million. And, how many EDA companies can go from 0 to profitability on $5 million?
Look at CoWare. How much money did they take? I don’t know the number, but I believe that it was over $30 million, and I think that Synopsys bought the company for about $85 million. [Editor’s Note: The terms of the acquisition were not officially disclosed.] Who made a 10x return on that money? I don’t think anyone did.
* Q – What about the reported $125 million that went into Monterey Design?
* John Sanguinetti – Yeah, but that was an outright failure. Look at Magma: how much money did they take in? – They were successful, went IPO, and at one point had a market cap of approximately $2 billion. Some people made 10x on the money they invested in Magma, although I doubt the institutional investors who invested in Series B, C, and later rounds made 10x.
* Q – If an IPO exit strategy no longer appears viable, today it‘s only exit by acquisition?
* John Sanguinetti – Considering how many companies in EDA have gone public in the last 15 years – if you’re a startup, [the reality is] if you are successful you’re going to get bought. You can pretty much count on that being your exit.
But again, with current acquisition prices [running] between $20 million to $50 million, nobody’s making 10x at those prices.
* Q – Perhaps a different VC model is needed?
* John Sanguinetti – That’s right, or maybe what VCs want has changed. Maybe now it’s 5x in 3 years. But how often do you see an exit in 3 years?
Along with it getting harder, it’s also taking longer and more people to get your product out, and that affects your exit. You won’t be bought in 3 years; you’ll get bought in 5 to 7 years.
* Q – To guarantee acquisition, do you need to make the product dovetail exactly with a specific Big Guy? Aren’t you taking a risk that they’ll do it themselves.
* John Sanguinetti – The big companies like to say that, but they know it’s not true.
Say you’re Synopsys or Cadence, and you’re looking at some little company that’s got a product that sort of works and seems exciting to some early customers. Do you say, “It took them 3 years to make it, but it will only take us 2 years.”
No. Even with all of that, it’s still better to buy the little guy than to do it yourself.
* Q – What are the biggest challenges of being acquired? Technology? Personnel?
* John Sanguinetti – The biggest thing is to have the proper expectation. If both sides are honest with each other, and they set out what the expectations are, and everybody kind of knows what’s coming – you’ll have a much better chance of success.
The technology part is not the problem. You have to exist within the design flow, and you have to integrate and coordinate with other products. It’s often the case that the little company can actually do a better job with that than the big company. As a small company, you make a single product and it has to fit within the tool flow. If you don’t, your customers can’t use it.
If you make a verification product, for instance, and it’s going to do something with the design source, you might make it work with only one commercial simulator because your resources are limited. If you then get acquired by someone else, you end up having to make it work with within the acquiring company’s verification environment, but that’s not that hard.
In mergers and acquisitions, it’s the human side that is so tough!
Typically, a handful of people in the acquired company sign an employment agreement and commit to staying for 1, or 2, or 3 years. But usually, not many people have those agreements. It’s pretty rare for most people to have to sign them.
If you are one of those people, however, if you stay in the company that’s acquired you for 2 years – that’s a pretty long time, at least as far as the time needed to get a product integrated into the acquiring company’s technology.
If the acquiring company is doing a really good job of promoting the product, and the technology and the people are happy, they won’t leave. It’s completely up to the acquiring company to do their part to keep these people on board.
* Q – If you sign a non-compete agreement after you’ve been acquired, what do you do during the time period of the agreement?
* John Sanguinetti – First of all, you don’t really have to agree to it. When Chronologic was acquired, I agreed to 4 years of non-compete that covered everything in EDA. It was a big mistake for me to sign it.
* Q – Didn‘t the sale depend on it?
* John Sanguinetti – No, it didn’t, but it was presented to me as a must-be-signed. Had I said I wouldn’t sign it, they would have changed it. [Actually], I would have been better off if it had killed the deal. This is old history now, but it was pretty much known at the time.
* Q – But what if the non-compete is required?
* John Sanguinetti – The first thing is, you have to decide how much you want the deal to happen, what you’re willing to give up to make it happen.
Then, the non-compete needs to be reasonable. More than 2 years is not reasonable, and covering the entire EDA industry is not reasonable. It should only cover the technology of what’s being acquired, and then you’re free to go off and do something different, but still in EDA.
And, if it’s an employment contract and you’re going to be locked up for 2 years [working for the acquiring company], your non-compete should be up at the end of those 2 years as well.
* Q – What‘s the right metric for a successful serial entrepreneur? 5 attempts with 2 being successful?
* John Sanguinetti – 5 is a lot, and can cover 15 years. Even George Janac doesn’t have that many! [Note: George actually might have done that many.]
* Q – Okay, so what‘s the right number?
* John Sanguinetti – If it was me, I’d say 3. Forte is my third.
* Q – And how many failures should somebody be willing to accept?
* John Sanguinetti – I really have to admire people who have had several failures, and keep at it. But, it’s also the case that you need to have a good idea, and you need to be able to execute on it. There are a lot of people who try to start companies with only marginal ideas.
* Q – Is the concept of the celebrity VC or entrepreneur counter-intuitive to innovation? Tom Perkins or Steve Jobs?
* John Sanguinetti – It’s orthogonal. High profile investors don’t mean diddly-squat to the EDA industry. I can’t think of anyone [like that] who has made a difference to an EDA startup. However, high-profile entrepreneurs – guys like Steve Jobs – can have a real effect on a company.
I worked at NeXT for a year, recruited by a friend. I knew I was in a temporary place, that I wanted to do something else, but I figured I’d [learn something] if I could work near a guy like Steve Jobs.
[During that year], I was in a meeting between Steve Jobs and Joe Costello – it was around 1990. Cadence was starting to really succeed and be a major player. Joe wasn’t a god yet, but he was in his ascendancy. Steve was Steve, and was trying to sell Joe on the idea of Cadence porting all of their software to NeXT machines and selling it as a bundle. Joe [on the other hand] was simply trying to sell Steve on buying more Cadence software on Sun machines. It was a heavyweight encounter.
Joe clearly had the upper hand, but Steve was Steve – even then, he was charismatic, as good a salesman as there is. It was a fun experience being around Steve Jobs.
But the thing I concluded after being around him for a while – it wasn’t that he was so brilliant (even though he is, in several important ways), but everybody in the world wanted to tell him what they were doing. He had more knowledge about where the industry and electronics were going than anybody else, because everybody was telling him. Motorola, Sun, HP – they were all telling him what they were doing. I thought that was a real advantage!
* Q – So there is a benefit from the high-profile, charismatic celebrity investor?
* John Sanguinetti – The Tom Perkins of the world, they’re kind of like that too. They have access to information, and to the extent they can share it with their portfolio companies, [it has an affect].
I haven’t seen that as an advantage in EDA, however. It just doesn’t matter.
* Q – Is that because EDA is so difficult to understand?
* John Sanguinetti – It’s not that the industry is so difficult, but it is very specialized. What your investor has just learned about new trends in social networking don’t really help all that much. And there’s certainly a lot of inspiration and education needed to start with.
* Q – The industry needs to recruit some new young PhD's?
* John Sanguinetti – We are doing that. For example, I know of a PhD coming out of the University of Michigan directly into EDA, and I know there are others at other schools.
But, it requires an understanding of computer science and electrical engineering, and there are only a handful of professors who are knowledgeable about the EDA industry. It’s their students who will end up having direct knowledge and exposure to the industry.
It’s more common [for people to come into EDA] after doing something different, getting to know the industry based on their exposure through work. In my case, I thought I was going to write operating systems after I got out of grad school. That was my goal in life. I thought it was the pinnacle of computer science to write operating systems.
* Q – What changed your mind?
* John Sanguinetti – I went to work for DEC, and was exposed to hardware design. I changed from being the guy who writes the operating system, to the guy who designs the computer. For a while, I thought computer architecture was the pinnacle of the industry, where I wanted to be. That lead into design verification and then, I guess I lowered my sights.
Back in the 1970’s and 80’s, when EDA got started as a discipline, all the design tools were done by EEs with no knowledge of CS. People were solving problems with only a limited understanding of the ways to do it, simulation being the perfect example. EEs had no knowledge of compiler techniques, things that were already common in a general CS curriculum, but not in EE.
That’s all changed today, but it’s still the case that you need to have a pretty good exposure to both disciplines, CS and EE.
* Q – Will other geographies start producing more EDA entrepreneurs and innovation?
* John Sanguinetti – If not here, somebody somewhere will do it. Nobody has a monopoly on knowledge.
Peggy Aycinena is Editor of EDA Confidential.