Outside Director/UTokyo Professor × Group CTO: Panasonic’s New Value Creation Through Physical AI

AI evolution is accelerating. In fact, the AI technology domain is breaking free of software. The era of “physical AI” is just around the corner, where AI creates value by connecting sensors, devices, spaces, and human behavior. What pathways does this trend open for the Panasonic Group? How can our accumulated manufacturing expertise and highly diverse businesses serve as a competitive asset in the physical AI space?

Yutaka Matsuo is a professor at The University of Tokyo Graduate School of Engineering. Since June 2025, he has been an outside director of Panasonic Holdings Corporation (PHD). Tatsuo Ogawa is an Executive Officer and Group CTO, and Hiroshi Saijo leads Panasonic Group’s DX/CPS Division. They came together to discuss the Panasonic Group’s strengths in the AI era and the future direction of value creation.

What made you decide to come aboard as an outside director, and what are your impressions of the Panasonic Group so far?

Matsuo: When I was approached for the role, I was eager to accept. Panasonic products have always been familiar to me, and it’s a brand I personally like. I think the Panasonic Group has tremendous potential in the AI era. AI is not confined to software alone. Ultimately, it will spread through society, connecting a very wide range of hardware and real-world environments. From that perspective, the technologies and business assets that the Panasonic Group has cultivated over decades have significant potential to create unique value that competitors can’t match.

Ogawa: When AI was just taking off, Professor Matsuo was already offering valuable insights. Since then we’ve collaborated on multiple initiatives, including Konosuke Matsushita AI reconstruction project*, developed jointly with Matsuo Institute, Inc. We’ve also collaborated on talent development and discussion forums. Now, as we consider what it is that we should accomplish in the coming era by leveraging AI, having him as an outside director is very reassuring. In business as well as development, how should we move forward? His perspectives as a leading expert on these questions will be extremely valuable.

* In 2024, Panasonic HD and Matsuo Institute, Inc. developed an AI inspired by the philosophy and writings of Konosuke Matsushita, the founder of Panasonic. The AI is not intended to replicate or represent Konosuke Matsushita himself, but rather to convey his management philosophy and way of thinking through modern AI technology.

Now, with an inside perspective on the organization, how do you see the Panasonic Group’s strengths? What about future growth areas?

Matsuo: I think the Panasonic Group’s strengths, and areas where there is room for further growth, are quite clear. First, I see a very strong commitment to technology and quality. At the same time, if sales and marketing data and customer feedback can be cycled seamlessly into product development and business decision-making, I see significant potential to become even stronger. The Panasonic Group’s range of businesses is a source of diverse information. If this information can be leveraged responsibly through a shared foundation, it can boost competitiveness across all business domains.

Where do you think it’s most important to begin when considering potential for growth?

Matsuo: It is important that sales and marketing data be fed back more actively into design and development at an earlier stage. Traditionally, sales networks and frontline sales teams were tasked with understanding customer issues and desires. Today the industry structure itself has changed, but with the power of AI, we should be able to interpret that information in new ways. Once this feedback cycle is taking place on a company scale, the Panasonic Group will become even stronger.

One of the key themes of this discussion is physical AI. How should we understand it?

Matsuo: I think the easiest way to understand physical AI is as a domain where AI creates value by connecting sensors, devices, spaces, and human behavior. Rather than just offer users information on a screen, physical AI registers real-world conditions, makes judgments, and changes the behavior of devices and services. This will become a major arena for competition.

Ogawa: The Panasonic Group has devoted itself to manufacturing hardware, but in the future it will no longer be sufficient to think simply of hardware as hardware and software as software. The point will be how we design and connect them as a single, unified, AI-driven customer experience. By leveraging AI and data, we can take value that has been refined in our frontline operations and business units, and elevate it into premium customer value and a premium customer experience. I think this is precisely where we should focus our physical AI efforts.

Why does the Panasonic Group have opportunities in this arena?

Matsuo: Since 2019, I’ve been involved in the National Colleges of Technology Deep Learning Contest (DCON)*, and I’ve seen firsthand how combining AI with manufacturing can spark astonishing ideas, even from students. At the same time, when you look at the professional world, there are surprisingly few companies that can create entirely new products from scratch, even if they excel at refining existing ones. In that sense, I think the Panasonic Group is a company that, while highly professional, also possesses the ability to generate new kinds of value. That’s precisely why it has the potential to make substantive advances into emerging domains such as physical AI. Moreover, quality and safety are ultimate requirements, and the Panasonic Group’s ability to implement solutions through to the social deployment stage, including verification of quality and safety, is a major strength.

* DCON is a competition for students from Japan’s National Colleges of Technology, higher-education institutions known for their practical engineering education. In the contest, students apply deep learning to develop product and business ideas for real-world use.

What should we be thinking about when translating that potential into actual product development?

Matsuo: In the era of physical AI, it’s important not to rely too heavily on existing product forms. Take refrigerators. Rather than simply making incremental improvements to the function of cooling food, we should reimagine the product, starting from questions such as what consumers truly want and how we can support different lifestyle values within each family. I like to think of this as an “exploration” mindset that searches broadly for possible answers. People live with diverse needs and desires. These include things that would be nice to have as well as things that could be improved. Moreover, the range of possible solutions changes rapidly with the times. So it’s essential to first explore a wide field of potential answers, identify the most promising areas, and pursue solutions from there.

Another important point is that while the engineering and scientific rigor applied to manufacturing products is extremely thorough, discussions about marketing tend to rely too heavily on intuition and past experience regarding customer needs. In fact, it should be possible to explore the kinds of value that customers are looking for from a broad perspective, supported by a scientific and engineering mindset. Adopting a broad perspective is essential to creating future products and services.

Ogawa: I thought we were doing some degree of exploration, but hearing your comments makes me realize that we often end up searching for answers only within boundaries we set from the start. If we broaden our perspective even slightly, we might find a solution somewhere completely different. I also think we may not yet be fully leveraging AI and data to deeply understand the living spaces and work environments our businesses serve. The places we must return to are, after all, the real settings of daily life and work. Those are where our long-standing commitment to sincere craftsmanship has been developed, and where the sources of enduring value are to be found. We should start from real-world contexts. From there, the key will be how deeply we can enhance our understanding of those contexts through AI and data.

While thinking broadly, how should we increase the speed of development?

Matsuo: It’s actually true that the broader the scope of your inquiry, the more important it is to run your initial hypothesis-testing cycle quickly, to draw provisional conclusions in a short timeframe. If that doesn’t lead somewhere useful, you shift your focus of exploration and try again. Repeating this enables you to identify promising areas at an early stage. Today’s tools and libraries are well developed, which allows exploration at a far faster pace than before. In fact, speed of exploration itself is becoming a competitive asset.

At that point, how should we think about the key data and sensing?

Matsuo: What we choose to capture from the real world becomes extremely important in physical AI. Depending on the task, we have to determine which information is truly necessary and what can be discarded. Hidden within unverbalized details, such as minor household habits, housekeeping preferences, and unspoken norms inherent to different workplaces, there are often new needs waiting to be uncovered. It’s fine to start by asking, “What can we do with the data we currently have?” But as you clarify your goal, you start to see where the data is insufficient. So you go out and collect new data. This iterative process is the key, moving back and forth between purpose and data, experimenting, refining, and moving gradually toward a better solution.

Ogawa: What you are saying overlaps with what we’re trying to do. For example, in scenes inside a home, we find a wealth of contextual details unique to that household, yet most of them have never been verbalized or captured as data. When we can detect those cues and align what people want next with the functions of physical AI, their reaction will be, “That’s exactly what I wanted.” Such alignment should result in genuine utility.

Matsuo: This doesn’t mean we have to suddenly change our existing practices to something completely different. What we have treated as fixed assumptions and constraints may in fact not be constraints at all, but variables. If we can see it that way, we can decide how to broaden the scope of our exploration. I think that will make it easier to leverage the strengths that the Panasonic Group has cultivated over so many years.

As AI continues to evolve, how will the role of humans change?

Matsuo: AI is extremely strong when it comes to processing massive amounts of data and capturing overall trends. But when it comes to that final moment when we sense that something is good, or interesting, or likely to be supported in the near-future market, human sensibility plays a major role. AI supports the upstream analysis and insights, and humans make the final value judgment. I believe that combination will continue to be important for the time being.

Ogawa: It’s very thought-provoking to hear a leading expert in AI say that human sensibility is crucial in the end. The idea that the final judgment relies on uniquely human intuition, what we might call “gut feeling,” is encouraging from a human perspective. At the same time, I think we face challenges in gathering and discerning valuable data from our direct points of contact with customers before we reach that final judgment.

What about human resource development and collaboration with startups?

Ogawa: In the PHD Technology Sector, we have established our Technology Future Vision, which outlines the ideal form of society in 2040 and the direction of R&D required to achieve it. Based on this vision, we are strategically strengthening our technologies and promoting co-creation with partners. At Technology CUBE, our new core research and development hub that began full-scale operation in Osaka in April, we plan to pursue open innovation and co-creation with external partners even more actively than before. Large companies and startups have different roles to play. Large companies should tackle initiatives requiring long-term commitment, while startups excel at rapidly exploring challenges. Since many talented individuals and startups gather around Professor Matsuo, we hope you will advise us on collaboration partners with whom we can generate new value.

Matsuo: In AI and data science, it’s essential to learn by actually going through cycles of trial and error. As more people interact with and study AI, they naturally develop a sense for when a hypothesis needs to be revised and how to speed up the exploration process. Startups have the advantage of being able to take risks and venture into new areas more easily. Large companies, on the other hand, have the strength to carry solutions all the way through to social implementation while ensuring quality and safety. I believe it’s best for each to make use of their respective strengths and collaborate through a clear division of roles.

Finally, tell us about future initiatives and what expectations you have for the Panasonic Group.

Ogawa: We have launched Panasonic Go, an initiative aimed at making roughly 30% of our total group revenue AI-driven by 2035. At present, various trials are progressing simultaneously, but it is essential that we not let them remain isolated efforts. Instead, we must ensure they are reliably connected to strengthening the capabilities of the entire group. How should we build our AI infrastructure, how should we establish direct touchpoints with customers, and how should we develop a research network for physical AI? Through these efforts, we hope the Panasonic Group will continue to be a company that contributes to society through AI.

Matsuo: I see great potential in the Panasonic Group’s future. After becoming an outside director, I visited the Konosuke Matsushita Museum, and I sensed the founder’s strong desire to enrich people’s lives and continuously pursue exploration toward that goal.

Now we are in the era of AI, and the scope of that exploration has expanded far more than before. The essence of deeply understanding consumers and everyday people, discovering what truly delights them, and delivering something good does not change, even as the times change. The Panasonic Group has the potential to realize that essence in a new form. As an outside director, I hope to actively contribute to making that a reality and helping to shape it with you.