? In which work steps can AI score and which activities will remain with the human developer?
! Diamantidis: Clearly the more repetitive, tedious, and laborious the task, the better the fit for AI. We see AI as a great human assistant that is able to augment designers‘ ability to deliver things and deliver them in time, and most importantly, make the entire team deliver at expert levels.
? Will AI-powered EDA tools make it easier for newcomers, developing an IC or ASIC for the first time, or will they place even higher demands on the design engineers?
! Diamantidis: Yes, it will make it easier for newcomers to be much more impactful compared to where they are today. With chip design, and particularly certain areas of chip design, like physical design for example, our place of expertise comes a decade at a time, not a day or a week at a time. Having the ability to leverage the knowledge of experienced designers and channel it so that newcomers and junior designers can actually benefit in the real live production environment is perhaps the biggest value that AI brings to the design workplace.
? How will the work change in the future for development engineers who develop systems together, with ICs, SoCs, assemblies and software?
! Diamantidis: Today, I don’t think it’s possible to develop systems together – we have a “divide and conquer“ approach to systems design because the complexity of each and every domain involved is huge. We have many silos that have impact on each other and optimizing those silos in isolation leads to sub optimal system level solutions when you bring it all together. In the worst case, the systems don't work at all.
? How will electronic systems be developed in the next years?
! Diamantidis: We are now entering the age of AI Assistants for electronic system design. AI Assistants will allow domain experts to gain enough visibility into other domains to drive co-optimization organically and in parallel. We see a future where domain experts not only become super experts augmented by AI, but are also able to work better with experts of other domains to develop and deliver optimally-designed systems.
Our CEO, Aart de Geus, has a great example of this. With the great cathedrals in Europe, many of which are in Germany, that were built in times where you wouldn't think it would be possible to build them. The architects not only had to understand the function the cathedral had to serve, and devise new architectural forms, but they also had a working understanding of the properties of the stone that they used for any part of the structure. With AI, we're going to have domain experts that may be an expert in optical design, but will have a working knowledge of how the rest of the system actually works at a pretty deep level.
? Is there a growing interest in Germany and Europe to develop ICs/ASICs themselves?
! Diamantidis: Absolutely, we see a growing interest in Germany and Europe. The development of custom ICs and ASICs is becoming increasingly important in a wide range of industries, specifically automotive, where Germany is one of those centers of excellence.
I think we are going through the most significant transformation the automotive industry has seen in decades, where we are beginning to think of the car as so much more than an actual mechanical system. It's becoming to be a more software driven, electronic system that has autonomy and can deliver services that can do things that could never be done before.
? What motivates companies today to develop ICs/ASICs themselves?
! Diamantidis: There are several factors that may be motivating companies to develop their own ICs or ASICs themselves. The supply bottleneck for semiconductors is certainly a factor, as many companies have been experiencing delays in obtaining the chips they need for their products. By developing their own ICs or ASICs, companies can have more control over their supply chain. Another reason is the workload-driven, software-defined architectural trend we discussed earlier.
Folks across the global economy are thinking big with AI and are developing novel solutions at an unprecedented pace. Specialized compute, storage, communication, and sensing infrastructural requirements for these solutions drive the need for more IC design. What we want to see is our customers innovating, no matter the system, to help manage production costs and leverage all the technologies, which is very much possible.
? Are there enough skilled workers for IC development in Germany and Europe?
! Diamantidis: Europe has an incredible education system that will allow it to deliver innovators for tomorrow. In the U.S., the semiconductor design industry is on track to face a shortfall of 23,000 highly skills design works by 2030. It's a significant concern for us and our customers about availability of new grads and folks getting into the electrical engineering subject. As for interest in our EDA tools, we certainly see interest from Europe in incorporating our new breed of AI driven design tools into their thinking process and into production.
? In the wake of the COVID pandemic and the war in Ukraine, there has been and continues to be public discussion about rebuilding production capacity in Germany and Europe – and not just in the semiconductor sector. What do you expect in the coming years?
! Diamantidis: I believe the introduction of the CHIPS act, which is focused on increasing investment in semiconductor infrastructure worldwide, is bringing more awareness to the issues surrounding the industry. It’s important that governments are sensitized to this because at the end of the day, the global economy is pursuing AI productivity and enablement – whether it’s in data centers, autonomous vehicles or other applications – so we’ll see a continuous drive and demand for chips. The attention we’re paying to this, which is very overdue, is encouraging.
? Thank you for the interview.