Deep Tech for Manufacturing: From Prototype to International Markets. A chat with Sabrina Sasaki Partner at Monozukuri Ventures

yeah and exactly uh my curiosity let's say uh follow the the your conversation in the at the intersection of let's say ai and legacy industries automation robotics and so it depends on the value chain and the supply chain but What do you think are the most exciting areas where today physical AI, let's say, is increasing the excitement of investors, but also the industry? And do you think that the next startups in the field are hybrid startups, hardware and software or software solution fitting in the robotics or, let's say, robotics standalone companies? What do you expect to see in, let's say, the next three, five years in the field?

I'm very biased because out of our 60 portfolio companies, only one is primarily software.

So my expertise, my networks and my experience has been with people who are building hardware and see hardware as a way to scale and collect and provide value from a data space.

point of view with predictions or for maintenance and other areas.

So I would definitely dare to say that in my experience, the hardware startups that are doing primarily only hardware without any acknowledgement or value provided from a software standpoint are already struggling to fundraise and to get their milestones because I think the real problem beauty of the technologies we have now with AI and everything in terms of data and potential ways to predict and prevent things to happen in the physical world is something that it's allowed by hardware, but without a very strong platform on the software side that can actually be useful, not only collecting data.

I think collecting data is no longer the case.

It's not a problem anymore because most companies, I mean, Japanese manufacturers tell me they have plenty of data.

They just don't know how to treat this data, what kind of data they should focus on and how to treat this data to provide some insightful business action steps.

I think that's the main thing.

So you as an investor, one of the first things you look at in a company, let's say at the intersection of AI and legacy industry, is the data set basically.

How do you train that? Because it's one of the most important competitive modes the company can leverage.

I think it all depends.

Of course,

we also invest in advanced materials or semiconductors and some other areas where this might be hard to state as a criteria.

But I think on the robotics side, it's very difficult to invest nowadays in a company that is only thinking about primarily the hardware aspects.

But sometimes their main focus is providing software companies some specific kind of inputs and data and I'll say data points for a software company.

I think that's also valid.

It all comes to what the real value proposition is.

So even if the company is not building from zero, it has to be very aligned with the data approach.

Otherwise, it becomes very easy to be commoditized, as we've seen with the hardware startup revolution and where we are nowadays since the makers movement.

Yeah, of course.

And we will be there in a valuable position in a minute.

The reason just I ask is that sometimes it can happen in the lab before, let's say, the market validation.

We can call it not traction, but almost having the first interaction with B2B.

The founders could ask, which is my, let's say, one single dimension, my unique competitive advantage? I need to focus on the data set, especially to, let's say, the task my robot can perform, or maybe a LiDAR technology, as we have seen in robotics for agriculture, where you can map, let's say, provide autonomous driving.