Soft robots are solving hard problems in Japan

Slithering robots sounds vaguely horrifying, but it turns out they’re incredibly useful.

Soft-robots lack the charisma of their more acrobatically-inclined rigid counterparts, but some startups are betting on the unique role they have to play.

Today we sit down with Toru Ichihashi, the CEO Solaris, and we talk about what it takes to turn a university research project into a viable startup, what happens when Japanese VCs decide it’s time for a startup to change CEOs, and the most likely path for soft robots to slither onto the world stage.

It’s a great conversation, and I think you’ll enjoy it.

Show Notes

• The creeping problem with pipes

• How to stuff a robot into an industrial pipe

• What’s driving the early adopters of soft robotics

• Turning a university spin-out into a business-focused startup

• When do Japanese VCs decide its time to change CEOs

• The importance of both funding and marketing partnerships in Japan

• Can partnerships help Japanese startups go global?

• Why we don’t see more soft-robotics in the real world

• Why Japanese investors focus more on risk than opportunity

Transcript

Welcome to Disrupting Japan, Straight Talk from Japan’s most innovative founders and VCs.

I’m Tim Romero, and thanks for joining me.

We’ve been talking a lot about robotics on Disrupting Japan recently, and frankly, that’s because there’s been a lot going on in robotics in Japan recently.

Today, we’re going to be focusing on soft robotics. And by soft robotics, I don’t mean that they’re cute and fuzzy, but that they are non-rigid. Soft robots are robots that can bend and curve and even slither.

So today, we sit down with Toru Ichihashi, CEO of Solaris, the makers of the soft robot Soha, which crawls through industrial pipes to inspect and to clean them. It’s a natural, real-world use case for soft robots. But of course, things are not quite so simple. Even in Japan, bringing a robot to market is a lot more challenging than you might think.

Now, during the podcast, both Toru and I make truly heroic efforts to try to describe how this robot actually moves and manages to crawl through these small pipes, but you really have to see it to understand. There’s a link to the videos up on the site of Soha crawling through transparent pipes, so check those out.

Toru and I talk about the real challenges of establishing a new market in Japan, why sales partnerships are much more important in Japan than they are elsewhere, and what happens when startups investors decide that it’s time for a new CEO. But you know, Toru tells that story much better than I can. So, let’s get right to the interview.

Interview

Tim: So, we’re sitting here with Toru Ichihashi, the CEO of Solaris, who’s making the Soha robot, a small-diameter pipe inspection and cleaning robot. And thanks for sitting down with me. I really appreciate it.

Toru: Yeah, I’m the same here, so I really appreciate this opportunity.

Tim: Yeah, in your brand new office too.

Toru: Yeah, yeah, just a week.

Tim: So, tell me a bit about Soha. What is it exactly?

Toru: Okay, so do you have a camera or not?

Tim: No, yeah, so for our listeners, it is sitting right in front of us, but this is an audio podcast, so you’ll have to explain.

Toru: I see. Our target is narrow pipe. Especially diameter is 150 millimeters or less. Just think about social infrastructures. So, many narrow pipes everywhere.

Tim: So, is it mostly like water pipes or pipes carrying like chemicals and exhausts in factories? Just all kinds of things.

Toru: Yeah, all kinds of. So, we could say this is a kind of veins for social infrastructures. However, these narrow pipes are very small. That is why humans cannot access. Conventional robots also cannot access. So, let’s take an example of a factory. When they need to inspect or clean, they stop manufacturing equipment and they...

Tim: They like disassemble everything.

Toru: Yeah, disassemble pipes. Then by brush, they clean manually.

Tim: Okay, so they take it apart and clean section by section.

Toru: Yeah, exactly.

Tim: And so what does Soha do?

Toru: Our robot is -- ket me show you how to move.

Tim: Well, it’s coming into the same problem of being an audio podcast.

Toru: Ah, yeah, that’s not good.

Tim: But no, no, I get it. I mean, you mentioned before that it was inspired by the way an earthworm moves. So, kind of like sticking out its head and pulling up its tail and sticking out its head and pulling back its tail again, right?

Toru: Yeah. So, to move into narrow pipe, conventional robots cannot move. For example, caterpillar, such kind of robot cannot move. People think, okay, that’s mimic snake or inchworm.

Tim: Because it can’t bend in the pipe, right? It can’t move up and down.

Toru: Yeah, you’re right.So, Snake move by winding, so narrow pipe, snake cannot move. Only earthworm can move into narrow pipe because earthworm moves, expand and moving this way.

Tim: Pushing, like head pushing forward, pulls up behind, push the head, pull behind.

Toru: So, our robot is mimic earthworm movement.

Tim: In practice, like how do you get it into the pipe? Can it go in through like an open valve or something, or do they still have to disassemble?

Toru: That’s a very good question. Such kind of easy operation is very important. So, we have station for the robot to start up. Then this station can be connected with customer’s pipe.

Tim: Okay. For example, you take off a section of the pipe and the robot can crawl in.

Toru: Yeah, yeah, yeah, yeah.

Tim: Okay. So you don’t have to disassemble the entire thing?

Toru: No. In case of factory pipes, some cases at the end, open area is there, and also some connection is there between pipe and pipe. So, just disassemble connection only, then connect our starting stations, then our robot can move into the customer’s pipe.

Tim: That’s cool. And how far can it go?

Toru: Currently, 20 meters. So in case of factories, 20 meters is enough.

Tim: Yeah, I can see that. It’s actually quite far. I mean, there’ll be some kind of break every 20 meters. And since it’s an earthworm, it can even climb straight up and straight down as well, right?

Toru: Yeah, yeah. Vertical climbing, no problem. And also these kinds of complicated curves, also we have no problem to move.

Tim: Right. So, it can move through very sharp turns and straight up, straight down.

Toru: Yeah, exactly.

Tim: Well, what I think is also really astounding about this is that it’s pneumatic. It’s all powered by air.

Toru: Exactly. No electric, only air.

Tim: I mean, I can definitely see the advantage in that it would let you go into explosive environments. Was that the main reason for going with air or there are other engineering concerns?

Toru: Our main purpose is how to mimic earthworm movement. Air is perfect. Let me explain how it works. So, this is artificial muscle.

Tim: So, each segment of this earthworm has a kind of an air bladder, an air segment to it.

Toru: Yeah. Then let me -- so, this artificial muscle...

Tim: We’re going to have to turn this off because you won’t be able to hear me out there.

Toru: So, you can see once air is inflated, this artificial muscle expands this way only. And this way contracts.

Tim: Okay. So, each segment would balloon out. And I guess when it’s ballooning out, it’s kind of gripping the side of the pipe. Okay. And then pulling everything up behind it. So, it’s not an autonomous robot. It’s controlled by the person...

Toru: Controller. Yeah. This is very simple. Move forward, stop, move back. So this robot, inspection and cleaning, we can do simultaneously.

Tim: So, tell me about your customers. Who’s mainly using this robot?

Toru: Our current main target is semiconductor factory. Because semiconductor factories manufacturing process by narrow pipes, such kind of dust, they push it out.

Tim: Okay. So, just the extreme demands of semiconductor manufacturing means the pipes have to be very clean.

Toru: Yeah, yeah. Once this pipe is sealed with dust, manufacturing line is stopped. So, this is a very critical maintenance process.

Tim: I can imagine. But it would seem like these kind of pipe inspections are needed, I mean, across all kinds of manufacturing.

Toru: Yeah, of course.

Tim: Do you have customers in more traditional manufacturing as well?

Toru: Yeah, yeah, exactly. Why I said semiconductor is the first target is because once manufacturing equipment is stopped, opportunity loss is huge. Currently, they are doing manually. They need to stop manufacturing equipment for a day or two days. But by using our robot, just one hour or two hours stop is enough.

Tim: Okay. So, their cost saving is not so much the cost of the cleaning itself, but of the opportunity costs and shut down the line. But again, I imagine that’s true in a lot of industries.

Toru: Yeah, you are right. So, also chemical industries, car manufacturers, all kinds of factories, they have same kind of issues. So, currently our deals, which we are working with, 70 or 80% are semiconductor manufacturers. The rest is chemical factory or car industries. As you said, other industries also cleaning pipes is very important for them.

Tim: Okay, but perhaps just not quite as urgent. So, they’ll take a little longer.

Toru: Yeah, that is why I believe we need to make a business model by semiconductor manufacturers. Then we expand our business.

Tim: That makes sense. I want to get back to the business model in a minute. But before that, I want to talk a bit about you. So, you’ve actually been in robotics for quite some time now. You were at Melton and ZMP.

Toru: So, my background is I spent about 30 years at Toshiba. And the first decade, I was software engineer. And the second decade, I did a global business. And seven years, I was in the United States to start new business. So, I was first headcount for this business, start hiring people or build our equipment. But finally, after seven years, Tokyo headquarter asked me to shut down this business. So, I think this is basic experiment for me to manage a startup business. Then last decade at Toshiba, I was existing leadership. So, general manager for global business, such kind of. After that, I dive into the robotics startup. First, ZMP, main achievement at this company is to significant revenue expansions.

Tim: Right. And as you mentioned, you’re not the original founder of Solaris. It was founded back in 2017 by professors Nakamura and Yamada-san at Chuo University. And actually, at first, they weren’t really focused on pipes at all, right? It looks like it took a while to find a product market fit for this.

Toru: Yeah. When Professor Nakamura and Yamada started Solaris, many products they were trying. Because university laboratory doing many projects.

Tim: What kind of things do they experiment with?

Toru: Basically, professor Nakamura is researching soft robotics. So, like earthworm or snake, such kind of soft robotics. However, as a business point of view, not research, but real social problem. Our product should solve that problem. And also we are small startups, so cash is not enough. So, we need to focus the product which has best opportunities. Then last May, I have been appointed as CEO. And also at that point, executive team, I also all changed.

Tim: So, what was the trigger? So, as you mentioned, you joined fairly recently, last year, about a year ago. What was the trigger for the investors to bring you on board?

Toru: I don’t know exactly that story. But our Solaris main venture capital is Antelope and Spaxks at Meltin MMI. Same two venture capitalists who are main investors.

Tim: Well, that makes sense on why the VCs would know you and select you. But was there any particular trigger for -- because the company had been around for seven years. So, was there any trigger for saying, look, it’s time to bring in a more professional CEO?

Toru: So, I was third CEO. First CEO is professor. Second CEO, Mr. Umeda, he was engineer. So, Antelope and Spaxk thought same way. So, three of them were trying to find next CEO.

Tim: Okay. And it’s something they’ve been working on for a while, but just it took a while to find the right person.

Toru: Yeah, yeah.

Tim: All right. So, I mean, I’ve noticed this pattern a lot in Japanese university spin-outs, the kind of full-time professor, part-time startup CEO.

Toru: Yeah. I think professor at university, their main purpose is research. Company business needs revenue, profit to run a company. But to do so, we need to meet with customer and our products must solve issues. And also, because startup tries to create new business, nice to have market, which is not target, at least at first. Yeah, because, oh, that’s very nice, but they don’t pay.

Tim: Well, that’s the problem, especially in something like robotics, which is interesting. Everyone loves to talk about it. But finding people to pay for it is always difficult.

Toru: Yeah. So, that is why we need to find the market who needs this robot as must-have. And the second important point is how to tenderize product. Because real world, to meet with each conditions, we have many customized product.

Tim: That’s so dangerous for startups. Especially in Japan. Japanese customers demand a huge amount of customization.

Toru: Then that does not make sense for business point of view.

Tim: It’s really hard, yeah.

Toru: Yeah. But for startups, it is very risky area because we are small startups. Let’s say Toshiba or whoever, big company asked to develop a product.

Tim: You don’t want to say no.

Toru: Yeah, yeah. And also people at startups are excited. Oh, we are called by such kind of big enterprise company. Let’s try to develop this product. Then many customized product and finally business does not run.

Tim: You know, I think that new CEOs make the mistake of thinking that the cost of the product is the cost of developing that product. Not the cost of supporting that product for the next four or five years and fixing little things with it. And it’s extremely expensive.

Toru: Yeah, exactly.

Tim: So, was the pipe cleaning robots something new that you brought in or was that one of the projects they were working on before you joined?

Toru: When I joined, a couple of projects were running. I decided to select this product only. The others I stopped.

Tim: So, you launched Soha Pipe Inspector last year and you’re calling it robot as a service.

Toru: Yeah, yeah.

Tim: So, explain how that works.

Toru: If we try to sell this robot, a price, let’s say a couple of hundred thousand dollars.

Tim: And so your customer would have some expensive asset they don’t know what to do with.

Toru: Yeah. So even though customers, oh, that’s great. I want to use. But how we can get approval for such is the brand new robot. Is it really okay or such kind of?

Tim: Yeah, that makes sense. It’s lowering the short term costs. It’s lowering the long term.

Toru: However, it’s very hard for us from a cash flow point. Because to manufacture the robot, many cash is cashing out. Then revenue is two years or three years. So, we are a small startup company. That is not easy for us.

Tim: Som for the robots to the service, are you providing just the robots or are you also operating them and going on site?

Toru: Operating is very easy. Of course, initial training, our employees would go to a customer site. After that, the customer itself and some cases, this maintenance operations kind of outsourcing company doing operation for this robot.

Tim: Okay. And when you’ve launched this service, you launched it in partnership with Mitsubishi.

Toru: So, Mitsubishi HC Capital is investor for us. And we sell this product to Mitsubishi HC Capital. So, we can get cash upfront for this robot itself. Then Mitsubishi HC Capital rent it to customers. Then customers monthly revenue Mitsubishi HC Capital and we share. Mitsubishi HC Capital is financial partner. In case of semiconductor factories, we have another strategic partner whose name is Suzuki. This is trading company targeting semiconductor industries.

Tim: So, Suzuki is currently your main sales channel.

Toru: Exactly. And because this is brand new product for worldwide, Suzuki and our salesperson together go to customers.

Tim: In the future, are you planning on maintaining that kind of arrangement? Do you plan on developing a direct sales force?

Toru: I believe such kind of partnership is very important because we are not expert for semiconductor industries. We need partners who really know their problem. So, that is why Suzuki is a perfect partner for us. So, if the next target, let’s say chemical industry, I think such kind of partner who has expertise for chemical industries we need.

Tim: You know, I find this point to be really interesting and really different between American startups and Japanese startups. Where American startups are obsessed with owning the customer relationship. With owning the customer. You never want to let anyone get between you and the customer. In Japan, that’s just not the case. I think it seems to be a much higher trust relationship.

Toru: Yeah, of course. Even though we are selling through partner relationship between us and customer is strong because maintenance is required for this robot. Customer needs to send back this robot to us. So, relationship with customers is continuing.

Tim: Right. But I would also think it’d be really essential to have that close customer relationship just so you can understand the deep needs of the customer and learn how to improve your product.

Toru: That is why customer side, not only our partner, but also our salesperson and the field engineering, or in some cases, engineer itself, goes to the customer side.

Tim: So, it really is a joint sales process.

Toru: Exactly. That’s very important. Otherwise, as you said, customer’s actual situation, our engineers should understand. Then our robot is more useful. I ask the engineer, go to customer side.

Tim: Yeah, I think it’s an advantage that Japanese startups have that U.S. startups don’t have, is that sort of long-term trust and transparency.

Toru: It’s really, it’s great. And also our target market is not Japanese only. We are targeting worldwide. But to expand outside Japan, partnership is critical for us.

Tim: But I would imagine that the semiconductor industry, the industrial chemicals industry, these are both global industries. And wouldn’t a strong reseller in Japan also likely have clients overseas?

Toru: Some cases, yes. Some cases, no. So, if we can find partnership who covers worldwide, that’s perfect.

Tim: Okay. So, Soha uses soft robotics. We were talking about that a bit before. But we don’t see many soft robots out there. Why not?

Toru: It’s not easy to develop.

Tim: It’s just soft robots are more complex and expensive to build in general than their rigid counterparts?

Toru: Yeah, yeah.

Tim: Well, I mean, this makes sense. So, you’d only use soft robots when you really need them.

Toru: Yeah, yeah, yeah.

Tim: Okay, well, that’s probably why we don’t see so many. Well, that makes sense. They’re expensive to build. There’s a lot more moving parts and joints and things.

Toru: Yeah, and also many soft robotics try to mimic living thing or animal or...

Tim: Sure, like a snake or an earthworm.

Toru: Yeah, yeah. Because snake or earthworm or human itself can move very flexible to target these kind of conditions. Soft robotics is more suitable.

Tim: Let’s talk about Japan and robotics in general. So, we mentioned before that these days, the U.S. and China seem to be getting all the attention. They seem to be far in advance of robotics globally. But where do you think Japan is uniquely strong in robotics today?

Toru: Technical point of view, still Japanese persons is very good to develop this kind of soft robotics. Kind of very precise to run through this kind of severe conditions. Very careful we need to develop. This approach, Japanese people is very good.

Tim: I do think, I mean, there’s no question Japan even today is still leading in like factory automation robots and some specialized areas like that. They’re still the best in the world. So, do you think that skill set is now being applied to other types of robotics?

Toru: So not only soft robotics, but other robots. Japan should be good.

Tim: Okay. Well, listen, Toru, before we wrap up, I want to ask you what I call my magic wand question. And that is, if I gave you a magic wand and I told you that you could change one thing about Japan, anything at all, the education system, the way people look at innovation, the way people look at risk at universities, anything at all to make it better for startups and innovation in Japan, what would you change?

Toru: Japanese people or companies, they like risk assessment.

Tim: Oh, yes.

Toru: So, when I try to do new business or new development, especially at Toshiba, many people were asking me, do you check this point? Are you really okay with this? Then our dream is cut.

Tim: So, I’ve worked in several large Japanese companies, and I have noticed that people will be, of course, you’ve got to look at the risk, but people seem much more focused on risk than opportunity.

Toru: Exactly. Yeah. That is one of the reasons why I jumped into startup world. But even in the startup world, many venture capitalists still like risk assessment.

Tim: Well, I find that venture capitalists have this reputation of risk takers, and they like to think of themselves as real risk takers. But in general, no, they’re pretty risk-averse people.

Toru: Yeah, yeah, yeah.

Tim: In their own way.

Toru: Against the Japanese enterprise companies, you are right. This round, because we are creating a new market by this robot, so there are some discussions we need to do with venture capitalists, but no report in the world. I discussed with some venture capitalists how to measure opportunity size. Then some venture capitalists invested. However, other venture capitalists, ah, I cannot imagine why.

Tim: Well, that’s the thing. Yeah, I’d say most VCs will not invest in a new market that other VCs have not invested in yet. I think it’s kind of human nature. Very few people want to be the first penguin.

Toru: Yeah, exactly, exactly. And also, venture capital has many LPs. So, they have responsibility for LPs.

Tim: Right, and if they’re investing in some risky new startup, it’s harder to explain than everyone losing money together.

Toru: Yeah, yeah. That is why this rounding is very hard.

Tim: So, do you think that aversion to risk is changing in Japan?

Toru: Because I have been doing robot startups, from this point of view, period of funding should be longer. This mainly 10 years period for funding. 10 years is not enough for robot startup.

Tim: Yeah, some corporate venture capital have much longer, like 20 years span. But do you think that Japan in general, not just VCs, but Japanese enterprise in general, and society in general, is getting less obsessed with risk?

Toru: I don’t know. But another point we need to change is IPO is main target. But I believe M&A should be target too.

Tim: Well, I think we’re seeing a lot more M&As now than we did before.

Toru: Yeah, and outside Japan, for example, US, M&A is much popular than IPO. Used to, in Japanese mind, M&A is kind of loser.

Tim: Sure, sure. It used to be distressed assets, right? But I think that’s definitely changing. There’s been some very successful M&As.

Toru: Yeah, yeah, yeah. So, we need to change because robot startups need money and time. So M&A should be good approach to have robot startups grow.

Tim: Yeah, I think so too. Well, listen, Toru, thank you so much for sitting down with me.

Toru: Okay, thank you very much.

Outtro

And we’re back.

Toru’s experience with Solaris really drives home the importance of partners in Japan. Both for a startup’s go-to-market and for their finances.

Now the co-selling agreement that Solaris has with Suzuki is not unusual in Japan. Neither side really fully owns the customer relationship. And the risk of disintermediation or substitution is pretty low. It’s a slower but less transactional business culture.

But let’s talk a bit about risk.

One interesting thing that occurred to me as I was editing this episode down and listening to Toru’s comments about Japanese VC behavior. And that’s how very different the impact of low interest rates is on American and Japanese startup investments. In the US, low rates means that everything gets levered up and that a lot more money flows into all kinds of investments. So, that even marginal, highly risky projects get funded.

To be sure, most investment in Japan responds pretty much the same way.

Startup investment, however, can be a bit different. While the pension funds, the banks and the like tend to be sophisticated investors, most CVCs are not. And those CVCs control a lot of startup investment. Both directly and as LPs. And these CVCs are often not benchmarking against stock or private equity indexes, but against their own cost of capital. So, when interest rates drop, they need to take less risk to beat the benchmark and they move to a more risk off posture.

So, in this small but important segment, low interest rates can actually make some Japanese investors more risk averse. Some things just work differently in Japan. That said, sometimes things end up working better for startups in Japan. A recent METI report on the impacts startups are having on the economy show that even though startup investment is decreased slightly over the past three years, the economic impact of startups has been rising steadily.

Startup activities now account for about four percent of Japan’s GDP.

So, Japan is not exactly a startup nation yet but every year, startups are contributing more and more to Japan.