Nearly four years ago, Dmitry Grishin launched a US $25 million fund to invest exclusively in consumer robots. Grishin, the co-founder, chairman, and CEO of Mail.ru, the Russian Internet giant, believed that robotics was going to be one of the next big technology revolutions, and he was willing to put his money where his mouth was.
Now the Russian investor is ready to double down on his vision. Or actually double double down. Grishin Robotics has recently announced a second fund four times as large as the original one. The new $100 million fund will seek Series A and B deals and expand its focus to include startups in markets like connected devices, collaborative and material-handling robots, AI and data analytics, and industrial Internet of Things.
Dmitry Grishin, CEO of Mail.ru and founder of Grishin Robotics.Photo: Grishin Robotics
Grishin’s previous investments include many different companies. There’s Sphero, maker of the BB-8 droid toy [pictured above], and Double Robotics, which sells telepresence robots. He’s also invested in a company called Spire that develops micro satellites, and in Petnet, which builds robots for feeding pets (Amazon recently became an investor as well). More recently, he backed Ring, which offers smart doorbells—the first investment out of the new fund.
Though none of his startups seem to be on the verge of a billion dollar or greater exit, Grishin still believes that robotics is going to be a massive market. “It’s still very early days,” he says. “We need more funding and more entrepreneurs and more products, and then we’ll see this market take off.”
IEEE Spectrum spoke with Grishin by phone last week to find out more about his plans for the new fund and views on today’s robotics industry.
IEEE Spectrum: What are some of the lessons you’ve learned from your initial robotics investments?
Dmitry Grishin: Something that became very clear to me over the last three years is that entrepreneurs need to focus on robots that solve one particular problem. Because robots are very complicated devices and even when you’re trying to solve one specific thing, like cleaning your house or feeding your pet, it’s very difficult and there will be a lot of surprises and challenges.
So don’t try to build a product that does too much. I believe that as the market evolves, you’ll have big, standalone verticals like home automation and telepresence, where robots solve particular problems. And then, as robotics technology improves and the market becomes really big, we’ll move towards more universal solutions. Look at smartphones. We didn’t get smartphones on day one. Before smartphones we had 10 years of just mobile phones. And we also had cameras, MP3 players, calculators, and other devices, each with their own markets, before the smartphone integrated everything. I think that robotics will repeat the same pattern.
Last time we spoke you said something that was a bit controversial. You said that “robotics has too many dreamers and needs more practical people.” Do you still think that’s true?
The situation is much better, but we still need to improve this area. We have excellent technical people who are great at dreaming up new technologies but, like I said earlier, many still don’t focus clearly on a particular problem or application. Problem number two is that people create very exciting products but don’t have experience building hardware and end up having lots of issues with manufacturing, quality, and customer support. Some people have a software background, and usually in software you can ship a product and if you have an error or bug, you can fix it quickly with an update. With hardware, fixing things is much more complicated. If your product is not working, people will just send it back to you.
My hope is that funds like mine and others, which already have a lot of experience with these issues, can help new companies to make fewer mistakes. Because usually they are very similar mistakes that we see people making over and over. Another thing we need is to have more hardware accelerators that can offer expertise with the design, production, and distribution of products. I already invested in one called Bolt, which is based in Boston and Silicon Valley, but there are several others, like Highway1, Playground, Lemnos Labs, and HAX.
And why did you decided to expand the focus of the new fund to include things that are related to robotics but aren’t exactly robotics?
My definition of what is included in robotics is broad. For me, things like connected devices and Internet of Things are part of robotics. The big reason to expand the focus is that, right now, there’s a big new trend where you have not only software and Internet startups, but you have startups putting together smart hardware plus smart software. This creates lots of new exciting possibilities.
So when you look at the tech landscape, what are some areas that you find most interesting for startups right now?
I think that right now there are several interesting trends around home automation. Startups focusing particularly in security, like door locks, that’s an interesting category that has just emerged. Another very big promising market is connected toys, which can offer a lot of fun and social features. For example, the more you play with your toy, the smarter it becomes. You can also have games between groups of children.
Not too many companies have found the right way to do this, but one example is what’s happening with Sphero’s BB-8. They plan to launch a new product called the Force Band, which is a wrist band that you use to drive your Sphero like you were using the Force. I think it will be super cool, because you can create special gestures and make it move in different ways.
Many people are trying to create totally new devices for the home. But we found from our experience that it’s very difficult to create new experiences and new devices in the home. People are very conservative. The approach that we think is the best right now is not trying to create new category of devices; you should just make existing devices smarter. Like a thermostat, or a doorbell. We see that it’s much easier for people to accept Internet of Things and robotics technologies if it involves a familiar experience. And this is how I think robotics will migrate into our lives.
What about social robots? What do you think of Jibo?
I’m very excited about the possibilities of things like social robots but I think you should be very careful in terms of what kind of experience you are offering to users. For many products, it’s very difficult to explain to potential customers what kind of problem the device is solving, or in other words, why people should buy it. I think it’s better if your product doesn’t have a long list of things that it can do, but maybe only one or two—the key features. And these features need to work really well.
I believe that in the long term we’ll see more and more types of assistants around us, but, again, you should be very careful in terms of the way your product works. For a social robot, for example, if you say it can do a whole bunch of things, then users will have such high expectations that if the robot doesn’t deliver, you will end up disappointing them. And solving things like voice and image recognition is very hard. I think that you should do the opposite: you should say that your robot does one thing, but in reality it can do more. So you’re not overpromising, and the user will be happy with the extra features.
And what about all the AI startups we’re seeing lately?
I absolutely agree this is an important trend. But I think that AI will be the next wave. First, we need to create devices that will let us get more data about consumers and how they interact with products. Then you can start to iterate and make the devices more and more intelligent until we finally can build powerful AI systems.
To give you an analogy, look at how Tesla has rolled out some of its new features. First Tesla shipped the car. Then, as more and more people bought the car, Tesla collected a lot of data and learned more about the users. Later it shipped more advanced technologies like autopilot and other stuff. So my belief is that artificial intelligence features will become a second wave after we ship a first wave of connected products.
Jumping to another area, self-driving cars, there are now a lot of startups trying to solve different problems related to vehicle autonomy. Is that something you’re interested in?
As I mentioned earlier, I like the Tesla approach, when you’re not bringing a fully autonomous car to the market and saying that it can drive everywhere. You should be very careful about the user’s expectation because once you say it’s a self-driving car that does everything, you could be in big trouble. I think you should try to make your car smarter one step at a time.
I think that for the first wave of self-driving cars you’ll see them in very structured environments, not on the streets, but in university campuses, parking lots, golf courses, hospitals. The right approach seems to do iterations, progressively adding more intelligent features to your car. And there are some interesting startups in this space that we’re looking at.
There’s been a lot of talk about unicorn startups, and one report published last year even listed a few robotics companies. Do you expect we’ll see more and more unicorns in robotics?
First, let me say that I think we should focus less on valuation and more on finding the right markets and the right problems for startups to solve. I believe that in general home automation will be a huge market, and there will be a unicorn sooner or later in this category. Logistics and e-commerce automation are also potentially very big and exciting markets too. And I’m still a big fan of telepresence robots. I believe that more practical companies solving specific problems will have a much better chance of becoming unicorns than others that try to do too much.
This interview has been edited and condensed.
Erico Guizzo is the Director of Digital Innovation at IEEE Spectrum, and cofounder of the IEEE Robots Guide, an award-winning interactive site about robotics. He oversees the operation, integration, and new feature development for all digital properties and platforms, including the Spectrum website, newsletters, CMS, editorial workflow systems, and analytics and AI tools. An IEEE Member, he is an electrical engineer by training and has a master’s degree in science writing from MIT.