Robotics is a field that seemingly hasn’t really delivered on its promises. Sure, plenty of robot arms and other robotic solutions in logistics and industrial spaces, but what about the rest… including, Consumer?!
Our show: Tech DECIPHERED brings you the Entrepreneur and Investor views on Big Tech, VC and Start-up news, opinion pieces and research. We decipher their meaning, and add inside knowledge and context. Being nerds, we also discuss the latest gadgets and pop culture news
Subscribe To Our Podcast
Welcome to Episode 64 of Tech DECIPHERED. Today, we will discuss robotics. Our new overlords are coming. Robotics is a field that seemingly hasn’t really delivered on its promise. Sure, plenty of robot arms and other robotic solutions are out there. But somehow it feels like we should be actually already controlled and dominated by our robotic overlords.
Today, we will discuss B2B Robotics, their growth, adoption, industry use cases, the drivers of innovation in that space, players in that space, challenges. We will also talk about consumer robotics, talking about also robots that have gone mainstream, personal and social robots, emerging trends that are happening in the home, market dynamics.
Then we’ll look ahead to the future of B2B, the future of consumer, and the societal and workforce impact, which obviously is going to be the last but not the least topic that we will address today. Let’s start with B2B. B2B Robotics.
Yes, thank you, Nuno. It’s certainly a growing space. As you say, it might not feel like it, but at the end of the day, we already have quite a few millions robots in B2B. It’s estimated that the stock of your operational robots, industrial robots, is around 4.28 million units in 2023, which was a 10% increase year-on-year.
If we look at China, and I think we talk about the fact that we are installing around 400,000 robots a year. China alone by itself has been installing 276,000 units in 2023. China is definitely a leader in that space and that might come as a surprise to some. That number, to put it into perspective, is five times higher than the second place, Japan.
I think China has recognized a long time ago that because of its population composition, that they’re going to have a lack of people to produce, to be in factories, et cetera, so they actually have been adopting robots now for many, many years. I think now, you were joking with me just before we recorded this, but now that we’re having trade tariffs and all that stuff, it might make sense to have a broad discussion around why robotics in an industrial environment are key.
Definitely. I think that actually I was surprised to hear that, when we’re making this recording on April 15, I was surprised to hear in the Trump administration, quite a lot of very positive support regarding AI and robotics. It looks like part of the plan is already acknowledging this is not like the old-school jobs that we are going to bring back. It would be a different type of jobs. There will be way more robotics than before. Obviously, more robotics is possible thanks to the latest advanced in AI.
There is some consensus that, yeah, it’s not just bring back the old jobs as they were, but it would be a new type of jobs. It would be a new type of industrial revolution and acknowledgment that robotics are here to make all of this not just more efficient, but even possible. Because it’s clear that if you take the US, for instance, there is actually not so much unemployment in the US. For an industrial revolution to happen, we need to bring our new friends, the robots.
If we go through the use cases where we see robots coming in, obviously manufacturing, automotive and electronics will probably come to people’s mind. A lot of robot arms are being used there. More than serve your classic solutions, et cetera. But it’s a lot of robot arms using on specific pieces on the line, a lot of pieces of robotics that we don’t recognize necessarily as robotics, but are actually roboticized systems used, in particular on automotive, again, in consumer electronics.
Basically, it has to do with the fact that there’s a tremendous lack of… One, there’s going to be, at some point, the limitation on how fast you can produce, so that’s very obvious. Two, there’s actually even lack of people that can do some of these roles. Some of it is actually high precision work.
In most of these environments, there are environments where humans need to coexist with the robots that are on the line, and that has led to the creation of this movement of cobots. Robots that are able to basically interact with the environment that they’re in, obviously be very careful around safety procedures that they don’t kill a human being that is around them. There’s been a lot of innovation around that using computer vision, using sensors, and using a variety of other elements around that.
That cobot segment in particular is growing quite rapidly, estimated to be 14.7 billion market by 2031. But obviously, we all recognize manufacturing, there’s robots being used, so no shocks there. One of the largest markets by far.
Definitely. Collaborative robots are a recent phenomenon made possible thanks to better activation of motors, electric motors possible thanks to new AI, vision AI. There is a lot of the latest technologies that makes this possible versus your big, bulky, robotic arms that used to be encaged. It’s not to kill anyone. Yeah, it’s a change from bigger to smaller. I mean, we can even say step by step, and we’ll talk more about it, but more humanlike in terms of how it looks.
Here we have around 113,000 professional services’ robot for transportation and logistics sold in 2023 alone. That was a big increase of 35% year-on-year. I think we have all seen how Amazon is doing with some robots in their warehouse, logistics systems. What some might not have seen is as impressive in China. If you look at the efficiency of some of these warehouse in China, it can be very impressive. I don’t think they’re holding anything back actually in China.
Yeah, definitely not. There’s a lot of amazing things happening on the logistics side. Obviously, Amazon bought one of the early granddaddies, which was Kiva Systems, which was bought by Amazon, and now it’s Amazon Robotics. There’s a lot of stuff happening on the lines, autonomous mobile robots for the transport of materials, robot extruders, mobile picking robots, automated forklifts, and a bunch of other things.
It’s also an environment where I know Amazon is obviously using a lot of co-robots, and there’s a lot of people around as well still. Very interesting environment. I’d say probably Amazon has been the company on the logistics side that innovated the most, the earliest, certainly in the US.
I still remember the Ocado ads where they use robots. In the UK, there’s always been a lot of this debate whether that was actually very useful, very impactful in Ocado’s operations or not. Also, retailer there. But definitely, I think we would recognize probably Amazon has been one of the players that has innovated the most in this space for a couple of decades now.
Yes, it’s definitely I mean, that has been key for them to grow, to expand their network. As you know, I mean, Amazon has a lot of people working in their warehouse, but they were still able to manage to grow so fast while not growing proportionally the number of people working there. It’s thanks to advance in robotics.
Indeed. One other area, obviously, of use has been in the healthcare and medical space, and particularly in high precision need environments. Surgical robots like Intuitive Surgical, their Da Vinci system, does provide the precision that you need for procedures that require minimally invasive play with high precision. Basically, things that you can’t really… The human hand actually might not be precise enough for it, even if you’re an amazing surgeon.
Beyond that, medical robots have grown quite significantly. It’s grown to the thousands units, which might not seem a lot compared to the hundreds of thousands that Bertrand, you were just mentioning. But obviously this is a much more specialized space, less mainstream in some ways of robotics. A lot of these devices are incredibly expensive, and then we’ve seen a lot of innovation around that.
There’s obviously the old-school players that are on the market, Johnson & Johnson’s of the world, Medtronic, that have also entered that arena, the surgical robot arena. Then there’s a bunch of other applications for robots that I would highlight. Hospitality or hospital delivery robots, so robots in hospitals that do deliveries.
Rehabilitation exoskeletons, which is really, really cool. If you guys remember Aliens, the second exoskeleton that she has and stuff like that. Obviously this is for recovery for people that are rehabilitating. It supports them in their rehabilitation without having necessarily to have a human there all the time doing that. Then obviously diagnostic robots. A lot of players in this space, CMR surgical as well, Ethon, Exobionics, Cyberdine. There’s a bunch of things happening right now that are quite cool.
Yeah. If you look at rehabilitation robots, the growth was pretty amazing at more than 128%. You can call it a surge. Some of these metrics are pretty amazing. As you said in rehabilitation, for sure, if you need a human by the side all the time, it will be way more expensive, and you might not have the staff to just do that. It seems like a very, very obvious option.
Going back to medical robots used for surgery, for instance, as you say, I mean, robots can do stuff that humans cannot. You can have a very small robot moving your own body in some ways, coming from places that would be impossible to operate. It also could end up being more safe. Ultimately, you decrease the need for specialized surgeons. You might need less of them, or you might simply do more with less.
Correct. Next, we go to agriculture and farming, AgTech. Actually, the name probably comes from use of robots in agriculture and drones, et cetera. We have some big numbers, basically according to our research, 20,000 units in 2023 that were used. There’s high precision agriculture that, again, requires things that are very, very precise. There’s also the base equipment that are used by humans that are being repurposed to become equipments for agricultural activities.
Autonomous tractors and sprayers would be a great example of that. Milking robots, the robots that do the milking of cows, et cetera, also makes a lot of sense. Again, replacing human in an activity that probably can be more easily replaced. Then you have, obviously, the drone space with a lot of things around field scouting, the surveillance, the pieces around how to create new ecosystems around the fields.
Last but not the least, obviously, we have picking robots, the robots that pick up stuff. Those robots are particularly complex. Obviously, some fruits, some vegetables, et cetera, need to be picked in a way that is more complex because they require either cutting or something like that. Its identification is pretty critical because you don’t go through the whole process of actually growing the stuff and then have the robots destroy it.
But yes, and there obviously there’s a lot of players, mostly the old guys, or the guys that have been around for a while, obviously DGI on the drone side, Yamaha, a motor, John Deere, then there’s a couple of startups as well. He’s emerging in this space, Farmwise and a few others that are doing a lot of activity. I’ve looked at this space quite a lot.
Just as a disclaimer, I spent a significant amount of years in a VC fund as a venture partner that was really very focused on Deep Tech, Frontier Tech, in particular, robotics. It was actually in the name of the firm. As part of that, ended up interacting with a lot of the B2B applications of it. AgTech seems like an obvious use for it because there’s searches of people. It’s been a huge debate around immigration and illegal immigration in particular.
The reality is agriculture has very, very thin margins in general. It’s maybe a broader point we can make at this point in the episode, but the problem with robotics is systematically, is this a big enough issue that you should solve it with a very complex solution. Because normally, robotics, even if it’s a verticalized solution, require complex elements, not just on the hardware side with sensors and a bunch of other things, but also on the software side.
I think AgTech is one of those areas where there’s been a couple of companies, there’s been one or two interesting exits with Blue River going to John Deere, et cetera. But still feels like it’s an area that we haven’t figured it out yet. Either we fully roboticized a facility or do something else, but we haven’t really figured it out yet. My theory is it’s just the margins are just too thin, and it’s too complex.
It’s also a space where the owners and the people that operate the fields are actually not necessarily people that have any technical knowledge of any kind. The solutions do need to be really, really simplified to that level.
That’s a good point. I have also seen quite a lot of startups in that space. At Red River, we invested in a startup called Robovision in Belgium. Some of what they do is definitely to serve agriculture needs. It’s not easy, as you say. First, it’s not like the car market, for instance, where you make millions and millions of cars every year, where you have a clear market and buyers are ready to buy.
In agriculture, there is tractors and stuff, but the competition is very different. The volumes are not the same, and there are always, as you say, questions about how do you deal with a space that is not actually so human-friendly. It’s not a road. It has not been designed for it. It’s a field. Two, the people around are not experts, are not technical, so the smallest issue can be a problem.
It’s really a big question. At the same time, I must say I’m quite and optimistic that robotics have definitely the ability to transform agriculture. If you look back, the story of agriculture is just a question of improving the methods, the tools and you increase the yield. That’s how we managed to move from 95, 99% working in agriculture and the fields to 1% today.
I’m still a believer. As you say, there is a big question around immigration, especially legal immigration in the US, probably quite a lot is happening to serve agriculture needs. Given that push, that might create some new incentives in order to stay competitive.
Yeah. I mentioned Blue River before, which was a company acquired by John Deere, apparently for $284 million or so. The company raised in total $30 million, so it feels actually like a decent exit. But it’s one of the very few plays in this space that has really gone to a fruitful exit. Also, important to magnify that a lot of the AgTech stuff for example, that’s been happening in vertical farming, for example, is around robotics.
It’s around almost managing vertical farming a little bit like a manufacturing facility with its own complexities, but a bit like a manufacturing facility. There we really haven’t seen a runaway winner. I mean, plenty raise a ton of money. From my knowledge, they haven’t done necessarily great. I’m not sure how roboticized they were. There’s a bunch of other companies in the market that we know well, like 1.1, et cetera. There’s definitely a lot of plays going on that depend fully on robotics around the AgTech that we see today.
Yes. You talk about a specific type of agriculture. I mean, the ones they try to do in some convert some old factory, for instance, do it even in the cities, urban farming. I think this stuff by definition, I’m not super optimistic that the yields can work. We’ll see, obviously, maybe if you control well water space, automation, but it looks like yields are definitely a big issue in this space. At some point, size and scale matter.
It becomes a unit economics issue at the end. I was very involved in one of the plays that I mentioned before, and it is clear to me that we will get there, but to your point, there is a need for scale that needs to happen. These guys need to get big contracts. They need to get really, really big contracts in terms of what they need to put in the market.
It’s also very clear to me, it’s a very difficult market for you to do stuff that goes one hop from consumer. Let me do a vertical farm and sell to supermarkets, and they’ll get fresh, produced from me and whatever. It’s very tough. There’s a lot of intermediaries in the middle, distributors, et cetera. A value chain that is much more complex than we think it is.
I think if you’re someone who’s not spent a lot of time around agriculture, you think, “Oh, there’s people in the fields that produce stuff. They sell it maybe to markets or someone in the middle, and that person in the middle somehow sells it to supermarkets and whatever and what we get.” No, there’s a lot more players in the middle that get involved for all of the stuff. There’s brokers and the wholesale players, and then there’s even in retail, there’s different types of players as well. It’s like, it’s my God.
There’s a whole communities feature market.
Yes, it’s an ecosystem that’s extremely, extremely complex. Maybe moving to construction, I think it’s probably an industry that, as promised a lot, that we would have construction robots replacing people. It’s a particular hazardous role. In construction, there’s accidents, people get hurt, and they might even get killed. Obviously, it’s an area where it’s high risk, high rewards. Robots, you would assume, would be great for this.
Honestly, I personally have seen some interesting innovation around 3D printing, which I would argue is it robots or not? Drones for surveying and for looking at buildings and a variety of other things. We’ve had this huge promise, I think, in construction of prefab, of the ability to do prefab buildings and prefab homes and all that stuff. There’s a lot of interesting players out there, so I’m not disrespecting them necessarily, but it hasn’t really scaled. Again, we haven’t seen a player that has really won that market. We saw a player that failed miserably, which was Katerra, which wanted to use stuff like that. I’m not sure they ever got to that level.
Yeah, I was going to say it’s been a decade we talk about prefab and stuff, and for some reason it has not taken off at all.
It’s not. I’ve seen amazing things, in particular around single-family homes and things like that. ADUs, right?
In particular, California with a change in less installation, et cetera, and certain markets, in particular, like LA County being super aggressive on opening up the ADU market, et cetera. Nothing dramatic. In this space, we see autonomous construction machinery, which in some ways could be used for what we were just talking about, the prefab market, robotic bricklaying, 3D printing robots, site inspection drones.
I think we are seeing more and more site inspection stuff. I think we are seeing more and more stuff around 3D printing. I don’t see this industry being disrupted by construction in the next couple of years, by robots, rather, in the next couple of years. I think that will be very optimistic.
Yes, it seems like it will take a while. Not saying it might not come, but it’s not as if we were at a tipping point at this stage. Going into retail, hospitality, we’ve seen a lot of robots that try to come to some sort of, I would say, awkward space, trying to welcome you SoftBank at some robots. Robots are acquired from a French company, actually, maybe at least a decade ago. It’s 54,000 units sold in 2023, so a pretty good increase, 30% plus. That also includes hotel delivery robots, restaurant food runners, concierge info robots. I think it’s a mix of everything, to be frank.
There’s a mix of things going on. Again, it’s a space that everyone’s like, “Obviously, this is going to be delivery robots and all sorts of things.” This is just going to scale very, very rapidly just if you think about the use cases. Inventory scanning, customer service, restaurant service robots, floor cleaning robots, delivery robots, all this stuff. It’s obvious.
Again, we haven’t had amazing winners in this space. There’s a bunch of players, question mark on when we’re going to get them. I think some of the problematics of this, I think in closed environments, I’m very surprised that we haven’t figured out closed environments. Why are robots not able to do inventory scanning at scale? Why are robots not able to deliver food to me in a restaurant?
That part seems like a problem that’s eminently solvable because floor plan is more controlled. I know there’s people, so that’s the cobot dimension to it. In some ways, the more the food delivery things, which seem to have taken off in some places, actually. Seemed like a more complex issue, but somehow I guess there’s more activity in that space. We’ll see.
If you remember for hotels, you had these delivery robots bringing your food potentially and not much happened to be frank. I remember seeing some startups trying to do robots for universities, colleges, and the same thing. I don’t think much happened. I think in many situation, it’s an awkward space because you’re like, “Okay, so I’m delivering because people don’t want to just go downstairs and buy at the counter or just buy an automated machine that is just delivering food.”
We know the machine to get your bottle of drinks, your snacks, it’s pretty straightforward. The robots providing not the last mile, but the last few feet. It’s a tough value proposition and it has so many constraints. For robots in restaurant… There is one chain that you probably know where I go once in a while, where you have these delivery robots directly to your table. It’s Haidilao.
I’m always impressed that the only restaurants in the US where I go, where we get served by a robot is actually a Chinese chain.
It’s a hot pot guy. They have pretty good food, actually. I’m always shocked because it’s really working. You see the robots coming up, not hurting anyone, finding their way around. You have kids in this restaurant. It’s a very kid-friendly place. You can see it’s not a replacement for human. It’s really making humans more efficient and for sure decreasing the risk of breaking something, having a plate falling off, that stuff. I think it’s a very smart use. I’m surprised why not more restaurants do that, but I think there might be a cultural issue, especially in Europe and in the US.
I think there might be. In Asia, it’s probably where we’re seeing a lot of these advancements coming in. We’ll talk about some of the geopolitics around this in a bit. Moving maybe within still B2B robotics to what are the key innovations and technology movements we’re seeing. One very obvious one, which is the one is artificial intelligence and the use of machine learning, deep learning methodologies, computer vision, and a variety of other things to optimise lines, to optimise how things are done, but also to optimise in-process things and what’s happening.
That is very, very obvious. There’s a lot of things happening around that space. Robots are not very useful if they don’t have some brain, if they are not doing something that is defined within the realm of some degree of operational intelligence, as I would call it. AI is critical to that. That I think is the thing we’re all afraid of. It’s when robots are actually using proper AI then they’ll come to kill us. Sorry.
Maybe we’ll talk later about the key players, but obviously NVIDIA has been pushing very hard in the robotic space and sees that as one of their big future markets.
Maybe moving along to something we already talked about, so co-bots, mobile and collaborative robotics in general. We’ve already talked about it. It uses a lot of the AI pieces that we were talking about, but it’s this notion that we will need to have things in this world that interact with us. There’s some cool things happening. Tesla with Optimus and Agility Robotics doing bipedal robots. We did have a bunch of… Boston Scientific was the dog ones, early on.
We’re having all of these things that are on the one hand, basically interacting with humans and on the other hand, looking more like animals and humans. Becoming a little bit more anthropomorphic on the human side as well. Interesting things that are happening in that space. As we discussed, those markets are real, they’re growing fast. Certainly in the manufacturing space, we know that’s the case. I think there will be a lot of crossover technology transfer to other use cases in the more consumer-related domain where there are people and little kids running around doing stupid things, et cetera.
Maybe to keep going on the technology, another way to make robots happen at scale is all the work done in term of digital twins and simulation. When you can build virtual replicas of your factory for testing, for optimizing, that will help you decrease downtime, that will help you be more efficient with your deployment. That has been a key part of the game in order to bring robots to the factory.
Indeed. Last but not the least, the things that we would like to highlight, more business models that are changing robotics. A robot can be a very expensive, cost for a firm, for a company, and for a startup. There’s a lot of things happening around robotics as a service where companies can lease robots on subscription rather than just purchasing them.
I’d say that’s more of a financing play, to say the very least, but there’s deeper things going on in the space with really solutions that are optimised towards outcomes-driven revenue sharing. People could say, “It’s just a business model innovation.” But there’s a lot of cool things happening around that. I think we might have some of the big manufacturing companies of the future, et cetera, merging because of these business models with a fuller stack where I have my own robots and that’s what I’m bringing to table. Making them available to you.
Startups like Formic that have been quite heavily discussed about, and they are trying to really go on a usage-based fee play. It’s solely based on usage. The jury is still out if this is going to really scale, if it’s going to be at the level of mass market. But certainly, if we’re thinking about small-scale manufacturing, et cetera, I would say it’s a great business model and we’ll create a lot of innovation around it.
I think business model innovation is quite critical. We have seen how software as a service, SaaS, has been able to transform the software industry and create basically a new dramatic wave of growth. For robotics as a service, could unleash a similar efficient model that will help scale to the next level robots, so it will be interesting to follow.
If we talk now maybe about leading companies and startups in the B2B robotics space and starting in industrial automation, industrial automation has the big boys, so ABB, FANUC, KUKA, Ishikawa, Kawasaki, Mitsubishi, Universal Robots, Teradyne. These guys are all super well-known. If we think about some of the big guys, they grew up with a lot of the solutions that we were talking about earlier, robot arms, but really expensive robot arms. That market is still dominated by these players, and they command significant premiums on their valuations, et cetera.
Let’s not forget that KUKA was a German company that got acquired by a Chinese company some years ago in 2016. I think that was crazy because it was one of the Chrome J-well of the German manufacturing industry. To let it be acquired like this was probably a strategic mistake. Germany decided to make some change in how they are going to let M&A happen, depending on the country of origin of the acquirer. The acquirer was a media group, Chinese appliance manufacturer. I think that’s really a fair question because when you know that the acquisition of Chinese companies is not something so easy for Western companies, I’m not sure why we should let it work in the other way.
Very good. Logistics and warehouses. We’ve talked already about some of these examples, Amazon Robotics with the acquisition of Kiva Systems. That’s how itself got pulled in. 6 River Systems, which I believe was acquired by Ocado to make their deployments due. Locust Robotics, Grey Orange, Geek+, Boston Dynamics. A bunch of players that have been done. What’s for me interesting, it’s an area that has had a lot of acquisitions by the retailers themselves, which is not obvious.
But not huge acquisitions usually as a result.
It’s not huge acquisitions, small acquisitions. Even Kiva, I remember was not a very big acquisition by Amazon. It was relatively important for them to go into that space. Anyway, I wrongfully, by the way, mentioned Scientific Boston Dynamics. Two different companies just to be clear. Boston Dynamics is the one with the anthropomorphic robots and also the dog-like robots. Boston Scientific, a total company altogether, but it is what it is. It’s a life science company.
There are other types of AI startups, collaborative robots, universal robots, Techman, Doosan. I think for me also, the big guy we should talk about it’s really NVIDIA. NVIDIA is making the brains of many of these robots. I have a lot of platforms to develop robots. It’s just open source, a new model. Recently, it’s called Project GROOT. GROOT, G-R-O-O-T, the foundation model for humanoid robots. We will talk more about that later on, but NVIDIA is definitely pushing very, very hard in the robotic space. They want to be the brain of all the robots in your factory, in your field, the medical space.
NVIDIA platforms everywhere. Obviously, we don’t have… There’s dominating legacy operating systems for a robotics operation, but most of them are owned by the firms themselves, the ABBs of the world, et cetera. We haven’t seen a prevalent robotics operating system that would work across domains that has killed dramatically. There’s a lot of interesting things, and it wouldn’t be in the realm of impossibility for NVIDIA to go up stack, do more software platform ownership, and even into who knows operating systems. We’ll see what happens. Not out of the realm of possibility.
Talking a little bit about challenges, just to put a bit of a bookend on the current conversation on B2B. We’ve talked quite a bit about geopolitics of it, et cetera. One of the issues clearly around robots are people. You can frame it as the integration of people and the integration of their skills with the use of robots. It’s complex because you have to reskill staff, not only to maintain robots, et cetera, but to work alongside robots.
In some ways, we people are unpredictable. We do stupid things at times, we try to optimise because we know the shift is almost coming to an end, whatever. I don’t know. In some ways, that piece of just integrating people with robots, integrating their skills with the skills of robots, et cetera, is something that we know for a fact has been quite complex and have led to a lot of safety issues and a lot of concerns around the use of robots, even in the B2B environment.
At the same time, it’s always easy not to say, “Okay, the robot is a danger.” Okay, sure, but we forget that many times robots are replacing work that is dangerous for human in the first place. Let’s be also careful about making sure we see the benefit of not having humans doing dangerous tasks or less dangerous tasks. A good example, actually, we don’t talk in that category, but if you think about robots for taking care of potential bombs, for instance.
Now you send your guided robot to try to basically make the bomb implode. No humans, neither. I can tell you that before that, it’s not a fun job for a real human. We know that there are robots as well for nuclear facilities, for instance, that can be used. I think that, yes, we have to be careful with human safety and at the same time, let’s not forget that a lot of these robots are decreasing risks for humans, ultimately.
Second big challenge is obvious. The high cost, the narrow eye. There’s a lot of things going on with robotics as a service, but the robotics as a service, someone’s paying for the robots. There’s still high costs in the space and there’s still a little bit of an argument on how do you think through return investment. It has to be a very top-down decision architecturally on what robots you go with versus not.
You can’t have a ton of trial and error in this space. I think that’s part of the issue. It’s not just that you have high cost and ROI, it’s just you can’t have a huge amount of trial and error. You have to get architectures right over time. If you made a significant mistake on your architecture, you’re going to pay for it, and it’s going to be very, very, very expensive.
Next, technical limitations. Very significant challenge. Robots can’t do everything under the sun. We know that. They have limitations in how they do things, the precision they have, what are they handling. There’s a lot of complexities around this. These machines. There’s a lot of complexity around it. There’s maintenance that needs to be done, make sure that it operates with the precision that people expect from it and all of that. That’s the third area of challenge in this space.
What’s interesting is that some of these technical holders have been solved by some of the industries. If you take EVs, for instance, we have seen the rise of the electric motor and more and more high-performance electric motors. Some of this is going back to the robotic space. You could argue that a lot of the current stage innovation in robots is half. Yes, better AI, better vision, specifically, better understanding of the environment, but I believe it’s also around motors.
Motors are becoming much better at doing the job. If you want to have a smaller robot, if you want to have a human-size robot, if you want to have hands that can do something, then yes, electric motors, actuators have changed what’s possible.
That’s a little bit of the key issues that are more on the core technical business model side. Then, if we move to other areas, the robotic space we’ve already mentioned has a huge issue around interoperability and standards, with ROS, Robot Operating System being one of the most common used protocols/robot operating system thing.
But honestly, the reality is there’s very little integration between systems. There’s a lot of pervasive systems out there that are dominated by the suppliers themselves and the owners of the hardware. There hasn’t been a huge focus on creating a platform that truly works across the world and across different domains as I mentioned before. Interoperability and standards are still very much an issue.
We just need robots that talk and listen and LLMs in the middle and problem solved.
We just need robots that do everything, yes and then it doesn’t matter.
I think we are going there. The reality is that the more you can naturally command a robot and ask it to do a task and have it indeed do this task without too much detail or at least the similar level of details to a human trained at a specific level, I think the more things will change versus having a team of programmers to program the robots and any time the stuff change on the other side, you have to reprogram the robot. That was a big part of the cost of running a robot in a factory.
Two last but not the least topics. Maybe we’ll bundle them a little bit together. The regulatory and safety, regulations that occur from market to market and how this is done, the regulatory framework by which safety is managed, and public acceptance in general, and supply chain dependence. It’s very clear that you need advanced components from all over the world. Going through a world right now where there’s all this geopolitical instability, and there’s tariffs, and there’s all this stuff, how does that work? That is a bit of a creating also encumbrance on development of B2B robotics.
As we said, China has seen a rapid adoption of robotics and that might surprise some because thinking about China, low-cost countries and stuff. But as you said, the reality is that China is a decreasing population, especially a population that can work, makes sense to automate. They also wanted to go more high-tech type of production. China is definitely big versus Japan, Europe, and US, where definitely there have been traditional struggles of automation. But as we just said, China is probably leading at this stage.
Just to be clear, the US is a big robotics market. If you’re listening to us, the US has to catch up. Maybe there’s potential to be growing, but the US from what we know, had 37,587 industrial robots installed in 2023. It’s a significant market, in particular around professional service robots. Some of the leading companies in the world. It remains hopefully a player in that space.
If some of the big brands we mentioned are not here, they’re in Asia, some of them are more in Europe, and now they’re owned by Chinese, as you were mentioning, Bertrandt. There’s still some stuff happening in Europe that’s significant, but I would say Asia has taken the lead in some ways across the board. China catching up now and going to the next level. Japan was one of the big, big first players a lot because of demographic pressure. South Korea has caught up as well. There’s a lot of interesting things happening in Asia in particular.
I still remember years ago, there was really a question that your… Cheap people working in factories was a better deal in many, many situations than having robots. I remember stories about highly automated South Korean companies having trouble to compete versus actually cheap labour from China, but I think all of this has changed.
Indeed. Shall we move to consumer robotics, which is probably what a lot of people, “When will we have this robot serving me at home?” It’s like, “Well, probably not anytime soon.” There’s good news and bad news, but probably not anytime soon. It’s growing fast. To be very honest, a lot of the home robots that we’ve had, one’s I think have been misclassified, which is the whole use of things in the kitchen, et cetera. Those things are not robots, just let’s be clear. Those things are very basic mechanical elements compared to what a robot needs to do.
In the household, the big dominant player that came into the market was Rumba. They reached a ridiculous amount of smart vacuum shipments as they say in the low tens of millions around 20 million units and all that stuff and then nothing. I don’t know if there’s anything public or not, so I’m not even sure if I’m talking out of place, but I heard that they’re not so much in great shape.
There’s how many rumours in the market that they are doing very bad?
I used it at the beginning and I really disliked it. It’s just like the experience wasn’t great. I think it’s a great experience if it’s a very contained environment, but if it’s, for example, if you have pets, it’s a mess. I just gave up on it. I tried twice and I just gave up. With the iRobot, just to be clear, a lot of people know the Rumba, but iRobot is the name of the company.
Same for me. I tried it very early on, maybe the first-gen. It didn’t work that much, to be frank. I think that’s what has been the issue. It’s very sad because it was definitely… I think the best-selling robot in the world by unit’s shipment of any robot. I don’t think there was any robot, autonomous, shipping at similar volume than iRobot. I can say that personally, not inside the house, but outside. I just bought two robots from Navimo to do my lawn, front yard, backyard. It has been working quite well, actually.
I must say I’ve been pretty happy with that because instead of having a gardener coming every two weeks to do the job, you have the garden clean every two days or every day, if you want. That has been a quite interesting solution, but obviously, you always have the question, how much is a robot? How much are you paying your gardener? How long is it taking? How long will the robot keep working? You have location issues. Usually, that’s the biggest issue with robots. You need to know where they are to know how to keep doing their job. It’s not totally problem solved, but it’s getting better.
It is getting better, but we just wish we had a vacuum cleaner for our house and a lawn mower for them whatever and it’s still not there. It’s like, surely these are the easier use cases. Then we move to the other use case, which we all want, the personal assistant, social robots, like the toy that plays with me, that interacts me, embodies Moxie. We had this… Was it Jibo? Jibo, the social robot thingy, whatever. Obviously, aibo, Sony’s aibo robotic dog, which we’ve seen a couple of iterations on. We all want that stuff.
I bought the first one 25 years ago now.
I want that stuff. Some of us grew up watching Battlestar Galactica, the previous one, not the dark one that came after, but the original one. We all wanted those little dogs, the little dogs that were in Battlestar Galactica. I wanted the dog since I was little. I wanted that dog. I wanted the dog that does stuff for me that I don’t need to train.
I can just tell you that my aibo was for a very long time my worst spend in tech gadgets I have ever had for at least a decade or two. I don’t know.
How much did it cost, do you remember?
It was very expensive. It was very expensive for the time. A few thousand euros, which 25 years ago was a pretty insane amount. For something that was, ultimately at the time, truly useless. I remember playing a few weeks with it. I think that’s a big issue. How do you make it useful while keeping cost reasonable? How do you make it so that it’s not just a fad where everyone got excited because that’s a gift for Christmas everyone gets to have, but next year no one cares?
I think there are two key issues around personal assistance and social robots. I’ve seen some really beautiful solutions. Things that look like it’s a toy. You know it’s a toy, but they’re really engaging with you. The ones that look the best and are the best are obviously, as I mentioned before, in the case of B2B robotics, over-engineered. They have a bunch of sensors, there are a bunch of things going on, and that makes them expensive. It means their bill of materials, what makes up for the cost of the hardware, is actually very expensive. Therefore, you can’t sell it at a very low cost.
I haven’t seen anyone at scale try to play around with robotics as a service for a consumer, where you have something, and you’re renting it out thing. That might be a play around that, but it feels to me that there has to be something around that. Because I’ve seen some solutions that are really endearing. You want to interact with a robot, et cetera.
That’s problem number one. Problem number two is to your point, I think maybe, Bertrand, I’m putting words in your mouth, is the level of engagement you get from it. It’s a novelty, you interact a little bit, and then what else does this thing do for me? If it doesn’t do anything else for me, then I’m like, okay, cool. I move on. There’s no interest. It has to have one, ability to be updated, to go to the next level, to become a better solution, and two, to actually do something with me that is of tremendous value to me systematically. If not, then why would I do it? Why would I buy and consume something that’s likely to be relatively expensive for me to have in the home?
There is always the question, you are fighting against all the stuff you can buy. You are fighting against buying a computer, buying a tablet, buying a phone, a TV. In some ways, you have the question, if the hardware side of the robot is actually not working that well, what are you left with? You are left with an Alexa where you can talk to that has some eyes. What will it do based on its eye? What action will it take that benefit from being equipped with two cameras, for instance? It can navigate your home, but what will it do by navigating your home?
Next question, can it hold something? If it’s small, it won’t be able. It will have to be bigger. It will have to be heavier and balanced. Then it’s a whole different ball game. Then we go back to cost. We will talk maybe more about it.
But if you look at Tesla and the robots they are working on for consumers, we are talking about the price of a car. They are talking about 20 to 30K. That might be the price so that you have a robot that indeed, practically can do things in your home. As what it takes to open something, close something, go up the stairs, go down the stairs, move some books, move some food, take a kid. It’s not an easy one. What do you do with a small size robot, like the size of a stuffy so that it’s cheap, but ultimately it’s useful? It’s not clear.
Innovation will come, I believe, but it’s not clear. I’m not sure we’re going to have Jeeves anytime soon as an anthropomorphic robot.
Of course. I’m not saying it won’t come. I was mostly in a mindset of the next 5 years.
It’s a little bit of a segue. We’ve been talking about personal assistance and social robots, entertainment and educational robots are in the same logic, SoftBank’s NAO robot. Things that have been used for education. We had Sphero. Everyone probably remembers Sphero. Unfortunately, the toy space needs to be even cheaper. If it’s a social thing to be doing stuff with your children, then that’s one thing. If it’s something that’s been there just as a toy or as an entertainment to play, it needs to be even cheaper. I don’t know.
It’s a space, again, I think the comments you just made, Bertrand, I’d make them my own. Very, very difficult to justify the cost unless there’s a clear beachhead in terms of the actions that the robot can do for the child or for the person that it’s interacting with.
I think what we see in the meantime, if you take Sphero, for instance, many of them are moving to the education market. Luckily, there is a big push for STEM in schools. As a result, my impression is that they realize that, hey, most parents have no clue. It’s not as if they are going to teach robotics to their kids. If we want to do stuff that is relatively cheap while interesting, let’s make them educational. Let’s have schools buy them. Let’s have labs with our robots. My impression is that a lot of what could have been from a robot actually end up being education for a STEM education type of robot that basically belongs to the classroom or the lab.
It feels to me, it’s a little bit like, let’s move business model. Let’s not sell to consumers because they’re price sensitive, and let’s sell to educational entities, academic entities, et cetera, who are not price sensitive. I’m like, you’re moving from price sensitive to price sensitive. You’re just saying, well, but the other guys have more money because they’re an institution. I’m like, yes, they do. But you have to share these across different people, et cetera. Different wear and tear as well.
But that’s the thing. You can share one robot for 30 kids versus 2.
I understand, but still, it is a very, very, very price-sensitive market as well. It’s very difficult to break into. If you look at the US, for example, unless you’re going to higher education, it’s going to be very difficult to crack into a lot of these schools. Very difficult. Even if you’re going to private schools, super price sensitive. I don’t know. I feel it’s an interesting pivot on business model that I have a lot of doubts will scale.
Maybe last but not the least, another demographic that obviously matters tremendously is the senior demographic, elder care, and assistive robots. When we know so many people don’t have someone that can take care of them and look over them. Even in assisted living, there’s lack and shortage of employees. Even if you are living in a home or assisted living facility, having these kinds of robots is particularly critical.
I know for a fact that the Japanese have tried a lot of things out there. I’m not sure if anything has scaled that we could talk about, but definitely for obvious reasons, again, around demographics, they’ve had to do it. The world is moving this way. We’re having more and more ageing population at scale. So having some elder care or assistance robots seems like a must. Again, if you have vertical solutions, in particular around assisted living, homes, et cetera. Now, one would say if you have personal assistance, personal social assistance, et cetera, that work across the board, that should take care of everyone. But as we’ve seen vertical solutions are the ones that emerge first. That’s probably where we are in the market at this stage.
Yes. I think that’s the right way to go. First you start with very specialized robots in terms of hardware and software. That’s your only option because the cost of everything, from the chips to the mechanics, is just very expensive. You have to pick something that you can solve where people find an ROI. But I think step by step, there will be a convergence somewhere to more standard models.
That’s why some are very bullish on humanoid robots, because the idea with a humanoid robot is that it can work anywhere a human can. It can do any task, at least physically, as that a human can, because it will have the same size, it will have similar hands, it has a similar balancing and walking mechanism. If a human can do some tasks there, physically, a humanoid robot can. I think there is some very strong logic.
Obviously, the negative is that it’s big, it’s heavy, so it will get expensive. But at the same time, if you move from, I can do just one task for cheap, or I can do 100 tasks way more expensive, maybe that would be the right combo.
Exactly. Really finding that sweet spot then leads to great unit economics and scalability. Talking, maybe moving a little bit into the key trends that we’ve seen in the consumer robots or robotic space. Really, the advancements are trying to make these more capable, affordable, user-friendly, as we discussed. First one, again, very obvious AI and the integration within put and output mechanisms like voice and others.
Cameras, vision, and all these things. These elements are particularly critical to what we’re seeing. Obviously, ability to recognize owners versus strangers, ability to generate speech back to people, to interact. I think the big part of engagement in all of that is also related to that. No surprises there. Obviously, one of the key trends that we’re seeing is all of the advancements that we’re seeing in AI and input and output.
Yeah, 100%. I’m a big fan of the input/output problem. That’s the problem we need to solve. If we cannot communicate easily with a robot, forget it. Forget it at home. Maybe we have seen in industries, factories, it’s okay, you might pay developers. But even then, I think the big scale will come when all of this become easier to communicate with and robots able to act on this.
Indeed. The second area is obviously around sensors and mobility, which enables a lot of these things. The ability for some of these robots to move around and the ability for them to have sensors that detect the environment around them and to have actuators to actually interact with that environment. It could be something that leads to a hand moving, to something moving, to something touching someone. Good luck on that.
Obviously, there are a lot of things happening in that space that are quite interesting and have led to tremendous innovations that are still being implemented in the market. In my perspective, it’s an area where we’ll need to deliver, deliver, deliver, deliver until we can have solutions to your point that are at scale, much cheaper, and then can be propelled into consumer robotics. It’s one of those things. It will ripple down the value chain to finally get to consumer robotics.
I agree with you. The same way that not everyone could get a car at first started with trains, actually, and then we went smaller and smaller. Definitely, I think some of this. Not just that, let’s not forget a lot of the robots we see today in demonstration of how agile they are and what they can do, actually are remotely operated.
People have to know that in many cases, this stuff is not real in the sense of a robot being somewhat autonomous. It’s just remotely operated. I was quite shocked how many companies, Boston Dynamics, for instance, are real pro of remote operating their robot and making you believe that… They won’t lie. They would just imply that this stuff is autonomous. No, it’s not in many situations, so nothing. We are not getting there because if you don’t have the hardware that is agile and has a power to do what it’s supposed to do, of course, you are dead in the water. But I think we’re not yet fully there in the autonomy of this robot, but it’s coming.
Maybe to highlight three more trends that we see in the space. Obviously, there’s the notion of emotional AI and AI interacting with people, reading their emotions, reading their things. It’s another field of AI that we’ve already mentioned. The aesthetics and form factors of robots are essential. In some ways, if you look more anthropomorphic, or if you look more like either pets or humans, et cetera, it’s easier to interact with them. Software designing versus more hardcore designing, friendlier design. It’s probably one of the key things that will lead to that attachment.
Last but not the least, obviously, the realm of connectivity and integration with smart home facilities. Maybe as a proxy to avoid precisely the problem we were just talking about, because some of these sensors are very expensive to have in the robot itself, but could the robot use other stuff that’s already in the house? Do you have cameras in the house? Do you have a TV in the house? Do you have whatever? I can interact with that.
Again, on that stuff, the more multimodal stuff beyond just connectivity, beyond me just connecting to Wi-Fi, the multi-modal stuff, I think, is probably the one that I’m the most sceptical about. If we think it’s going to take decades for us to have some of the stuff in our home working at scale, the multimodality of it is going to be even more difficult because it assumes interoperability, it assumes that the systems talk to each other, which we know doesn’t happen today. They don’t talk to each other. They’re all specific to the manufacturers, to the owners, in many cases of the hardware, as we mentioned before.
I’m somewhat hopeful for the integration because even today, we can see you have apps to integrate, and you build up step by step, pieces by pieces, your smart home. There are standards, so there are ways. It may take time. Of course, there’s no real unifying effort. That’s true that the humanoid size robot at home is not there in 5 years, probably not 10. But at the same time, I can see that it’s coming. Maybe it’s 15 or 20 years that it’s really practical. But the tech is slowly… I would say slowly, but now it’s getting there faster, is my personal perspective.
Yes, it will take time to really do useful stuff, to be cheaper so that people can afford. But at the same time, I think it’s important to share my belief that it’s getting there. Yes, we are complaining because it’s not yet there, but it’s moving in that direction.
Maybe switching for a rapid fire on preferences and cultural factors that we see around the world. Obviously, the US, not very shockingly, has a strong preference for utility-based robots, so things that do stuff for me, like vacuum lawn mowers. Obviously, the US, we also have more space, so we have homes and fields. It’s very different than living, for example, in Tokyo.
Less focused on emotional AI, maybe.
I think we should have more emotional AI in the US. Maybe the vacuum cleaner can also talk to me.
Maybe I can pet the vacuum cleaner.
The vacuum cleaner, yeah, come here. United States is, as you would imagine, probably one of the most utilitarian in that sense. We want robots to do stuff for us so that we can then just chill by the TV and ask the robot to pick us a beer and do stuff for us. That’s probably the best way to look at it. The opposite is a little bit, I think, Japan. Japan is more, in some ways, even embracing robots, maybe too much. We know that some funky dynamics in Japan where some robots are treated like pets and characters and whatever. By the way, if any of our AI overlords are listening to me, I love robots. I love robots, please don’t kill me.
But obviously, there are a lot of things happening in Japan that have to do with ageing society, a society where human connectivity has had some challenges because of demographics, because of how cities have been built, because of cultural aspects of Japanese. In those environments, robots are really more tailored towards personal societal links, pets, characters, all that stuff. That’s maybe the most extreme opposite. Then coming a little bit to the middle, we have South Korea, which is both a mix of utility, and we also have some social interaction.
Yes. Of course, Europe will focus on data privacy and safety and will regulate its way out of robots.
No robots for Europeans. As we discussed, I think China is definitely very technologically optimist, both on the industrial side and the personal side. I remember there was in China at some point, I forgot the brand, a brand of robots for kids, educational, that was doing very, very, very well. I was very surprised, and that was years ago.
Yeah. Maybe moving to forward-looking and what’s happening in the market. A mix of what’s happening in the market that is structural and what’s happening in the market as we look forward. We’ve already looked at a lot of these partnerships, already mentioned them, not just acquisitions, but also what happened with Amazon and iRobot. The deal got scrapped in the end. As we know now, iRobot, who are the makers of Roomba, may actually falter now. We’ll see, but Amazon was very active early on, in particular in the B2B space. A couple of things also on the consumer side.
We’ve seen other things happening a lot in the B2B space in terms of acquisitions. We’ve already mentioned several of these examples. Alphabet has come back into the market to do things around particularly the operating system, the stack. Obviously, they acquired a bunch of things like Open Robotics, which maintained ROS, the robot operating system, and Vicarious. They’ve done a lot of movements that are a little bit more connected to the software platform side with operating systems.
NVIDIA has done a bunch of things that we’ve talked about. I know you are a big believer that NVIDIA will become one of our overlords. Bertrand, do you want to talk through some of those?
Yeah, I don’t know if I would put it this way, but of course, I welcome our new overlord if he listens to us in the future. But look, I think NVIDIA is realistic that they need to keep looking for new paths for growth, and the robotics is a perfect match for what they do. You have vision issues, you have audio issues, you have communication issues, you have mobility issues. It’s a perfect match for the NVIDIA stack.
I think for them to propose solution on hardware, on software, on chips makes a lot of sense because it’s going to be a big market. Maybe we don’t talk much about that, but it’s clear if you are thinking 10, 20, 30 years from now, it’s one of these big markets that’s going to change the world.
Don’t let the fact that we didn’t start at that point… That robotics is going to be huge. Robotics is a next era of AI in some ways. I don’t think we’re there yet. I don’t think we’re at the level of rapid disruption that we faced in AI more on the software layer. But robotics, we will get there. There’s no doubt we will get there. It will be, again, a major seismic shift. It’s going to be one of the next big industrial revolutions or technology revolutions, if we want to call it as such.
To maybe go actually as a segue into that, obviously, we could talk about partnerships, we could talk about all the money from government that’s going into a lot of these activities. But maybe looking a little bit ahead in the next 5 years, maybe even beyond that, what do we see as happening?
The first piece that everyone’s particularly excited about is this convergence of AI software, so to speak, in robotics. It’s when AI comes into robots that can do things with that, that they can do manipulation of pieces, interaction with people, interaction with other objects, et cetera, in a very meaningful way. There the world just takes over. In particular in a world where the robots can then do things that they haven’t trained for, that they haven’t done any piece of supervised learning or unsupervised learning, that they can go to the next level in really framing how they interact with their environments and with the world. That’s a big deal.
Once we start having that conversion, it’s accelerated. It’s also the part that concerns us. Just to be clear, this is the part that if we get to a convergence of AI and robotics super quickly, we should become a little bit more concerned about what they do because then they have physicality. They do have sort of a brain. At that point in time, we should worry a little bit more than we do today.
Definitely. That’s why AI progress has been fast in the pure software side now. AI hardware, robots basically, are downstream of this trend. As you say, that’s true, that it’s a different story when suddenly you have a body, you are more or less independent because I guess that most robots would be designed to be really some edge computing. I would be surprised if robots are designed in a different way.
Not saying they won’t be upgradable through the internet, not saying they won’t be able to look up for something on the internet for some reason, either looking for information or looking to increase capacity on the spot. But I would be shocked if anything is not designed to be fully working locally. You need to work around a network disruption, you need to work around internet disruption. I think that’s a good example.
That’s a very good point. I mean, if we’re looking at humanoid and general purpose robots, they definitely need to have most of their sensorial, their senses’ interaction with the world—be it the sensors or the actuators—controlled by the edge. I don’t see at scale how they can just be communicating and all of a sudden there’s no network. What happens, right? What state do they go to?
It’s like smart cars. If you remember in the past, there was the idea that self-driving cars would be remote controlled. I mean, no way. It has to be fully automated. Maybe it has to go back to the human owner in some ways for some needs, but ultimately there is a need for real local autonomy.
I think it’s funny that you mentioned that because we didn’t talk about self-driving cars. But for me, self-driving cars are a robot. I mean, they fall under the definition of a robot. Machine capable of carrying out a complex series of actions automatically, especially one programmable by a computer. They’re our transporters in some ways. I do think that’s going to be one of the big ones that will be available to us maybe within a decade.
I didn’t talk about self-driving cars as robots because they have another name, cars. But you’re right, they are 100% robots. Again as we said, you could argue actually they are at the forefront of everything. They have been at the forefront of electric motors that we need for any other type of robot. They are at the forefront of AI. The self-driving part has been the forefront of AI for 10 years, easily, way before LLM. Of course, people are willing to spend 30, 40, 50, 60K to buy their car, you have financing mechanism available for that.
The whole economy is designed to help you buy and finance a car in the first place. It’s definitely the reference. It’s good we end on this in some ways, because it’s a critical piece of the puzzle. You can argue that not all, but many robots are downstream actually of the car innovation.
I feel that might become actually the first real full-on consumer robot that consumers buy and use.
It has a specific use case. It’s not trying to do groceries or take care of the kids, or this, or this, or that. No, it’s driving people. It might be driving stuff as well, but it’s driving on the road, going from point A to point B. So it helps.
If I was an automotive OEM, original equipment manufacturer, I would be thinking through the car and how the car becomes actually the next centre of innovation because the car itself becomes its own spot. If I have that amount of intelligence so to speak, AI so to speak, linked to the car moving around, which is the most complex activity, think about all the stuff I could do inside the car.
Honestly, the rocket science part is the part of the car actually moving around and doing all its things. I don’t know. I feel sometimes I have conversations with some automotive executives and people that are consultants to automotive executives, et cetera. I feel the industry is still very limited in how it thinks about self-driving cars and how that changes the world, et cetera. It really becomes key element of mobility interactions. If you don’t need to drive it, there’s a lot more space to do a bunch of other things, really.
I think that’s the reason why Tesla has been moving so fast, so hard on Optimus, their humanoid robot. A lot of things are very similar. Yes, the use case are less clear. It’s definitely not… We already have a robot, it’s a better robot, like cars. Where it’s a car. We have better car. The use case is less obvious for humanoid robots, but at the same time it’s clear that there will be a huge innovation from the car manufacturing, the self-driving cars at least, that are going to provide downstream innovation to especially humanoid robots, but not just. You can imagine other type of robots. Xiaomi, who recently launched cars, is also into robotics. So I would expect to see more and more of this actually going forward.
Maybe, obviously, just to name a couple of things that we do know will happen over the next few years. Obviously, B2B robotics will have a lot more applications across sectors, will go into the small and medium business arena. It goes without saying as well that we believe that the scaling in economics of robotics is going to be critical, that unit economics can only work at scale once you have materials sensors, actuators, et cetera, that are much, much, much cheaper coming into the market.
The effect of robots in society and in public discourse, technology, philosophy, are looking very deeply into it. How do we interact with robots? How do they interact with society? What’s their role? What rights do they have, et cetera. Regulation and standards, evolution, so all of that. Overall, I feel there’s a lot of topics that we didn’t go in detail on, but that you guys can probably get a little bit of a feel on where we stand on them just because of the description we had earlier.
Last but not the least, obviously we always talk about robots as something that is separate from us, but there’s a piece of robotics that is very focused on us and them which is obviously human augmentation. There’s obviously the discussions we already had around co-bots, collaborative robots, exoskeletons, and all is related to that. Which I think it does lead us to a brand-new world when we have the ability to complement ourselves with elements that are extraordinary to us, but then make us much more either productive or have skills that we wouldn’t otherwise have, et cetera. A lot of cool things, I think, will happen around augmentation. I think that’s more science fiction, to be very honest.
The reason why I think that’s more science fiction is because these elements need to interact with us. The moment you start interacting with humans, with the humans themselves, there’s a level of complexity that you need to deal with. Not just in terms of technology, but also in terms of regulation, standards, et cetera.
I don’t know if exoskeletons are a real robot in themselves, but at the same time I can see how it can become a huge, as we say, co-bot complement to a human. Either because you need a rehabilitation or because your work or your labour is very hard, or because simply you are too old, and that’s a big change. With an exoskeleton, suddenly you’re able to walk. Or normally, you don’t need a stick, you don’t need something else. That might help in many ways.
Let’s talk about societal impact. I mean, we’ll adapt unless the robots start killing us. I think that’s societal impact. That’s my sentence. It would be great. We will all be great unless they start killing us, then we won.
You have seen too much of Westworld or Terminator, I guess.
Or the movie iRobot, actually.
On a serious note, I think there’s a lot of complex implications of it because as robots became more and more humanoid, they become more and more… Look like us. For example, material innovation will lead us to have robots that might actually look just like us, not anytime soon, but in decades. What happens? What’s our interaction with these beings, these elements, et cetera? I do think that produces a very interesting dynamic societly. We’re already starting seeing things like that. If you want to see at complexities of society, look at Japan to see what’s happening already.
But can you get married to an avatar? Can you get married to a robot? How does that work? Do they have rights? Do robots have rights? What are the rights of robots? Will we have a constitution for robots? At some point we say, actually, the robot can become citizens. How do they become citizens? All of that stuff is like a minefield. I don’t know. I feel we still have much to go, and it will really depend on the implementations that we have of robots that scale.
In all honesty, in particular on the consumer side, once we have robotics on the consumer side that scale and start becoming more and more anthropomorphic. More and more like us, we will have to deal with, based on those use cases, what are the regulations we need to think about? What are the issues we’re having now with these machines as they are scaling in the market. That’s my two sense on that.
I was wondering, we cannot do an episode on robotics without talking about the three laws of robotics from Isaac Asimov.
Yes, we have to talk about Isaac Asimov.
Isaac Asimov, a famous science fiction writer, has written extensively about robots. I actually think he’s the inventor of the term robotics more than now, 80 years ago. He built some laws for robots, that robots are supposed to follow. First law is a robot must not harm a human or allow a human to be harmed through inaction. Second law, a robot must obey human orders unless his orders conflict with the first law. Third law, a robot must protect its own existence unless that protection conflicts with the first or second law. From these three laws, Asimov managed to write a lot of sci-fi explaining ultimately what it means in real-world application to follow these laws.
I think we sometimes forget that. We always say, “The science and technology of people are changing the world.” I mean, in some ways that’s true. In other ways, actually, there’s other things.
It’s like writers change the world, people that do movies change the world. A lot of the things we’re seeing implemented today in our tech realm in the future, both in terms of AI and in terms of robotics, for example, just to quote two very big visible spaces, we are imagined by people decades ago. 80 years ago, 70 years ago, 60 years ago. So the imagination was already there. And in some ways, we’re now implementing them, which is interesting.
I feel sad for some people like Isaac Asimov who didn’t live long enough to see the conclusion and the execution on some of his dreams. To conclude, B2B Robotics is scaling rapidly, fuelled by labour needs, AI breakthrough, cost declines, but also geopolitics. Consumer robotics is more fragmented, but has had a breakout success in cleaning robots, with social robots on some cusp of border acceptance. I think we should keep an open mind on how robot can transform the workplace and the everyday life. It’s an exciting moment. Robots are shifting from the realm of sci-fi to practical reality, and we reshape industries and how we live. Thank you, Nuno.
