The semiconductor industry is emerging as a crucial enabler of the forthcoming AI-driven era: Dr. Aart de Geus, Chair and CEO of Synopsys
In an exclusive interview with Express Computer, Dr. Aart de Geus, Chair and CEO of Synopsys, delves into the transformative dynamics shaping the semiconductor industry and its pivotal role in the advent of AI-driven technologies. With a focus on India’s burgeoning electronics sector, Dr. Aart sheds light on how Synopsys is driving the industry’s transition from monolithic systems to multi-die configurations. He unveils a compelling vision of systemic complexity through AI systems and big data management, projecting a future that capitalizes on exponential opportunities while adhering to Moore’s Law ambition. Further, Dr. Aart discusses Synopsys’ strategic approach in leveraging its expertise and collaborative efforts to foster the growth of the ecosystem in India, paving the way for innovative software development and positioning India as a global hub for cutting-edge technology.
Some edited excerpts:
Can tell us in brief about the company and its role in India?
We play a substantial role in the semiconductor industry here in India, not only because of our software development but also because of our successful partnerships with many of the multinationals and India based companies. There are many different verticals as well – one can go to manufacturing, or into assembly and testing, which is a very different but relevant area, especially in an age where we’re going to go to multi-die as the whole next generation of complexity.
How do you see the transition specifically, with respect to your company, leading this industry to transition, like you said, from a monolithic system to say multiple systems?
This is a profound transformation and let me try to explain why – semiconductors have fundamentally been governed by this magical law called, Moore’s law, which was essentially an exponential drive towards smaller and more densely packed devices on a single chip. Today we are working with devices that have single, indigent, angstrom size transistors, so it is still being pushed forward and is also clear that the cost is becoming quite high. Hence a question naturally arises as to why we are not moving to multiple chips. When multiple chips are placed on another chip, it is commonly referred to as an interposer substrate. This term is used to describe the intermediary substrate that facilitates the connection and integration of these chips.
The difficulty in achieving connectivity between chips has posed significant challenges in the industry. However, an intriguing development over the past few decades has been the remarkable improvement in chip connectivity by several orders of magnitude. This progress is evident in the ability to accommodate a significantly higher number of pins with greater density and increased speed of connectivity. This has opened up a new horizon, prompting engineers to explore options such as utilizing larger chips or splitting them into multiple interconnected units.
Moreover, the concept of stacking chips vertically, known as multi-die configurations, is gaining traction. Currently, there are more than 100 multi-die designs being pursued by our customers. This movement marks a significant shift from the traditional pursuit of scaling more transistors on a single chip, known as Moore’s Law, towards embracing a paradigm of systemic complexity while still adhering to the ambition of Moore’s Law.
And that systemic complexity comes about in two ways. One is multi-die which is really the embodiment of complexity. And the other is the fact that all the architectures are now designed with the software in mind. So you have more and more specialized chips or systems, such as AI systems or big data management systems and that are all aimed at just a narrow set of tasks in order to get higher speed and efficiency. By combining a functional perspective with a form perspective, a vast realm of exponential opportunities unfolds, projecting the next two decades to witness a significant surge in the application of transistors to process extensive amounts of data.
Consequently, the semiconductor industry emerges as a crucial enabler of the forthcoming AI-driven era and the proliferation of smart technologies in various domains.
In terms of software trends and considering the significance of semiconductor chips, can you tell us how Synopsys is poised to capitalize on its expertise and collaborate with relevant stakeholders to foster the growth of the semiconductor ecosystem in India?
The good news for us is that we have been involved extensively in the entire design stack that enables the ecosystem. Our tools have been used on the most advanced chips since the beginning of our existence and that will continue. We have also proven our capability to optimize transistors all the way down to the atomic level. We have tools that help optimize manufacturing so truly the whole stack is impacted by our tools.
As one of the leading providers in the industry, we hold a prominent position as one of the top two providers of IP blocks. Our extensive portfolio boasts around 7,500 IP blocks, supported by an impressive range of 340 distinct silicon technologies. This particular segment contributes approximately 25 percent of our total revenue, underscoring its significant financial impact.
IP blocks are one of the biggest efficiency and productivity drivers’ period, because when you don’t have to design something, you can take it out of the catalogue, and you can very quickly build very sophisticated architectures that bring you to the next stage which is architectural design. And here we have quite a number of tools as well and then you go to the next level up, which is really the intersection between the hardware and the software. Another exciting area is the role of semiconductors creating digital twins. We are in a prime position to support this advancement using our acceleration machines, and emulation and prototyping technologies.
And so you look at all of this, it’s a stack that goes all the way up to the software and beyond and we have about 10% of Synopsys dedicated to the security that has to build in. It’s worth noting that our approach to security penetrates all the way down through the IP into the silicon, and we have a number of capabilities to build in pre-encryption root of trust and other techniques to increase security.
AI will change completely their destiny, and we are in the midst of helping make this all possible and therefore ultimately, we too, will become more and more specialized with the different vertical markets. So that’s sort of a long story. But we have many puzzle pieces and are very unique in connecting really all these players that are driving the future.
Do you see the emergence of niche players in line with the emerging trends?
My prediction is that when you have such a broad set of opportunities, some existing players will do phenomenally well, and some brand-new players will emerge totally out of nowhere. They know how to assemble some of these pieces to create something for an end market that itself ends up changing the end market. This is where the promise of advanced AI, supported by sufficient computation and data has the potential to bring quite remarkable productivity and efficiency improvements to many fields. And you know, all you have to do is, look at, for example, the transportation field, where, if you can learn from the needs of your customers, the needs of your supply chain, a small change in efficiency has an enormous economic impact. The same can be seen in advanced research for medicine, new chemistry, and so on, because many experiments can be done on a computer instead of in a lab, and then analysed by AI to see what next experiment you want to try.
This is a whole new wave for mankind of what can be done on the computer and when you have high discontinuities, you have also a high degree of new entrance and we’ve seen some of that already, because, I want to say seven or eight years ago a number of the hyperscalers all decided to start investing in designing their own chips. Now the same is happening in automotive. The car companies want to define their own architecture because they understand that it will be critical moving forward.
The Indian Government has made substantial efforts to promote domestic growth in the electronics sector. However, there is a recognized concern regarding the adequacy of the talent pool and skill sets in India to establish a robust ecosystem similar to the existing software services space. Are there any plans or initiatives from your organization to address this challenge through the implementation of skill development training programs or industry-specific training series?
I would not underestimate the amount of talent that is available in India because you have certainly a schooling system that develops outstanding engineers and it is only increasing. This is precisely why I see there’s a lot of opportunity in India because India is also realizing that it has an opportunity to evolve in the global market. Synopsys intends to continue our growth in India, and we’ve grown substantially. We have 5,000 people here today but the growth that I’m really more proud of is not just the number of people, but the quality and the sophistication of the team that we have, which continues to evolve. Which is why I believe, India can absolutely be part of many of these new opportunities that are opening up.
Considering the increasing adoption of cloud technology and the advancements in AI, can India effectively leverage the synergistic potential of these two technologies to propel its competitive edge? In other words, is it possible for India to emerge as the global hub for innovative software development, capitalizing on the unique combination of cloud computing and AI?
Our company operates across various geographical locations, thanks to our long-standing investments in these regions, over many decades. Achieving a consistent level of quality across geographies requires substantial time and effort. As for whether a specific location will emerge as the central hub, I believe that talent exists worldwide in numerous places. One notable advantage of India, for instance, is its strong connectivity with our other hubs scattered across the globe. The teams we have here possess the expertise to effectively collaborate with partners located in diverse locations. This morning, I personally witnessed the seamless collaboration between teams here and teams situated in different parts of the world. The only issue with a different place in the world is the time zone. It’s not distance but time zones, and that never goes away. There’s a shortage of top-notch engineers throughout the world at this point in time and we’ve done well in India – with our facilities both in Bangalore and Hyderabad.
What are your future plans in the next 6 months to one year, or maybe the short-term and the long-term plans for India?
The short-term and long-term plans have always been the same. We are in this field for 37 years and what we do today may have an impact in the next 4 or 5 years. If you don’t invest in it today, you won’t be there in 4 or 5 years. At the same time, it’s one long succession of sprints because every customer is in a hurry understandably.
As far as India is concerned, we’ve also been able to build a strong team that has supported a number of customers at very different places in the world. There is also a collaborative element here in India that works particularly well in difficult technical situations, because our team here demonstrates a high level of technical expertise and perseverance, and truly go above and beyond for the customer. And that really matters in our field where trust and relationships are as important as a technology company.
