speaker-0: Hi everyone, welcome to the Robotics Stack Podcast episode number seven. My name is Leo and I am live from Paris.

speaker-1: My name is Oscar, I'm live from Berlin.

speaker-0: So today we are first going to talk about some lesser known publicly traded robotics company. Then we discuss Anjuril, the Defense Tech Company, who recently released a very advanced piece of software for industrial automation. After that, we are getting into Terminator-looking humanoids robots, the T eight hundred. You have probably seen videos of it in the past few weeks. And finally, we are going to discuss a robotics hand startup from San Francisco that got sued by Tesla. Stay tuned for that.

speaker-1: So let's jump right into the stocks because I've seen that's a very prominent search term on YouTube. People seem to search a lot for robotic stocks. Of course, the disclaimer: this is no investment advice. If you lose all your money, that's up to you. But it's worth digging a bit because there are quite a few companies in the robotic space. I have five in total, and I will always give you the pros and cons. And the first one. Is Stereo Texas traded on the New York Stock Exchange, and they are a pioneer in robotic magnetic navigation technology primarily used for cardiac electrophysiology. Sounds like a mouthful, but it's actually very helpful technology helping in the treatment of rhythm disorders of the heart that basically allows doctors to perform more complex procedures with better precision and less radiation exposure by remotely steering. Catheters using magnetic fields. So super advanced technology, and their key product is something called the Genesis System, which is like a robotic navigation platform for these catheters. And they also have the compatible catheters, their own brand called Magic Sweep for high density mapping. So very interesting that you not only see robotics in the humanoid space. Most of us would probably not consider this a robot on the first side, but these are actually robotic technologies. So what do we have about the pros of this specific company? Obviously, they have a strong clinical advantage in safety and precision, and it's a minimally invasive heart procedure. So it's not a big procedure in that sense. They have a potential for recurring revenue from disposals, for example, or service contracts. They are positioned in a growing field. I mean we've seen other robotic assisted surgeries, even Neuralink could be considered as one. But that is obviously also a con at the same time because there is big competition from larger players in the broader surgical robotics market. Of course, their revenue has been somewhat volatile with the recent earnings, and they also have a pretty small market cap that makes it very sensitive to news and funding needs. Now this is company number one. Number two is I mean Leo, you can always chime in if you have a question.

speaker-0: No question about that. I think what I like about this one is that it's in the medical field. So it means there has been FDA approval processes and all of these very strict guidelines that they have to abide by. So if they are authorised to enter a market, it means they have figured reliability of automation at a very high level. Because reliability and repetitability of a task for now is one of the main issues for humanoids manufacturers, for example. And you know that in the medical field the guidelines and the norms are very strict on that. So it means they are pretty advanced on reliability, which is a pretty good signal to me.

speaker-1: Yeah, 100%. There's also another field which is growing in robotics, and that is safety, of course. And there's this company called NightScope. They're on the Nasdaq and they specialize in autonomous security and public safety robots. They were founded with the mission to reduce crime through technology. And the company deploys AI-enabled robots that patrol sites, detect anomalies, and provide real-time data. To security teams. It's basically like the neighborhood version of Palantir. They have some interesting products. You can see them on their website. These are all autonomous security robots for outdoor or indoor patrol, emergency communication kiosks, and AI power detection software. They even have some sort of a car. It's like a mobile security unit, basically, that drives around the neighborhood and it can read license plates. It can even do, I'm not sure if this one's. in particular can do it, but they even have products that have built-in infrared or heat cameras and stuff like that. It's not like the average camera from Amazon that you order, Blink or what is it called. So this is very much more advanced. You can see it on construction sites. For example, I think in in Berlin I've seen a couple construction sites that have these poles and at the top you see this like sensor system. And this is the kind of stuff that they also build. And of course, speaking about the pros here, they are addressing a chronic labor shortage in the security sector while also offering 24-7 operation and data collection. They get data from this as well. I'm not sure about the specific laws that apply here, but it is a potential revenue stream, of course. And many people they don't want to work in these security jobs often because they are high paid sometimes, but they're also risky. They also require you to have night shifts, which is generally not. Something that a lot of people want to do. Their recovering revenue model is through leasing and monitoring services. So this also comes with a software suite. They have visible deployments in the real world already, like campuses and parking facilities, and they also have some maintenance service that you can get or that you probably have to get if you use them. Now on the cons side, it's a very small market cap still with ongoing profitability challenges. They have also regulatory and societal scrutiny around autonomous surveillance, of course, and data privacy. We know other countries are more pro this, so this can limit their ability to penetrate new markets. And of course, they also have competition from traditional security firms and other tech solutions. They don't have any humanoids. And to me, obviously, it also makes sense to have humanoids in the security sector, of course. Patrolling around, scanning stuff. You can do all of this with a humanoid. So this remains to be seen if their approach with other form factors is better or if it's worse. So yeah, that's night scope.

speaker-0: I think beyond labor shortage, I think it's also just a net increase in security overall because even with a large security teams they cannot cover, they can patrol around but they cannot cover everything. If you have automated sensors all around a very large construction site, it's always better to have more sensors everywhere than to have a blind spot somewhere. You know, it can be pretty dangerous for everyone. Because I've seen some company do construction site progress tracking. using drones, for example for very large solar panel farms. I think this could be applied to this company as well, right? Drone surveillance, basically have a drone going around over a very large site. Yeah yeah.

speaker-1: Absolutely. And they position themselves more like we give you the insights that your cameras don't see. They're actively saying like, okay, the camera is potentially not enough. So we have this car driving around or this little mobile platform driving around, checking all the angles and all the different spots that you might not see. So the next one is surf robotics. And this one focuses on AI-driven autonomous delivery robots. They're also driving around, but these are not for security, these are for delivery. And they're designed for specifically sidewalks rather than roads. Maybe you've seen them because they're quite popular. It's spun out from Uber, and the company aims to revolutionize the last-minute delivery with efficient and eco-friendly robots. So I think on their website they have the tagline: why do you need a Two-ton car to deliver a 500 gram burrito or something like that, you know? And it does make sense. Their key product is a compact, low-emission sidewalk delivery robot. Kinda looks like a little cart on four wheels, and it has, of course, advanced AI for navigation and obstacle avoidance. And it can also interact with a customer, and it is deployed actually already with a partnership of Uber Eats and also others. So this is a more known candidate here. And they have a couple of pros. For example, there's strong tailwinds from e-commerce and delivery demand, a lot of people ordering more. So these types of solutions have still a lot of potential to grow. People just get more and more lazy, unfortunately. Then we have Notable partnerships that they have. As I mentioned, Uber, but also 7-Eleven. They have tech backing, for example, from NVIDIA, and they are environmentally friendly, so they don't stink, you know, they don't produce any crazy emissions, and they can also reduce vehicle traffic by using the sidewalks. It's a pretty good solution. Now on the cons side, Again, regulatory hurdles and these vary by city. So some cities might have permits for sidewalk operation and some will not even allow it at all. At the same time, they're still in the early scaling phase and they're currently operating at losses. So no profit inside yet. And their operational challenges that they have are obviously weather, or I've seen these drive around certain areas and people would kick them, you know, there were thieves trying to get what's inside of the robot. Because the robot once it's out and autonomously driving, it's just exposed to all these outside factors, of course. So it's probably not something that you will see a lot in crowded urban environments in New York on a busy street or something. It's more for calm neighborhoods where it can safely drive on the sidewalks.

speaker-0: Yeah, that makes sense. I I've seen videos of them around. I think it's pretty straightforward.

speaker-1: Okay, then let's go to the next one, which is rich tech robotics. And they are in the Nasdaq and they develop a range of different robots. So service-oriented robots targeting hospitality, healthcare, also light industrial applications. ⁓ it's weird on their website they say something like 18 years of development, but at the same time I looked them up and it said they were founded in 2016. So I don't really know how that fits together, but they also have 11 patents, which is good. Their key products are more like bartender robots. I don't know if you've seen these. These also used to go quite viral. So a robot like handing out drinks or a waiter robot. I've actually seen it in a restaurant here in Germany. So it's more like a little platform that walks around and you can put on your drinks, order new drinks, and it will drive to you. Something like that. So delivery and hospitality robots as well. They have a cleaning robot and also newer industrial offerings like a pallet handling system. And they are also in early humanoid robot programs, although they don't have a humanoid yet. I've never heard of them before. I mean, even though they have been around for quite a while. And their pros, in my opinion, a broad portfolio. So they're very broad, right? They have a lot of different robots. They're not focused on one thing specifically. So they can attract different needs in different sectors. For example, restaurants, hotels, and warehouses. They're also expanding their product line. They have recent moves into warehouse automation. Potentially, there's faster adoption in these commercial settings compared to highly regulated areas, like we've seen it with the medical stuff or defense robotics. And they also have a high tier collaboration with Nvidia and already over 450 robots deployed. And that is at the same time a pro and a con because 450 is a lot, right? But we see other companies having deployed much more and producing much more robots. And then on the cons side, relatively small revenue base, and they still operate on net loss. They are in a crowded competitive landscape. Of course, they directly compete with humanoids, and that's why they want to get into this sector as well. And there are risks in scaling manufacturing sales and this customer support if they want to expand globally. But that of course applies to almost every robotics company. So that is rich tech robotics.

speaker-0: And yes, these robots have been around for a while. I think it's a pretty stable business. Even though they are not profitable, there has to be a path to revenue for them.

speaker-1: It has to, yeah. And the last one of the bunch is called Miyomo and they are traded at the New York Stock Exchange. This is kind of a wearable medical robotics company, so also something we didn't yet dug into on this podcast or on our other podcast as well. And they create powered orthosis, so kind of like braces to restore upper limb function. So they target patients who lost their arm, mobility. Due to stroke or spinal cord injury or neurological conditions. So this is again something in the medical sector. The key product is the Mayo Pro, and this is like a family of myoelectric controlled braces that detect muscle signals and then they amplify them to enable functional movement. Comparable to EMG that Meta is doing with their recent bands that they released. They also have a training app, which is always good, you know. They have Kind of like a product pipeline around it, and they also have support programs. Originally, this technology was developed by MIT, or not this technology, but this specific one was developed by MIT. And the first myoelectric prothesis was actually created in 1948 by Reinhold Reiter, a physics student in at Munich. But it's quite fascinating that you can use this type of electrical signals to help people. Now on the pro sites, Obviously, it's a very direct and meaningful impact for the patient's life because you improve their life basically immediately. You have improving insurance reimbursement pathways there. You have a growing addressable market due to the aging population again and rehabilitation needs. You have high gross margins once the devices are scaled. Least that's ⁓ kind of how it looks, and it's non-invasive, so no implants, no nothing. You attach it basically and then you're good to go. Now on the cons side, you have long sales cycles usually with these types of products. You don't just like go and say, Hey, look, it usually requires some education proving that this thing has value for the clinical evaluation because you need the payer approvals and stuff like that. They also have a pretty small market cap with ⁓ dependency on the healthcare system adoption, of course, and they competing with other exoskeleton companies and also rehab technology developers. Exoskeletons for me are quite interesting because essentially they are also robotics, but just more merged with the human together. So quite interesting.

speaker-0: Exoskeleton ⁓ reminds me of you know Boston Dynamics, the humanoids made for military. And you know, military is a pretty hot spot right now in tech funding. And I'm pretty sure you know about Anduril, the defense tech startups by Palmerlocky. Recently they reached a 60 billion valuation and they published a demo of an internal piece of software developed. And it's pretty interesting because it seems very advanced, but it's also only for internal use. So it's not really clear why they are communicating about it. But yeah, on June 17 they released a video demonstrating what they called Enduring's digital manufacturing backbone. And it's one entire software platform, kind of an ERP for enterprise resource planning. It's a kind of internal software you use at your company to manage all the resources. Track the stock, everything the company owns, the company owes, the goods they have in stock, the goods they have to deliver, and so on. So it's pretty much like a centralized database of everything your company has to do at a specific point in time. A company like Salesforce, for example, offers such software. So it's very enterprise-driven, very typical type of software for large companies. And Andrew. They don't do things you know like everybody else, so they develop their own software, which is something you don't want to do usually because it's incredibly complex and it takes a lot of resources. So basically their software covers design, computer-assisted design, supply chain, instruction for assembly in the factory, and also sustainment and delivery. So it covers the entire production chain and lifecycle of all their products. They also developed another software that should not be confused with, which is called Lattice. Lattice is a software developed by Angerville and sold to the military after that. Lattice is used in the battlefield pretty much. It's a software to plan, fight and use on the battlefield. And obviously it has different secrets behind it and everything. This one is called Arsenal OS and it's only for internal use. It's live already at Arsenal 1, which is the main ⁓ factory that Andrew owns. It's a 500 acres factory in Pikaway County, Ohio, and Andrew invested 900 million, close to 1 billion, on it. It has more than 4,000 people working on it. And they hope to increase it to 5,000 square feet at some point. And the demo video is very interesting because it demonstrates the software with one of their flagship products, which is the Fury. I don't know if you've heard of this one. Theory is like their main drone that is incredibly stylish and that they managed to ship 150 of these drones already. So it's like one every couple days, pretty much over one year. Obviously, it's a production rate smaller than what we could see from Russia and Ukraine, who use drone warfare quite a lot. But it's also a much more advanced and larger drones. So they saw that they needed like

speaker-1: Yeah, I think so.

speaker-0: an internal tracking tool. And if you look at the demonstration videos, it has like some kind of automation in tracking of advancement of production. Like every time someone picks up a part and moves it around and ⁓ works on it, the software will update kind of automatically. And ⁓ I think it's super interesting because when you got when you are a company that size, things like managing and tracking resources can sometimes slow you down even more than the actual production takes you when it comes to time. Right. And Angry Real says that they want to automate tens of thousands of systems per year at full scale, which is pretty ambitious. So far Angreel has delivered, but keep in mind that so far it's only automating one line that produces a hundred and fifty drones per year. We are not quite there yet, but slowly getting there for sure.

speaker-1: Okay, so so they're using it internally but they also want to sell it.

speaker-0: No, no no no. ⁓ they use it internally to automate the drone production line that is a hundred and fifty PCs per year. But they say eventually we are going to automate tens of thousands per year just internally because they don't only produce the fury drone, they have missiles, they have a bunch of other things. So eventually they pretty much want to apply it to all of the products. And they openly say it is not for external sales. But then it raises the question, if it's not for external sales, why are you making you know, so much communication around it, especially external communication. It's not a leak from an internal memo or something. It's like Andrew openly talking and publishing video about it. What I think is that maybe eventually they want to sell it or maybe they want their they want to develop a network of supplier who relies on it. And if your supplier relies on it, then you know exactly where the production line is at and when you are going to get your supplies. So maybe they want to integrate you know, everybody into that and they want to become some kind of Salesforce ERP for the defense industry. I think that's one vision. Again, that's my opinion, my analysis, not at all what Angie is saying. And another thing I think they want to do, and they're already doing, is that they want to communicate on the ability to scale their factory. Just like Tesla did. I don't remember if Tesla gave a name to its software, but Tesla talked a lot about Industrial automation host their production line was updating constantly for more efficiency. And even recently, Tesla said we snake ⁓ Optimus robots now. And they can turn their factory from one type of product to one other very quickly. And that's where their strength lies right now. And they are right about that. And it's exactly what China started as a movement that we see now happening in the US reindustrialization movement. We talked about that in previous episodes. In China, you see a company that makes, you know. vacuum robots and suddenly they get five billion of funding and they start making electric vehicles and after that they get another ten billion in funding and they start making humanoid robots. And when you see it from the outside you're like, this does not make any sense. But when it comes to what is actually their product and their skills, it actually does make sense. Their skill is not making a specific type of product. It's not expertise in cars or whatever. It's expertise in having an extremely flexible factory line. And I think that's what NGRIL wants to do, just like Tesla has already demonstrated in the US as well, right? We have to keep in mind that NGRIL is literally abiding by the startup motto move fast and break things, right? Andrew had many public issues in their product, especially in the battlefield. I mean it's at the expense of a few issues, but it's also on the plus side, it takes them like less than three months to ship a new missile or a new drone to a live battlefield somewhere. Something that would take five years to ⁓ legacy defense industry, right? So of course you go faster, but you are not as relevant. It just makes sense. And overall this piece of software is probably an investment into long-term relevant for them as well, right? I really think they want to become the de facto software production line for the flexible high reactivity defense industry in the US. Because again when you think about it, it doesn't make sense. for them to communicate on that if it's only an internal tool, right?

speaker-1: Yeah. And it also reminds me a little bit from Prometheus. I'm not sure if it's comparable in any sense, but ⁓ we talked about it in last episodes. I don't know if if if you think it's somewhat comparable.

speaker-0: I don't think it's exactly comparable, but I think Prometheus will develop its own software similar to this one, for sure. ⁓ what Prometheus wants to solve is the part that is from design to computer assisted production, because this part takes a lot of manpower hours, you know, especially from engineers and technicians. And this is the part Prometheus wants to automate. But then it only makes sense for them because we know we covered that already, Jeff Bezos is also acquiring through ⁓ holding fund, many legacy factories and industrial companies in the US. So if you want to apply your Prometheus like production chain everywhere, you need standardization. And you're not going to use Salesforce for that. And if Andrew is not selling its software, Prometheus will have to develop its own. And I think it makes sense because ⁓ Andrew is in the defense industry. And they will remain in the defence industry. They have shown no signs of going elsewhere, they have no reason to. So I'm thinking Prometheus will develop its own equivalent piece of software for their internal use as

speaker-1: Yeah. I mean to me all these feel like we are also heading towards humanoid robots on the one hand, but also these automated factories on the other hand. I mean these are the early signs, but I think along the line, you know, these systems beat something like Prometheus or this system from Underreal that they use internally to automate their production line. These are to me like the early signs that manufacturing and production and factories They are basically becoming these giant self reproducing things, right? That just like run twenty four seven where everything is automated, no human potentially in place anymore, just overheading things, yeah.

speaker-0: Yeah yes and no. Yes and no. ⁓ I think what we say in this field is the factory is a product. You know, the ability of the factory to be flexible is a product pretty much. And I'm not sure for example if you take the Tesla Optimus factory, many humans work in this and there is not necessarily a lot of automation, but the ability to change a production process is i incredibly fast and valuable. You know? And that's what they need right now. Either for example if you manufacture hands that we talked about earlier and you want to change something in the hands design, it doesn't take you three months to change the production line, it takes you one day. And that's where the strength really lies eventually, I think. Yeah.

speaker-1: mean I'm not sure. I think Tesla is already automating quite a bit. I saw they have like the Giga Press, have you seen this? Like where they print or they adjust this chassis. They they definitely have some ⁓ automation going on.

speaker-0: Yeah, I mean ⁓ it's again very surprising to me that they are communicating on that. And we talked a lot about companies trying to set standards in robotics lately. And I think that's pretty much what they want to do here. We maybe will see a company doing the same for robotics applications, right? That if you manufacture robots and robots always need the same parts, ball bearings, actuators, ⁓ whatever, tendons if your hands has tendons. And maybe we will see some kind of production line tracking software mostly for robots. I think it's a possibility. And I think it will also probably come from China, but that's only my opinion.

speaker-1: Yeah, China always the keyword. Most mentioned keyword in this podcast, I believe.

speaker-0: ⁓ I am the China maxer of the podcast, so of course.

speaker-1: Yeah, I mean if you're ready, I have another China topic for you if you want. Okay, so the next one is engine AI, and that is obviously a Chinese company. They're also known as Chongqing Robotics. And ⁓ I can actually use my Chinese skills. I think Chong means middle because Chongoren means Chinese, land of the middle.

speaker-0: ⁓ yeah. Go ahead.

speaker-1: Or the people from the land of the middle, but anyway. Engine AI was founded in 2023 in Shenzhen. So another company that was just founded three years ago. Yeah, we had in the last episode we had AGBot three years ago, having tons of different products, software and hardware. And now we have another one just founded three years ago by Zhao Tongyang, who previously worked at XPeng. XPeng You know, the cars of XPeng, very advanced cars. I think they have the most advanced autonomous full technology stack integrated into cars. And they also have some humanoid efforts, I believe. And he worked there. And then, yeah, in this very short time frame, he became one of China's fastest rising players with his company through their aggressive full-stack development. They design their own joint modules, they have motion algorithms. They have embodied AI software and they also have their manufacturing processes. A couple of milestones from their past. I mean, they're not that old, but they already have quite a bit to show for. They did a really rapid funding, a cumulative 1 billion Chinese yen, and that is around 142 million dollars by late 2025. And this one was followed by another 200 million Series B in April 2026. And now their valuation is over 10 billion Chinese yen, which is around 1.5 billion dollars. So just in three years, you can get to this place and robotics. It's insane. Then in June 2026, they confidentially filed for a IPO. In Hong Kong, and they're working with the CICC. So of course, I did mention some small companies already being enlisted on the stock exchange. And yes, China's also coming for the stock exchange. Yes, it might be in Hong Kong. Yes, it might be a bit different, but still they're going for the stock market. And yeah, they're working with the CICC and the CITIC securities for that. And they're not very much details about the size and the timing. This is all still pending, but it's definitely a part of this wave of Chinese robotics companies going public. Their manufacturing scale up is also pretty remarkable. They have a new Shenzhen intelligent base. They launched this just a couple months ago in May. And this one will enable or is already enabling high volume production. And they also plan additional facilities, for example, in Henan and Cheng Chu to support their broader scaling. So we've seen it with these other robotics companies. Mass production is one of the key milestones that they reach. I'm not sure if they have already gotten there, but it's just a matter of time, I believe, for them. In February, they also launched something called the Ultimate Robot Knockout Legend. So the URKL. And this is the fighting. So this is maybe where a lot of people on the World Wide Web have seen this robot or one of the flagship robots. Because it is actually pretty strong when it comes to fighting skills. You know, a lot of different companies they show the dancing from Unitry, all this kind of stuff. And they have a more straightforward approach. They are saying, okay, we are making our own fighting league, and they're showing like low kicks, roundhouse kicks, boxing combinations, and hooks and stuff like that performed by the robots. And they're really serious about it. And honestly, to me it's highly entertaining, and I'm I can't wait to see the first fights of this league because they actually have a quite a large price pool, so 1.4 million for the championship belt. And they kinda want to use this also to stress test their hardware, of course, and their algorithms. And if you are watching right now and you have a robotics company and you think fighting is something that your robot is doing very good at, then you might want to check it out. You can register right now because the global registration is open. And then also their robot which you've most likely seen with these kicks is the T eight hundred. You see it on their website. You know, they had some kind of a shitstorm where people say it's CGI. And to be honest, some of their videos look very, very CGI. Either it's some weird cinematic effect that they put on there or their CGI. And honestly, I don't know why they keep these videos open because even if it's not CGI, you know, you don't want to have this type of creative on there. That is potentially interpreted by people as CGI.

speaker-0: Yeah, I've seen videos like this. Are you sure they were posted by Ngin AI themselves?

speaker-1: Yes. You can go on the website. I've ⁓ the main videos on the website still look quite CGI like in my opinion.

speaker-0: I've seen videos on Twitter not posted by them that were AI generated but still very, very well made, of like a T eight hundred fighting and destroying a Tesla Optimus, for example. Which has not happened yet. It probably will in the next few weeks to months. But yeah, I've seen that directly. But it wasn't going from them, it was like, I don't know, AI slop from the Twitter space, you know.

speaker-1: Yeah, I'm not really talking about that. I'm I'm talking about the actual videos that they have on their website. Because if you go on their website, I'm gonna show one here maybe. ⁓ it's a really good video, don't get me wrong, but it just has this like ambiguity to it if you watch it. And you just think like something about the lighting that that kind of feels a bit CGI. But at the same time, the CEO of the company, he took a direct kick from the robot in one of his videos in protective gear. to prove it was real hardware and not CGI. I'm really not sure what to make of all of this, but their T eight hundred is definitely well kind of a terminator because that's what they call him after, right? So yeah, even on the website they have like born to disrupt and then they have like these fighting scenes with this robot and it really gets you going. Like you really want to see this robot fight and punch somebody in the face. So yeah. It's quite quite entertaining. Yeah.

speaker-0: I've also seen videos of it in the streets of New York and ⁓ people like playing some kind of fake boxing with it. And they were actually scared. You could tell the people were keeping a safe distance from the robot because the robot has like a lot of power, you know, ⁓ in its hips and in its arms. Even like I think there was like a video with a professional boxer doing like shadow boxing in front of it, and ⁓ the guy was not that confident in front of the robot for sure.

speaker-1: Yeah. I mean you ha know that of course a robot is made out of metal and made out of this very, very strong materials. So it's even very, very painful to do a low kick on somebody that has a really strong leg sometimes, you know. But if you low kick a robot out of metal, you probably will just break your ⁓ legs. But maybe at some point they will create a new leak where you have humanoids that are basically with artificial skin and muscles and they look like like normal humans and then they fight humans, that would be insane. But for now it seems they're mostly focusing on robot versus robot, which I think is the right approach.

speaker-0: Yeah, robot vs. human will take some serious insurance company to get involved and some serious medical attention.

speaker-1: Yeah. They also showed this T eight hundred at the CES in Las Vegas this year. They also have a lighter version which is called PM Zero One. Also super dynamic movements, pretty interesting mobility and performance. And yeah, they also got high-level praise, of course, from the Chinese premier Li Jiang during a Guangdong inspection, for example. So they're also not a small player necessarily. And They're trying to expand applications in, of course, also other areas beside the fighting. So they want to go to logistics, manufacturing services, the usual areas that you want to go into when you are building a humanoid, which it is after all. They also got some backing from investors like JD.com. Also, they are already selling it to the US. I think the T eight hundred has appeared ⁓ in the US because it was imported by the startup called REK or REC and they import it for entertainment and combat events. And that's I think where you see these viral clips sometimes where it's around the United States and people fight with it because that definitely shows interest in non-industrial use cases like VR controlled fights. Because I I'm not sure some of these fights you have two autonomous robots potentially fighting each other, but you can also have one where both are teleoperating the robot. Right, then they're fighting each other with a motion tracking suit or VR suit, basically like with a controller. And that must be so fun if you just fight another person. There was this show, I don't know if you remember it, I don't know if it still exists. It's called Robot Wars. If you know that, do you know that? Did they have it in France as well? Yeah, it was a show where they had this center stage and ⁓ you know there was like protection border around and

speaker-0: I don't tell

speaker-1: Two teams fighting each other and they would let their own constructed robots go inside the battlefield and fight each other. And they had all sorts of different mechanics in these robots. I remember there was one who had like a spinning plate. So every time another robot would come near it, the spinning plate would hit it away. There was one with with a like an axe that goes up and down like that. I mean, it was far from a humanoid, but it was super entertaining. I used to watch it sometimes.

speaker-0: Videos like this on YouTube, but I think they were from the US. But it was basically ⁓ teams of geeks, robotics geeks, setting up a robot. And what was interesting to me is that there were mostly wheeled robots, but all of them had different styles. One of them, like you said, had like some kind of projectile or something going around it to kick the other robots away. ⁓ I remember one robot looked like a shovel, and its goal was pretty much to flip the other robots over so it could ⁓ kid it, so it could So it should stop moving. Another one had some kind of very strong clamp or hammer to crush the other robots. So you had different philosophies basically of fighting. So that was interesting. One thing about what you said is that you said robots are teleoperated. ⁓ so far I've only seen teleoperations through like a gamepad. Have you seen for real teleoperations through ⁓ VR headset and exoskeleton, for example?

speaker-1: I've definitely seen demos and their demos were even from engine AI where they have someone in the back and he's like fighting and the robot is copying the the moves in real time. And it's also something that you've seen with the dancing videos that were very viral. So I guess it already exists, yeah.

speaker-0: I think it is visible, but I don't think I've seen it live. Because for example, you said it's REK, the company that makes robots fight in San Francisco. I think it would be really fun to have like you get a UFC type octagon with the two robots, and on the side you get two guys who are far away from each other in like a safe zone, each of them, with like an exoscoreton, a VR headset, and they fight each other through the the view of the VR headset. It needs extremely low latency to work, obviously. But then you can imagine to have like ⁓ cameras from the outside of the octagon, just like a normal fight, but you could also get POV camera from the robot's eyes, you know? And I think it could be extremely fun to have that.

speaker-1: Yeah, I think it's gonna be super viral because some people hate fights, of course, ⁓ but some people love it. I'm more in the category, I I really love fights. I think it's also athletics and if you have robots, well, it's even better. Nobody hurts themselves and ⁓ you push technology with this very entertaining little thing that you have going. So yeah. I kinda like it. And if you look at the T eight hundred it's actually pretty massive. So it's around one hundred and seventy three centimeters tall. Weighs seventy-five to eighty-five kilograms, so very much human like proportions. Their actuation is twenty-nine degrees of freedom, excluding the hands. It has a custom high torque planetary joint or multiple with up to four hundred and fifty newton peak torque and fourteen thousand watt instantaneous power. So that's pretty massive. I'm not sure if A human would survive that in some cases. They have active cooling on the leg joints for sustained high-intensity work. I've never heard this before, which is quite interesting. And they are up to four to five hours of continuous operation with them. They have a ternary lithium or an optional solid state battery. Their sensing and control is pretty similar to the other humanoids as well. Omnidirectional awareness, motion control. They have hands three-fingered for those, grippers with multiple joints and For practical manipulation. And this robot, the T800, comes in different editions as well. So they have a base one, an entry-level one, then they have the ecosystem, the open source-friendly one. So if you wanna have a very advanced fighting combination that ⁓ you normally need to jail break your robot for, then maybe you get this one. Then they have the Pro and the Max. And these are like the flagship models. And when it comes to price, they're starting around twenty-five thousand US dollars. But in China, so it always depends how much you then pay for the tariffs and the logistics. And the mass shipments began around mid twenty twenty six. So basically right now.

speaker-0: It's funny, so I just started shipping and we see it everywhere, including in the US already. I think it's ⁓ I have one comment and one question for you. So it's an incredibly cheap price tag for such a powerful robot because it has, I'm guessing, one of the most powerful actuators embedded for humanoids, considering the strength it has. And my question is you said the hands have three fingers, is that accurate?

speaker-1: Yeah.

speaker-0: Okay. So three fingers plus a thumb, or is the thumb included?

speaker-1: Yeah, they have two fingers and a thumb. And ⁓ they also have other versions of course. They have other robots, so they have a couple of different products and some of them have more advanced hands as well. But for the T eight hundred that's what they have.

speaker-0: I see. So I I'm presuming they are probably actuators based hands and not tandem based for reliability and how solid they are over time. I'm guessing they have less fingers to have more strength per finger, probably since their goal is mostly strength, right? Yeah.

speaker-1: And they're probably focusing on reliability, I guess, if they wanna use it for fighting.

speaker-0: And strength per finger, which is one of the main issues when you make robotic hands. And it's a it's very interesting because it's a very peculiar design we haven't seen so far. So far, hand design is basically just a basic clamp or a five-finger hand that we covered, you know, in a previous episode. So one good thing about having a three-finger hand is that it's a very different design than the one used by Tesla, right? Tesla has five fingers, human-like hands. tandem based. I'm specifying that because Tesla recently sued a hand robotic startup from San Francisco. And they also settled the lawsuits very quickly after. On June twenty ninth, so like a few days ago, there is a company in San Francisco called Proception. Have you heard of them?

speaker-1: No.

speaker-0: Okay, so it was funded by two ex-Tesla engineers, Jay Li and Jack Su. Both of them are ex-Tesla. And Li is actually ex-Tesla Optimus hand team. So he was exactly on the topic already. Li was working on that until September twenty twenty four and six days after quitting, he created And in June 2025, so like one year ago, Tesla sued him for I don't have the exact legal term, I'm not sure if it's infringement, I'm not sure if it's ⁓ theft of ⁓ intellectual ⁓ property or something. But what the lawsuit makes clear is that Tesla is arguing that Lee, right before quitting, downloaded a lot of files he had technically access to within Tesla. But that he had no actual work reason to look for. And all of them were about hand research and development, hand sensor on the Optimus, right? And all of a sudden, Lee quits Tesla. Six days later, he incorporates a company that makes tandem-based humanoids hands with 22 degrees of freedom and a bunch of specs that are exactly the specs shared by the Tesla Optimus hands. So you have to admit that this can look a little bit shady, right? Yeah, just just a little bit. Tesla sued them ⁓ like in June 2025, and we don't know at all what happened, so ⁓ we are doing a lot of speculations here. But one year later, in June this year, Tesla settled the lawsuit and the terms remain unknown. And Lee appeared publicly at the same time, right when Tesla settled, and he announced two things. Obviously, that Tesla

speaker-1: Yeah, a little bit.

speaker-0: Folded the lawsuit and that it was a good resilient test for them. It looks like Tesla's legal department is extremely aggressive when they are suing you, so you don't want to get sued by Tesla, I'm guessing. And he also announced an 11.5 million seed round. Why combinator is part of it? And it's also it's interesting because in the case file, Tesla specified that on its work computer, Lee search Google for terms like. those YC fund robotics company and stuff like that. So it it was as if he was looking for an exit, you know. So yeah. But again, I think ⁓ Tesla is probably folding the lawsuit also because they don't have that much to win over it. And also when it comes to image, it's usually a bad image when a very large company tries to, you know, ⁓ make a smaller one cave. No matter what can be argued about whether or not documents were stolen from Tesla, the two engineers are still very skilled in this field and Lee was you know working a robotics hand so he has also his own knowledge in the space is very strong. So even if you were going to start a company without documents, he would have been able to make some pretty reliable hands, you know, to begin with. To tell more about the product, it's called the Pro Hand 9.0. It has 22 degree of freedom, ⁓ five fingers, tendon driven, and it has some kind of tactile skin over it. For pressure sensing. It also has another product which I think is really interesting. It's called the Pro Glove. It's pretty much the same hand, just without all the mechanic called parts inside. So it's like the skin of the hand you use as a glove for pressure sensing. And you use the two of them and you can work with them and it acts as a data collection device. And I think it's extremely smart to have the same pressure sensing skin on both the hand and the glove such records your data to train it, right? So yeah. You have a comment on it? No.

speaker-1: It it it just reminds me a bit ⁓ of A G Bot and their A two data collection unit.

speaker-0: Yeah, of course, data collection is going to be like crucial. It already is. But I think it's interesting that they make a data collection device that matches exactly how their hands work and behave. One thing I read that I found interesting is that ⁓ Tesla maybe had little ground to sue them because the 22 degrees of freedom, tender-based, five fingers, those are specificities shared by many hands on the market already, right? ⁓ Some companies choose to have actuator-based hands. like ⁓ the TI Intel Rate from NginAI or like Figure AI, you know, it's also doing actuators-based hands. We covered hands in a previous post. And also a company like LinkerBort from China, who is valued at six billion and makes hands exclusively, they have like 10 different hands already on the market. And they both have tendon-based and they and actuators-based, and they have also one that is five fingers, 22 degree of freedom, and so on. Because it's a typical hand structure. that is the right balance between being similar to human and being easy to configure in a factory, in a factory setting and in a tr data training setting. It just makes sense that this architecture of hand is appearing. And if we look at the videos of you know these hands, their demo videos are extremely ⁓ interesting and impressive to me because their hands, since they have some kind of this glove made of sensitive skin over it. They can pick up some very small parts and make some very precise like ball screwing and task tasks like that. That other hands that have a very how to say hard or slick material at the tip of their finger cannot really do. And that's where the edge is. It's a very interesting company and I like the approach of a company that does one thing and does it extremely well, which is the tactile sensor that allow for very precise object manipulation. And I think they are really onto something here. And also that maybe Tesla realized there is no point in suing them. Tesla is valued at I don't know how many billions this company just raised 11 millions. If they end up making a good product, Tesla will just acquire them in like a heartbeat. That just makes sense for them. And ⁓ the vision of the founders is to become some kind of hand providers for all the robots manufacturers, just like Linkerbot again. And they also argue that in the car industry, if you look even at very large car manufacturers like ⁓ Mercedes or Volkswagen, most of their sensors they don't make. Right? So they buy the sensors from companies such as Valio or companies like this. And they say humanoid robots companies, they will have to, you know, outsource the hand as well. Because ⁓ Elon Musk also said it in a separate and previous interview that hands was like 80% of their research and development human work time. So if a human company can focus on scaling their production line and not on doing R and D on hand and just outsource the hand R and D, I'm guessing it's a win win for everybody else.

speaker-1: And also if they do the lawsuit, they always give attention to this other company, you know. Just like if you if you do a lawsuit and then you have to fold it then that's a bad play to be honest.

speaker-0: Yes, yes, yes. Yeah, there is that. And there is also I think the image ⁓ battle is also very strong here. Tesla is an extremely large company. it has been pub it has had public image trouble lately, Elon Musk as well, you know. So I think Tesla suing a small company that was incorporated like a year ago, ⁓ is not really a good look for them. So they know it's not worth it. And again, yeah if they have a good groundbreaking product, they will be acquired in no time, I think. Possibly. But yeah. In the end I would be very curious to see like a comparison of or competition between perception and Tesla's hands in like a parts manipulation comparison of challenge between them.

speaker-1: Robot in a robot fight. In a robot fight, please.

speaker-0: And a robot fight? I think a robot, you iron arm competition or something. Yeah, arm wrestling, exactly. I hope you guys enjoyed all the topics. Really enjoyed being here with you today. ⁓ please subscribe to our channel, subscribe to our Twitter. We post daily on robotics news. And see you in the next time. My name is Leo.

speaker-1: I'm wrestling. My name is Oscar, bye bye.