EDA startups of the future will look very different. They have no choice.
I have been involved in several startups during my career, both true startups and startup ventures within a large EDA company. While they both demand innovative solutions, they have a very different goal and structures.
When researching my story about innovation within EDA (yet to be published), I expected to receive a lot of large company bashing. While some pointed to sales strategies that made it difficult for startups to be successful, and other the difficulties created by highly conservative design companies, they also were able to look beyond this to other reasons, including systemic failures within the ecosystem that have made it difficult. One message that came through strongly was that attempting to displace existing solutions was not the way forward.
My first startup was a result of my academic experience. I had been sponsored through university by an aerospace company. As part of my final-year dissertation, I took up a problem I had experienced during my time spent in one of their development teams — debugging software in a physical prototype. It was early days for many of these systems and they were so expensive that we only had one to share between both the hardware and software teams.
I set about trying to design a re-targetable in-circuit emulator. The university commended the aggressive nature of my project but informed me that it could not afford to let me attempt to build one. They referred me to a group of post grads who were working on a logic simulator. I became their guinea pig and set out to find out if both my design would work, and if the simulator worked. Both were successful. Soon after graduating, that simulator was spun out of the university, and I was asked to join the team.
Back then, there were no large EDA companies to compete against. In fact, we were the clear technology leader, but that still didn’t make it easy. There were many hurdles to overcome to make a sale. Like many successful startups, we had a solution that solved an existing problem (gate-level simulation), but also invited users to try something new (RTL simulation, even though the term RTL had not yet been coined.)
It wasn’t long before we were bought by a larger company in America, which had its sights on becoming a major entrant in the emerging EDA marketplace. That company didn’t make it, limiting any further innovation in the product that did not fit in with its long-term corporate goals (which was test.)
After a couple of less successful startups later, both of which were related to hardware assist for EDA problems, I found myself in a small EDA company that was destined to become the largest company in the industry. Within that company, you might have expected to find a highly innovative culture, but that is not what I experienced. Even at that young age, there was a reluctance to do anything except incremental innovation by many of the senior technical people. They did not want to see money that their early tools were generating being used for the development of new and more speculative ventures. Management was in favor of a more diversified revenue stream, but the engineers were reluctant.
Later I would find myself in another of the major EDA companies. A new market opportunity had been identified for which there were a couple of exiting startup solutions. We evaluated them and determined they did not meet the market goals. We created a business plan and submitted it to company management. They allocated us some capital and sent us on our way.
This was an experiment because we had been given total freedom. There were no corporate ties, no constraints. We could use the money however we wanted, but if it ran out, the gig was up. We set our own development and sales goals, we hired wherever we wanted, we did almost everything differently from the rest of the company. We even integrated into competitor’s solutions before internal ones because that is what we needed to do to be successful.
It worked, and was the closest thing I have ever experienced to a startup that was not a true startup. The one exception was that we didn’t have the same fear. It was unlikely that the parent company would fire us if we didn’t succeed. The company did try and replicate this success, and some of them were quite successful as well. In fact the entire company almost became a VC, allocating capital to good ideas, with the expectation of a return on capital.
Marc Swinnen at Ansys spoke of the ways in which things are happening within his company today. While Ansys is a fairly large company, the EDA portion of it remains quite small. However, that portion has quickly become a very important part of the EDA landscape, and the company needs to be nimble in the development of new tools to retain its competitive advantage. Swinnen talked about the startup within a larger company culture, and it reminded me that in change is opportunity. Ansys has an established sales channel, the market knowledge, and the capital to take away many of the risks and difficulties faced by startups. In addition, the structure of its solutions allows external players, be they customers or academics, to extend and modify their solutions. In effect, Ansys is allowing them to participate in the innovation process.
As Rob Knoth at Cadence said, we should not perhaps try and replicate the ways of the past, but look for new ways in which to be successful, new ways to foster innovation and collaboration.
Innovation in business structure may be as important as technical innovation.
Leave a Reply