Last of three parts: How to measure success, failure, and everything in between.
EDA, arguably more than any other industry, has been built on the backs of engineering breakthroughs by startups. In aggregate, those startups are the backbone of tools that have made cell phones smart and which helped improved gas mileage on automobiles. Through an almost continuous stream of acquisitions, these startups have added to the top-line valuation of big EDA companies, and despite their usually narrow focus they often have played pivotal roles in automation across broad areas of design
But what determines whether an acquisition is good or bad? There are two ways to look at this. One is statistical. Does an acquisition add to the bottom line of the company that bought it, and how quickly does that happen? The second involves people. How many of the acquired company’s employees stay on after the acquisition—considering engineering teams are often the real assets of a company—and were they more or less upbeat about the deal six months or a year later?
Both are critical components to the success or failure of an acquisition, and while the former is the one that gets the most attention by venture capitalists and analysts, the latter has a big impact on the long-term value of the acquisition.
Reasons for signing on the dotted line
Part of what makes acquisitions successful is why companies are willing to be sold in the first place. In the case of public companies buying other public companies, it’s all about money and return on investment to stockholders. Employees, including board members, frequently don’t have the ability to stand in the way of a good offer.
Ironically, public companies buying other public companies are frequently less successful than big EDA companies buying smaller private companies. While EDA requires scale, it doesn’t require the kind of scale that physical manufacturing or consumer electronics does. And in EDA, there are rarely deals that fall apart at the last minute. In fact, the only one that achieved any public recognition in recent years involved Synopsys’ 2004 announcement that it was considering buying Monolithic System, aka MoSys. After months of due diligence Synopsys decided against the deal. No reason was made public.
Still, in comparison to deals done in the 1990s, the majority of EDA acquisitions these days are a bargain. Prior to 2001, the preferred avenue for EDA investors was an IPO, which jacked up the price of startups. But with the collapse of the IPO market after the dot-com bubble burst, the only remaining exit strategy has been selling companies to one of the Big Three—or to outside companies where there is some synergy, such as the sale of Apache Design to Ansys. In fact, the last EDA company to successfully launch an IPO was Magma Design Automation in 2001. Magma was bought by Synopsys in February 2012.
This isn’t the end of startups. But it has changed the reasons why at least some engineers get involved in EDA startups. Particularly for bootstrapped companies, acquisition or rapid growth isn’t always the goal.
“When you take on venture funding, you have a boss and that boss wants you to grow fast,” said Srinath Anantharaman, CEO of Cliosoft. “When you bootstrap, that’s not true. The initial few years are a struggle, for sure. We started as a different company doing consulting services, and the product was a side project. But when we moved to a subscription model, we could guarantee at least 90% of what we were going to make each year.”
That takes the financial burden off the company, and it allows engineering teams to do what they do best—and what they most enjoy. It also removes external deadlines for when new products need to roll out because sometimes those development efforts can be shifted mid-stream to more strategic use of resources.
“Most startups begin with a hunch that they believe they have something useful,” Anantharaman said. “Steve Jobs did that. We never had an exit strategy, though, because we enjoy what we do. We’ve had no turnover because of that.”
It’s hard to tell how many private startups there are out there with this kind of attitude, but the consensus among investors and other EDA companies is that there are more than you might expect given the lack of activity in the IPO or acquisition world. It doesn’t take a lot of engineers to form an EDA startup, as long as they’re not working in areas such as layout or in direct competition with established companies.
But complexity in design also is taking its toll on the startup community, similar to what has happened to self-automotive repair. Most people stopped doing it when the complexity and cost of entry began increasing.
“In the past you could do a startup with a very narrow solution,” said Andrew Yang, president of Apache Design Inc. “Now, your skill set needs to be bigger and the problem definition has to be very crisp. Opportunities are still out there, but execution is harder.”
Longer waiting periods
How long these startups remain independent is uncertain, too. While most profess to want to stay independent, that isn’t always the outcome when someone dangles a sizeable offer in front of them.
“You’ll see two things happen in the future,” said Jim Hogan, a long-time EDA VC. “One is the creation of strategic companies, which fit very neatly into existing companies. The second is the aggregation of lots of little guys into bigger companies.”
Startups still have a reputation for developing innovative point tools from scratch, but they don’t have mature channels for getting those products sold. The synergy of an established channel, coupled with one or more products that solve very specific problems in design or verification, can be a powerful motivator for startup leaders who historically have turned around after acquisitions and started up new companies. But with less VC capital flowing to startups, the overall number still has dwindled, which is why the acquisition targets of the Big Three EDA vendors are getting larger.
In the past three years Synopsys bought Virage Logic, Magma and Springsoft. In the same period, Cadence bought Denali, Tensilica and Altos Design. Those larger acquisitions also take longer to digest, slowing down the appetite for more acquisitions of any size.
This isn’t necessarily good news for the big established companies, or for engineers looking to dig into interesting and challenging problems. But leaders of the Big Three EDA companies, which have done the bulk of the acquisitions over the past couple decades, still believe there is plenty of opportunity for startups—particularly those that get their funding from angel investors or by bootstrapping it themselves.
“We are in the fastest-moving, deepest part of high tech,” said Aart de Geus, co-chairman and CEO of Synopsys. “This is a meritocracy, so it doesn’t matter where you are. If you move the ball forward, you have an opportunity.”
He said huge opportunities are opening up in artificial intelligence, automotive electrical design, and software at every level. Startups will be needed for that as well as established companies working to build integrated tools and IP. He’s not alone in that perception.
“I favor a lot of startups,” said Wally Rhines, chairman and CEO of Mentor Graphics. “Most of the revenue growth in EDA comes from new problems and new capabilities. We’ve seen that with RTL simulation and place and route, but those are mature problems and they’ve been flat for 10 years. The real growth is new problems. EDA has the advantage of not driving up the costs the way fabless semiconductor companies have driven up costs because you can’t put that many people around a problem.”
Rhines said that where acquisitions go awry is when there is a cultural incompatibility—difference over ethics, for example, will ruin an acquisition—or where a company pays too much or too little.
For an industry that has been built largely through acquisition, though, the shrinking number of acquisition targets raises some interesting questions. EDA’s success with acquisitions is unparalleled, and the question now is what will happen when there are fewer acquisition targets. Can the level of innovation continue at the same pace and for the same cost?
At least part of that may be answered by what happens in semiconductor design. Startups may play a lesser role at 7nm, for example, than with stacked die and IP—or they may play a greater role at older nodes and the Internet of Things. But wherever they appear, the EDA industry certainly will be watching with interest.