- Published on 02 Jan 2019
Interview: HiPEAC19 keynote speaker Alberto Sangiovanni Vincentelli
The HiPEAC conference takes place on 21-23 January in Valencia, Spain and features three keynote speakers. Alberto Sangiovanni Vincentelli, who will be speaking on Tuesday, 22 January, the Edgar L. and Harold H. Buttner Chair of Electrical Engineering and Computer Sciences at the University of California at Berkeley, is the founder of leading electronic design automation companies Cadence and Synopsys. In this interview, Alberto explains how he got into computer engineering, where he would start a business now and what the future holds for chip design.
How did you first get into computer engineering? What then attracted you to your areas of focus, like cyber-physical systems and design automation?
I actually got into computer engineering in a roundabout way. I started as a theory person working on systems and numerical analysis. When I moved from Europe to California in 1975, I was lured into more practical side of engineering: integrated circuits. In 1983, the algorithms I developed with colleagues such as Richard Newton and Bob Brayton caught the attention of industrial concerns such as Intel, Motorola, ATT and others, so we decided to start two companies: Cadence and Synopsys. These two are now the leaders in electronic design automation and are public companies traded on NASDAQ.
In 1988, I decided to move towards cyber-physical systems with a foray into the automotive domain when I started collaborating first with Magneti Marelli, and then Mercedes, BMW and General Motors. Recently I have been involved in the aerospace domain with United Technologies Corporation and in construction with Lendlease.
Where do you think chip design is going to go over the next ten years, given the end of Dennard scaling and Moore's law?
The most interesting direction in my opinion, given my interest in cyber-physical systems, is the development of sensors and actuators (the ‘more than Moore’ direction) whereby the integration of different technologies and physical domains is taking place. The increasing cost and complexity of moving to new nodes is making life difficult for most of the semiconductor companies, hence the Global Foundries announcement that it won't be pursuing 7nm technology. I personally believe that the strong movement towards consolidation is going to substantially reduce the number of viable semiconductor companies, even if governments are trying to oppose some of the mergers and acquisitions in this domain on the ground of strategic national interests.
Would you rather start up a business in Europe, the United States or somewhere else? On the flip side, where would you rather study?
Today, I would still start up a business in the United States, although I have started businesses in Europe. The Silicon Valley ecosystem is still the best in terms of offering opportunities to entrepreneurs, even though the cost of locating in Silicon Valley and the scarcity of human resources is making this area of the world less appealing. Of course, China is the big land of opportunities even if it is quite difficult to start a business there considering regulations and the weakness of business protection legislature. Singapore is another country where the government is very business friendly and it can be used as a gateway to East Asia.
On the education side, I am very grateful for my Italian high-school education and for my years at the Politecnico di Milano. This education gave me a broad cultural base that has served me exceptionally well both for my academic and industrial career. However, I am sure that combining this background with a PhD in United States would have given me insights and competences that I had to build on my own.
What advice would you give someone thinking about setting up a technology business based on their research?
Be sure that your business is ‘defensible’ – i.e. based on strong scientific foundations to protect your ideas beyond the protection that can be provided by patents. Intercept the future needs of your chosen customer base to build a large business over time. Do not think that small/medium enterprise (SME) status is the final game. Think carefully on how to scale up your business over the years.