Karel Masarik has been appointed head of the Czech Semiconductor Center (CSC), which began operations in April this year. He seems to have been predestined for this position due to his extensive experience in the semiconductor industry. In an interview, he shared his experiences from the academic and commercial environments and high-level grant policy, which led to his appointment to this important position. This interview has been published in News at BUT magazine.
When you were deciding on your career path, was microelectronics an obvious choice for you?
My high school studies were already focused on microelectronics, and I continued in this field at the Brno University of Technology, originally at the Faculty of Electrical Engineering. After the faculties were divided and the Faculty of Information Technology was established, I chose computer science because of cyber security and embedded systems. In my final year of engineering studies, I wrote my thesis as part of the Erasmus program, focusing on hardware at the German University of Hagen in collaboration with a Munich-based company. As a doctoral student, I chose a topic on which I collaborated with an Austrian company designing microprocessors for video processing in cars (digital rearview mirrors). In 2005, this was quite revolutionary, and this technology was only available in premium car brands. In my dissertation, I had to design a language for describing microprocessor microarchitectures and methods for automation in the so-called Design Space Exploration process and then implement them in an EDA (Electronic Design Automation) tool.
That sounds like a great start to a business. How long did it take to establish Codasip?
With the help of two grants from the Ministry of Industry and Trade, which were managed by a private company, we incubated the technology to the point where it was possible to attract venture capital for further development and establish a private company. In 2014, Codasip was founded, and I became its CEO. The company offered an EDA tool for automating microprocessor design. In 2017, with the contribution of Codasip, the non-profit organization RISC-V International was founded, which focused on developing the open microprocessor standard RISC-V as an alternative to the ARM monopoly. At that time, Arm was found in all smartphones, car infotainment systems, and consumer electronics. Customers began to ask if we could deliver reconfigurable RISC-V microprocessors as a replacement for the aforementioned Arm processors in addition to our EDA tool. However, we lacked the necessary know-how in Czechia to design these RISC-V microprocessors. I then found microprocessor talent in France, where ARM has large development centers. I convinced German, French, and American investors to invest in another investment round, and we established a German parent company for global expansion. From that moment on, the company’s growth accelerated significantly, and several new development centers were established in other European countries, as well as sales teams from the US to Asia. At the end of 2021, another investment round was closed for the company’s growth, and a new CEO from the US took the helm. At that time, Codasip was experiencing slower innovation and was primarily focused on product stabilization, so in 2022, I moved to the position of president and later director of innovation. At that time, research and development began in areas such as cybersecurity, automotive, AI/ML, and HPC/data centers.
And that’s when Codasip also started collaborating with universities and becoming active in European grants, is that right?
Yes, our university program aimed to train new engineers for advanced microprocessor development in the RISC-V field. In addition, we built very strong teams in Europe for the aforementioned verticals, which, with the help of
European grants, were able to deliver highly innovative solutions to companies such as Bosch, Infineon, and NXP. On grant projects, we collaborated extensively with universities such as Cambridge, Lund, and the Barcelona
Supercomputer Center. Three years ago, market needs and ARM’s actions indicated that more powerful microprocessors were needed, mainly due to the rapid emergence of AI/ML and Nvidia’s very close cooperation with ARM. We focused on preparing a project for a European chip for supercomputing centers and AI Gigafactories, which, after two years of preparation, was launched in March this year. In this project, Codasip became the main supplier of the general computing part, with the aim of developing a RISC-V chip using 4nm TSMC technology. Codasip was a founding member of RISC-V International, so I contributed to the creation of the community as a member of the organization’s technical steering committee and later as a member of its board of directors. My role was also somewhat political, as during the aforementioned projects I had to actively communicate with, for example, the European Commission and the German and Czech governments regarding finances. The European Innovation Council also participated in several investment rounds in Codasip, where I, as the founder, led negotiations with the commission on investments.
Please specify how the need for Europe’s independence from foreign semiconductor markets arose.
Europe began to feel this need due to geopolitical changes. The COVID-19 pandemic disrupted supply chains, which was far from limited to medicines and medical devices; semiconductor chips were sorely lacking in the automotive industry. This was followed by increased geopolitical tension in the world, with all regions of the world wanting to secure reliable chip supplies, so Europe also began to strive for diversification of production and certain capacities of its own. In recent years, thanks in part to the defense sector, it has become clear that we cannot do without a renaissance of European industry, especially the semiconductor industry. I think that, just as in the US or Taiwan, the time has come in Czechia when engineering will be considered by society to be very prestigious and beneficial for the good of the country, whether from the perspective of its positive impact on the country’s GDP or its strategic added value.
What is the history of the semiconductor industry in our country?
It depends on how far back in history you want to go. There was Tesla Rožnov, the first computers were created in Slušovice, even though it was an agricultural cooperative. Czechoslovakia and subsequently Czechia have always been strong in mathematics, physics, chemistry, and other disciplines that are a prerequisite for chip design and manufacturing. Czechia therefore has a very good engineering base with a very long tradition, which has been strengthened over the long term, in particular by the Brno University of Technology, the Czech Technical University in Prague, and other leading universities that focus on technical and scientific education in the Czech Republic.
How did the preparatory phase proceed before the opening of the CSC?
A few years ago, I personally participated with other European companies in the preparation of the European Chips Act, when there were extensive discussions among member countries about the concept. I think the result is a solution that will serve to distribute the allocated money transparently. The US and China have their own versions of the Chips Act – their budgets exceed the European one several times over, but at the same time, they do not define such clear rules for the distribution of funds. Competence centers in individual member states are therefore one of the main pillars of the Chips Act. Partners who were able to offer complementary services attractive to startups and small and medium-sized enterprises became members of the CSC. According to the center’s definition, it is precisely these companies that need to be targeted, not only in the given member state, but also abroad. The Czech National Semiconductor Cluster, which brings together over 44 semiconductor technology partners, is also a member of the CSC, and each new company can thus indirectly participate in the center’s activities when it joins the cluster.
What is the main mission of the CSC?
As I mentioned, semiconductor competence centers are part of the European Chip Act, the implementation of which is also enshrined in the Czech Republic’s National Semiconductor Strategy. For the first four years, funding for the centers’ operations will come from both national and European sources. This will significantly reduce the barrier to purchasing services for small and medium-sized enterprises through large discounts on the services offered, thereby greatly simplifying small companies’ access to highly expensive technologies. We firmly believe that this will result in a significant acceleration in the development of highly innovative chips. The aim is to accelerate the emergence of young innovative companies in the semiconductor industry in particular. The CSC will strive to
become self-sufficient in the long term.
Why did BUT find itself in the role of coordinator of the preparation and headquarters of CSC?
Brno, and by extension the South Moravian Region, is rich in technology companies that contribute to the global
semiconductor ecosystem. Electron microscopy is one example. The South Moravian Region also cooperates very intensively with other regions, such as the Zlín and Central Bohemian Regions, which are home to other successful semiconductor companies. The South Moravian Region is also home to the JIC innovation agency, which has achieved very good results in supporting the creation of new start-ups, which are the main target of the CSC. The largest technical universities in Czechia, VUT and ČVUT, are a guarantee of highly qualified talent for the semiconductor industry. Brno was therefore a logical choice for the establishment of the center. In addition, VUT Rector Ladislav Janíček acted as a unifying element for the consortium partners from the very beginning, helping to solve the problems we encountered in forming the consortium on a daily basis.
How will the CSC’s activities be reflected in teaching at BUT?
When the European Chip Act was being drafted, there was much discussion about the European ecosystem’s heavy dependence on EDA tools and IP from the US. This is a “fabless” (a product that does not require factories to develop, only engineers, ed.) situation, where more than 90 percent of these technologies are purchased from locations outside Europe. Therefore, I think that study programs expanded to include megatrends in industry, such as IP design or chip design for AI/ML, are the right direction. In terms of teaching, it can be expanded to include the basic building blocks of European chip law for inspiration. One of the building blocks, for example, is the aforementioned open microprocessor standard RISC-V, which is something like a cookbook. It tells you how to cook soup, or what the processor should look like, but the actual production of the product is up to you. There are a number of commercial and academic suppliers of RISC-V solutions, and any courses focused on use, programming, or design will be a huge benefit to the European ecosystem in the form of graduates with high added value for future employers. This is because all large companies, including the European Commission, already have financially supported plans to transition from Arm to RISC-V within a few years. This is just one of many similar examples. The company onsemi will build more factories for new technologies in Rožnov, and engineers who know how to work with these technologies will be needed.
How long will it take Europe to become independent of the Asian and American markets?
Let me use the example of the European chip for supercomputing centers and AI Gigafactories. Originally, the EU talked about wanting to replace Intel and Nvidia. However, for such an ambitious plan, the one billion euros that the EU plans to allocate represents a budget for these companies for a few months at most. As a result, this is more of a Plan B for the European Union. But let’s be glad that after two years of preparation, something has finally gotten off the ground. It was a very difficult task to bring together the interests and priorities of 27 member states. It’s like building Airbus: once you reach an agreement within Europe, you have a stable concept in which Europe can invest in the long term to become completely self-sufficient. I could find more similar examples, for example in the defense industry.
What can you expect in your position in the near future?
First and foremost, we need to complete the preparatory phases. The services provided by the consortium members need to be packaged and offered in such a way that they are attractive to clients and the center becomes a lighthouse for them to follow.
Source: Jana Novotna (News at BUT Magazine)
Photo: Václav Koníček
