Episode 01 - Verobotics | NVIDIA GTC Interviews: Tech CEOs Around The World

Added: 2026-05-23

[00:00:00]Hello everyone. I'm here at Nvidia GTC [music] and I talked to tech CEOs around the world.

[00:00:06]>> As a company, we don't do back flips or dancing. We are here to do real work.

[00:00:11]>> Being able to do the parallel processes and the billions of calculations that are required.

[00:00:14]>> So it's like a exoskeleton upon your hand.

[00:00:18]>> No one need [music] to risk his life.

[00:00:19]It's very simple.

[00:00:20]>> But also how do they take advantage of AI with their existing data centers?

[00:00:23]>> That's why I started human and to free people from dangerous, hard and repetitive tasks. Tell me, who are you, where are you from and what are you doing? My name is Ido Kenan Sar. I'm the CEO and co-founder of the robotics.

[00:00:35]We're building a vertical robotics for the building environment. Basically robots that traverse the building from the outside and we're doing cleaning, inspection and predictive maintenance.

[00:00:45]Wow, I just saw that beforehand and I was impressed. What does a vertical robot mean? So high-rise building. Yes.

[00:00:52]>> The way their exteriors are being handled today, whether if it's for cleaning or for inspection, people go on the rope or on the crane on the top of the building, which takes them hours and it's very dangerous and it's very risky.

[00:01:04]So we've built a robotic platform. You can think about kind of an iRobot that traverse the building from the outside and is doing all the tasks the building need to have in a much safer, efficient way. So there's no human needed anymore?

[00:01:17]No. No one need to risk his life. It's very simple. It's very straightforward.

[00:01:21]How dangerous is it? Like do you have numbers how many people hurt or like Yeah. There is there is thousands of accident. Luckily not a lot of casualties, but a lot of accidents and the main reason is a lot of old infrastructure. Imagine a building with a crane on the roof that's been installed like decades ago. Yes. And our goal is really to automate and to make the process really seamless end-to-end.

[00:01:44]How did that come? Let's just start with this. That's a good question. [laughter] Have you been doing this before? Like >> In a way, I would say that I'm an I'm an insider on the industry. We started >> an insider in the industry? We've started the robotics from the problem.

[00:01:57]So my family is the biggest facade construction company in Israel. I led the innovation there. Um we looked to bring robotic solution to do this task because it's very it's a huge market.

[00:02:05]It's like $40 billion per year.

[00:02:07]>> Wow. Uh only around exterior services.

[00:02:10]And nothing changed there for the last 80 years. Yeah. Started this process of okay, there is so many good technologies out there. Someone probably built something. So, it started very organically. Yes. To speak with companies from all over the world to solve this problem. And we saw a lot of solutions, but no one was actually able to solve it because it's a very complex problem. And the solution that we've seen were like 100 kg, 200 kg, and they could barely fit to most buildings. So, my understanding was okay, if someone will build a lightweight solution that would be automated, >> Mhm. and it would be smart that it can actually collect data on the facade and give a lot of insight, there might be a huge win. Decided to leave the family company and to do it myself. You have really have a purpose there. I see that.

[00:02:49]>> Yes. And how's that working? So, the robot is not Spider-Man, right?

[00:02:52]>> It's in a way it's a Spider-Man.

[00:02:53]>> Spider-Man. It's kind of a Spider-Man.

[00:02:55]Tell me like how's it So, there's no rope?

[00:02:57]>> There is rope for safety.

[00:02:59]>> the safety of the robot.

[00:03:00]>> Yeah, exactly. And it's the cable is also for energy and it's also ethernet cable. So, you can imagine kind of a robot to traverse on the building with two arms. It's not a humanoid, right?

[00:03:08]Okay. It doesn't need to be, right? It doesn't need to be, exactly. We made this to solve a specific problem, right?

[00:03:15]So, the robot traverse on the facade like literally kind of a Spider-Man. And then he has a payload, and the payload is there for cleaning and also collecting the data. Everyone coming into our business, they look at the video and they look at the product and they're like, "Okay, that makes sense." Right? And all of the work of the robot is completely automated. Then the operator is there, he's pressing a button, and he doesn't need to do anything, right? He can Yeah.

[00:03:34]scroll his phone, he can smoke a cigarette, he can drink coffee. No one need to risk his life. Yeah.

[00:03:39]>> And making the entire process much more efficient and smarter because we're collecting a lot of data on the facade.

[00:03:45]For building owners, they have no idea what's going on with their properties only when it's too late.

[00:03:49]>> So, is there still a human being down there steering the robot. Can't you also like map the the building and say like okay this facade has like like 100 m high and >> So this is what actually happens, but eventually there's always the last mile and in the last mile you need to have a person in the loop. Yes.

[00:04:06]>> And it's something that I think that you can hide, right? I mean it will take time until everything will be completely 100% automated, but I think that for us specifically working on a very dangerous and complex environment, we still have a person in the loop. We are reducing the labor in like 70-80%. But we call them kind of a babysitter. They're monitoring if there's any issues or something like that. They know how to take the lead.

[00:04:28]Well, that's good. And you're not only applying that in Israel. How is the ecosystem like Inception helping to scale global? So we got a lot of help also with introducing of Nvidia. We're working in Israel and in the US and we had some project in Japan and in Hong Kong. I can share with you that in Israel we are serving Nvidia HQ. Yes.

[00:04:46]>> So it's kind of a funny to think that you see the robot on Nvidia HQ campus in the north of Israel Wow. where they're actually building all of the technologies. So we have a contract with them for the services that we're doing and also opening a lot of doors. We see it now especially in the event with partnership on the technology side Yeah.

[00:05:02]and the people from Nvidia team that can help you see what's the latest technology because we're building something quite complex, right? We're building the simulation environment ourselves because it's not like a robot that hovers the floor that you can have so much Yeah. data sets going on. We're building a very unique data sets that no one has before on the vertical world on facade with so many limitation and challenges. So luckily we're getting a lot of help from the team. So tell me more about the simulation and how you like work with Nvidia tech. Yeah. We're using obviously the AI stack. So we're using edge GPUs for the real-time operation. When you think about the real-time operation, I mean imagine a robot that works at like height of like 50-60 floors. Yeah. Right? So with so much is going on at the same time and this is probably the only solution today in the market that can actually support us. Another way that we're using is Edge on on cloud for the training itself. And here in the event we've met a lot of people from Nvidia that we didn't know before, that we presented what we do and they all of them was like, "Oh, you should always you should also do the this tool and this tool and this tool."

[00:06:03]So it's like it's good to have this open communication that we can actually be exposed to the latest that can support our activity. Yeah, you accelerate your knowledge network here like like crazy.

[00:06:13]Yeah. And so as we train robots in factories, you know, we we use real-time data for simulations, but also artificial data. How can you also use existing data that Nvidia provides you to improve your product?

[00:06:28]>> Yeah. So definitely the use of the data is quite unique, especially because and I think that what is very interesting, it's working both sides, right? Because we're generating data that is quite unique. Yes. Right? And it's kind of a super cool data cuz every building is that we are going on, we generate around 50 60,000 images Yeah. of the built world. Wow. Right? So eventually we're using it not just for like improving our models for the locomotion and the autonomous walking, but we're looking to train this data to identify different issues on the facade. And this is huge because no one is doing it today. Are you planning to adapt what you are right now doing to other fields? Are there other robots planned? Yeah. So currently we're focusing on our world, which is big enough. I mean, if you're looking on exteriors, we are I think the biggest thing around what we're doing on the business opportunity Yeah. is that we are constantly adding more and more building, right? And the built world is crazy, but we're not adding people willing to do this job. So for us it's like it's an endless industry and our goal is like to be super focused on that. We want to stay there. This is what we believe. This is what we are good at and this is kind of the industry that why we developed this solution. I love that. So maybe you can finish my sentence. The problem we solve in the world is walking on heights, dangerous jobs.

[00:07:42]Perfect. like the easiest to to explain and the visibility to data that no one has before. Wow. Thank you so much.

[00:07:49]Thank you very much for your time. Thank you.