Which current developments are the biggest driving forces for research into chip design?
»One topic that appears in almost all areas of chip design is the use of artificial intelligence in electronic design automation (EDA). Today, AI primarily helps to analyze complex design spaces more efficiently or to resolve optimization tasks more quickly. It’s important to emphasize here that AI currently acts as a useful addition to classical EDA methods, rather than replacing them. At the same time, we’re seeing a fundamental change in system architecture. Chiplet concepts and advanced packaging are becoming increasingly significant as performance improvements can no longer be achieved through smaller structure sizes alone. Instead, different functional units are combined in a modular fashion. This paves the way for new possibilities – but equally presents new challenges for developers.«

Are there any particularly dynamic areas of research at the moment?
»There are some really exciting developments in terms of the automation of analog and high-frequency design. For a long time, these areas were thought to be hard to automate because they rely heavily on experience and expert knowledge. Today, we see that data-driven methods can increasingly help with tasks such as parameter optimization or layout decisions. At the same time, there’s a growing focus on verification. New methods of post-layout analysis allow potential problems to be identified sooner and expensive development iterations to be significantly reduced. Although many of these approaches are still transitioning from research to industrial practice, they have enormous potential.«

What is currently proving to be the biggest challenge for the community?
»The biggest challenge right now is the steadily growing complexity of modern integrated systems. Today, digital, analog, and high-frequency components are combined within highly integrated systems. This means that the requirements of the simulation, verification, and the overall development process increase considerably. We’re therefore seeing a significant expansion of major EDA tools toward the system level. At the same time, the question is how innovative research results can be transferred into industrial design flows as reliably as possible. Many academic approaches show impressive results but need further development as regards robustness, scalability, and tool support before they can be used widely. This is in addition to an aspect that is often underestimated – namely, the growing global demand for qualified specialists. Particularly in the areas of EDA and semiconductor development, well-trained young talent will always be a decisive factor in the technology’s success.«

What made SMACD 2026 a particular success in your view?
»The community! It was so inspiring to see the huge level of interest in the different topics and the many conversations taking place. The conference clearly showed the direction of travel for our field. AI-based methods featured in many contributions – sometimes as a replacement for existing process steps and sometimes as an extension to established design tools. Research is currently dominated by precisely these hybrid approaches. Other standout topics at the conference included chiplets, advanced packaging, and hardware security. The fact that these particular topics were the subject of such intensive discussion at a scientific conference shows how closely research and industrial requirements are intertwined. I was also especially pleased to see how strongly young scientists were represented. They contributed many of the most innovative methodological approaches – a key indicator for the future of our field.«

How will chip design develop over the coming years?
»I expect we’ll think more in terms of overall systems in the future. The focus is shifting from the individual chip to complex architectures that require a holistic view of chiplets, packaging, interconnects, and heterogeneous integration. In parallel, AI will become a commonplace tool in the development process – not as a replacement for engineering expertise, but rather to provide support wherever there is currently a need for complex, error-prone, and repetitive tasks. These methods will increasingly help to analyze large volumes of data and make complex optimization decisions. It is precisely this combination of established design methods, data-driven procedures, and a system-oriented perspective – delivering real added value for users – that I think will be the direction of travel in chip design over the coming years.«