Building a Future Beyond Boundaries with Honda and Cadence

We at Cadence are proud to be a long-term partner of Honda R&D (HGRX), and our collaboration with Honda's research efforts emphasizes our commitment to advancing electronic design automation (EDA). We are honored to have Dr. Anirudh Devgan, our President and CEO, engage in an insightful discussion with HGRX's Director, Atsushi Ogawa, and Yoshikatsu Shiwa, who leads their semiconductor initiatives, about the future of AI and semiconductors.

Join us as we explore the question: "How will AI take root in the physical world?" through this comprehensive, in-depth conversation.

Co-Creation Power Born from Mutual Learning

Shiwa: Welcome to Honda's Innovative Research Excellence, Anirudh-san. Honda and Cadence have had a long-standing relationship, and we collaborate across a wide range of areas. Today, I'd like to hear your insights on semiconductors and AI technologies.

From left: Yoshikatsu Shiwa, Atsushi Ogawa, and Dr. Anirudh Devgan

Dr. Devgan: Honda is certainly one of our customers, but the relationship goes deeper—we're also learning from each other. It's an honor to conduct joint research with a company that operates real-world hardware and systems through mobility.

Ogawa: I feel that Cadence and Honda engineers share a similar spirit. That might be why we resonate and create good research together. Thank you for joining us today.

Shiwa: While AI is creating tremendous value, it also poses challenges such as high power consumption. What do you think is the most important technological breakthrough needed to overcome these issues?

Dr. Devgan: AI fundamentally relies on dense matrix operations and linear algebra. When I entered the EDA field back in the 1970s, the algorithms were quite similar to what we see in AI today. They were dense, inefficient, and computationally heavy. But through optimization, hierarchical design, partitioning, and latency handling, we achieved more than 100X speed improvements.

Dr. Devgan: I believe a similar evolution will happen in AI software. Over the next 5 to 10 years, we'll see increasingly efficient architecture and optimization techniques, and methods like multi-head latent attention (MLA) and mixture of experts (MoE) will become standard.

Ogawa: Honda has been working on AI and machine learning for more than 20 years. We've obtained numerous patents and began exploring 3D chips around the same time.

Dr. Devgan: What's important isn't just software evolution. We also need hardware innovation—AI-specific architectures, physical integration through 3D-ICs and chiplets. For example, if each area—software, hardware, and packaging—improves by 10X, we can achieve an overall 1000X performance gain.

Ogawa: As AI chip performance grows exponentially, so too does the value that vehicles and robots equipped with those chips can provide to society. Our hardware is evolving at an exponential pace as well. Believing in that growth and continuing to push forward is essential to our development efforts. The challenge, however, is power consumption—which has a huge impact on mobility.

The Key: Merging AI with the Physical World

Shiwa: To tackle that issue, it's no longer sufficient to separate hardware and software—they must be designed in harmony. What do you see as the key to that?

Dr. Devgan: That's exactly why I find our R&D with Honda so exciting! I believe the key lies in "physical AI." Until now, AI has mainly been about software—language models, for instance. But in the future, AI will integrate with physical phenomena in the real world, becoming part of vehicles, robots, drones, and agricultural machines. That's where true value will be created.

Ogawa: Absolutely. For example, in automated driving, the "last mile" or, for robots, the "last millimeter"—these are domains where simulations alone aren't enough. You need to deal with unpredictable situations, grasp soft objects, or handle deformable shapes. These challenges don't occur in fixed patterns. You can only solve them by actually experimenting in the real world. We call this "physical AI" or "embodied AI"—AI that collaborates with real-world phenomena.

Dr. Devgan: This domain consists of three pillars: "Silicon (semiconductors)," "systems (hardware, software, mechanisms)," and on top of that, "data." Only by integrating these can we create real-world value.

Ogawa: We've delivered results in the system domain—cars and robots, representing the real-world physical systems. That's why our collaboration with Cadence allows us to bring out the best in each other.

Dr. Devgan: Cadence started in the silicon domain, but through partnerships with companies like Honda, we've come to deeply understand constraints and bottlenecks in mechanical and electrical domains, as well.

Ogawa: That's why, in this kind of collaboration, it's not about dividing tasks, it's about deeply understanding one another. Rather than saying, "you take it from here," we need a mindset of crossing boundaries and truly understanding each other's specialties.

Engineering Mindset Solves Any Challenge

Dr. Devgan: What we enjoy most about working with Honda is the direct dialogue and deep collaboration between engineers. Real breakthroughs come from engineers speaking candidly and working hands-on together. That's where the real magic happens. Honda has a rich legacy of research and development, and the teams have an incredibly strong engineering mindset. That's why this partnership is so valuable to us at Cadence.

Ogawa: I always tell young people, "Don't aim to be a generalist first—aim to be a specialist." Dive deep into one domain first. Once you've dug a deep vertical hole, it's easier to expand it horizontally later. Shiwa-san is a great example. He started as an automotive engine engineer, then transitioned to semiconductors a few years ago. Now, he's able to have deep technical discussions with your engineers at Cadence.

Dr. Devgan: When you drill deep into one domain, you build the strength to apply your skills to others as well. I believe the value of specialists will only increase, while generalists will see their relative value decline. People with vertical expertise in multiple domains are exactly who we'll need in the future. In fact, there are many fields that are harder than chip design. But if you have an engineering mindset, you can face any challenge. That's the true strength of being an engineer.

Driving Transformation and Shaping the Next Decade of Mobility

Shiwa: When we pioneer the future like this—what lies ahead? What do you think mobility will look like 10 years from now?

Dr. Devgan: I believe it will be completely transformed compared to today. That's exactly why I'm here. The next decade will be the golden age of AI. Especially, physical AI and embedded AI will become central to all system products and mobility technologies. And the same is true for semiconductor products. This is the most exciting time, right at the start of that transformation. As AI, semiconductors, and mechanical systems merge, an entirely new world will emerge.

Ogawa: In the automotive industry, we often talk about a "once-in-a-century transformation." But I see that as a great opportunity. That's because players from other industries entering automotive are accelerating the evolution of functionality and performance more than ever before. It's great for customers too, and it's a big advantage when co-designing hardware and software. Of course, there's a risk that software or AI companies will take the lead. But we have strengths in engineering and a culture of collaboration with companies like Cadence. That's the true value of Honda.

Dr. Devgan: Over the past 10 years, we've helped drive transformation with smartphone manufacturers. And in the past five years, we've worked with the leading hyperscalers. And I'm confident the next big wave is in cars and robotics. The next five years will bring huge opportunities. That's why working with Honda is truly exciting.

Ogawa: Honda isn't just about cars. We can help drive transformation across entire industries, including robotics. This is a massive opportunity, not only for Honda but for the entire industry. I believe we are now at the center of that transformation—together.

Dr. Devgan: Absolutely. We are one team, Honda and Cadence, moving together as one.

Shiwa: "One team"—I love that. Semiconductors are like the engine that drives AI. They're a technology that makes a major impact on society and our customers. Let's keep working together!

Learn more about Cadence Automotive Solutions.