What we talked about
Hiten Sonpal is the CEO of RISE Robotics, an MIT-founded company targeting the $60 billion hydraulic actuator market with the Beltdraulic, a fluid-free actuator that is three times faster and up to 90 percent more energy efficient than traditional hydraulics.
Show notes
Hiten Sonpal spent 16 years at iRobot growing from engineer to director before arriving at RISE Robotics, where his team is replacing a century-old technology with something inspired by how human muscles actually work. The company’s BeltDraulic actuator eliminates hydraulic oil entirely, drawing on belts originally developed for high-rise elevators to pull pistons rather than push fluid, and the physics advantage turns out to be enormous.
What we covered
- RISE Robotics invented the BeltDraulic by replacing hydraulic oil with polyurethane-and-steel-rope belts originally designed for elevators, where they were so reliable they lasted the lifetime of a building. The founding team, mostly MIT graduates, started by trying to build an Iron Man-style exosuit and turned to this belt technology after realizing muscles work by pulling in tension, not pushing with pressure.
- The actuator is three times faster and 75% efficient from electricity-in to linear-motion-out, compared to roughly 30% for typical hydraulic systems. Because the only place force is applied is the piston and the belt anchor points, there are no pressure spikes across hoses and seals that cause leaks and heat buildup in conventional systems.
- Autonomy compatibility is built in from the start. BeltDraulic actuators are digitally controlled and provide position, force, orientation, and health feedback to a vehicle control unit, making it straightforward to build a deterministic digital twin, the same drive-by-wire foundation that makes systems like Waymo possible.
- One unexpected use case: in an unloading cycle for a container ship, as long as the cargo weighs more than the empty fork, the system can recover enough energy lowering loads to lift the empty end effector back up, allowing the machine to run continuously without charging. Hiten described this as a “perpetual unloading cycle.”
- RISE targets industries with 24/7 pain points rather than slow-adopters. Oil and gas runs continuously, so efficiency and reliability gains multiply faster than in construction, making customers more willing to pay a premium. Marine and maritime are a second focus because hydraulic oil contamination at sea is a significant liability, fishing boats cannot afford a leak destroying a day’s catch.
- On leadership, Hiten described a moment he stepped out for a phone call and returned to find his team had independently reorganized the company, canceled first-generation projects, and restructured around a cylinders team serving internal and external customers. He quoted a leadership principle that stuck with him: a great leader isn’t moving people like pieces on their own chessboard, the team is moving the leader like a piece on theirs.
About Hiten
Hiten Sonpal is CEO of RISE Robotics, an MIT-founded company working to replace hydraulic actuators with its fluid-free BeltDraulic technology. He spent 16 years at iRobot growing from engineer to director, overseeing hundreds of people and millions of units shipped, before founding, leading, and advising robotic companies across autonomous, outdoor, and industrial sectors.
- LinkedIn: https://www.linkedin.com/in/hiten-sonpal
- Website: https://www.riserobotics.com
Episode 143 of the PreVetted Podcast.
Full transcript
Federico Ramallo (00:00.64) Welcome back to the pre-vetted podcast where we spotlight extraordinary people and remarkable talent reshaping our world. Today, I am joined by Hiten Sonpal, his CEO of Rice Robotics. Hiten started his career writing code to Self Fund College, spending 16 years at iRobot, growing from engineering to director, leading hundreds of people across millions of units shipped.
They went on to fund, lead, and advise robotic companies across autonomous outdoor and industrial sectors. Today, he’s at the helm of a company working to replace hydraulic actuators, a technology that has powered heavy machinery for over a century with something three times faster and up to 90 % more efficient. Hitam, welcome to the show.
Hiten Sonpal (00:55.041) Thank you Federico, this is incredible. I’m so glad to be on this podcast.
Federico Ramallo (00:59.7) I am honored to have you here today.
So for people meeting you for the first time, can you tell us more about what you’re doing?
Hiten Sonpal (01:09.536) Yeah, so I’m CEO of RISE Robotics and what RISE has done is invented this new technology that we call BeltDraulic. And what BeltDraulic is, is a fluid-free version of hydraulics. We’ve taken out the oil from hydraulic systems and instead replaced it with belts. So instead of using oil to push pistons back and forth, we’re using belts to pull pistons back and forth. And by doing that, we’ve created a linear actuator that is three times as fast as hydraulic systems, that is three times as efficient as hydraulic systems,
that is three times as durable as hydraulic systems and supports autonomy and electrification out of the box.
Federico Ramallo (01:50.397) Interesting. I haven’t thought about autonomy with this actuator. What makes them more compatible with autonomous compared to hydraulic actuators?
Hiten Sonpal (02:02.633) When you think about a hydraulic system, typically has bang, bang control. The operator is opening a valve, closing a valve, and the output is open loop. And usually the operator is closing the loop with their own eyes. So imagine a construction site where you’re running an excavator. The excavator is lifting some weight on the bucket. The operator is looking at the bucket, seeing how much, how the machine is behaving, feeling the whine of the pump, and then using that to decide how to drive the hydraulic system. And ultimately, they can press
the hydraulic system full and if there’s too much resistance the hydraulic actuator won’t necessarily move so they have to feel it all out and that kind of system is very hard to automate because you don’t have closed loop system closed loop control at the subsystem level you only have it at the full system level and it’s done visually what our actuators do is because they are digitally controlled they’re multi-position from the get-go and the high level system the vehicle control unit can
control the actuator very precisely. And then we provide a lot of feedback back to the vehicle control unit. So we can provide position feedback, we can provide orientation feedback, we can provide health feedback, and we can provide, did I say force feedback already? Let me start again.
Federico Ramallo (03:22.528) Yes.
Hiten Sonpal (03:24.677) We can provide force feedback, we can provide position feedback, we can provide orientation feedback, and we can provide health feedback. So with all of this information, it’s very easy for the vehicle control unit to build a digital twin of the full system. And that digital twin first enables teleoperation at the basic level. It also enables semi-autonomy. And then that can be increased all the way to full autonomy as the software foundation models get better.
Federico Ramallo (03:55.099) Interesting. So the precision comes that you don’t have any slippage, right? There is no slip on the mechanism, right?
Hiten Sonpal (04:06.598) Yes, the precision comes exactly. There’s no slippage. And also there is closed loop control happening inside the actuator. So the actuators are drive by wire. You just tell them, I want you to go to this much position, use this much power when you get there. And the actuator will just move versus a hydraulic system where you’re putting in liquid at a certain pressure. And then the output is not really controlled, right? A human operator looks at the output and decides if it’s done well.
Sometimes people will add like a switch at the end so the hydraulic system doesn’t bang into the end aggressively. Other times folks will try to add sensing systems, closed-loop control, on the outside of the actuator. But you can imagine because there’s liquid in the path, those control systems can be very complicated and can’t be very precise. And they’re expensive. But for us, it’s super simple because you have a motor and typically if you’re driving a motor you’ll have an encoder. We use brushless motors so they have hall effect sensors which are effective.
encoders so position feedback is free which is really exciting and then in terms of measuring current in terms of measuring force you just measure the current on the motor and that tells you how much torque is being applied and then you can turn that into force so even force is free
Federico Ramallo (05:11.882) That’s amazing.
Federico Ramallo (05:29.226) Right, you have force, you have position, and you can determine how much, you can determine when is the end of the career for the arm or whatever mechanism we have. And also you can detect if there’s any unexpected resistance, you can measure it real time.
Hiten Sonpal (05:40.452) Yes, that’s right.
Hiten Sonpal (05:54.17) That’s right. That’s right. And we also have, because you have electronics inside the sensor, there are other things you can do, like health monitoring, which we provide so people can know that this actuator needs to be serviced even before it needs to be serviced. You can’t do that with a hydraulic system. A hose will burst and then you’ll have to go service it then. But we can tell you when we get to a point where some service is needed. And then orientation is very easy to put in there for free.
Federico Ramallo (06:20.616) Right, right, interesting. So why belts and not gears?
Hiten Sonpal (06:32.236) Yeah, so.
Federico Ramallo (06:32.34) I I’ve seen bolts, mechanisms that access linear actuators as well, right? But on every gear you have a difference, right? Every time you push back and forth, you have a difference that could affect the precision, right?
Hiten Sonpal (06:51.714) Yeah, so.
Yes, definitely there are multiple kinds of linear actuators. Besides hydraulic systems, the most common ones are screw type linear actuators, right? And the issue with screw type linear actuators is that there’s metal on metal contact. So anytime you have shock, you can end up deforming the metal and your actuator is basically done because it will get stuck in that spot and you can’t move it anymore. So those actuators are not really good for shock applications. And also these screw type actuators can’t go very fast because if you move them very fast, they start to oscillate.
There’s a lot of mass in there. And then lastly, you can’t get a lot of force out of it because to get force, you have to make them strong. And if you make them strong, then you have to add more metal, which increases the weight and increases the cost of these systems. The way our company was founded is that the founding team, the majority of whom went to MIT, wanted to build an exosuit. So think Iron Man without the jet pack. They wanted to build something like that.
Federico Ramallo (07:48.96) Yeah.
Hiten Sonpal (07:50.785) That was their dream. And as they started building that system, they realized that they couldn’t use hydraulic systems because they were too heavy, couldn’t move fast enough, that required too much space. And so they started to look at other ways to solve this problem. And they looked to nature for inspiration. And when you think about muscles and humans and mammals and really all kinds of animals, they’re not really using fluids and compression to move anything. Because you can imagine if you tried to do that, you’d end up bursting through the epidermis and you’d have blood all over the place, right?
What humans do, and this is very efficient, is they use fibers in tension to pull things, right? Even when a human is pushing something, there is a muscle inside the body that’s contracting instead of, yep, exactly, it’s pulling instead of pushing. And so they were like, okay, we need to look for fibers that can accomplish the same thing. And they started looking and they found out that about a decade before they were looking at this space, there was a breakthrough
Federico Ramallo (08:33.313) That’s pulling. Right.
Hiten Sonpal (08:50.516) through in the elevator industry, where they started replacing steel cable with these new belts that were polyurethane on the outside and had steel rope on the inside. And these were very, very strong. And they replaced the steel cable and they eliminated any kind of environmental damage to the steel and they also eliminated any steel on steel friction. And so they ended up with these belts ended up
being so reliable that they would last the life of the building. So once you install them in elevators, you would not need to do any servicing. And because they were flatter than steel cable, you could take the elevator room, which used to be on the top of these skyscrapers, and really turn it into an elevator box. So they’re very compact, and they were used for safety-derived applications, like moving people in lifts. So they found these belts, and they were like, okay, let’s see if we can innovate around these belts. And these belts have constraints on how
you can package them. But the founding team found a way to package them in a very compact form factor and use these belts to pull pistons back and forth and that’s how they came up with this technology that we now call Belt-Draulic.
And once they had invented this technology, they realized that they would have to invent the exosuit market next, right, because no one was buying exosuits at the time. And they said, you know what, there are all these systems that already use hydraulic technology. And if we could just use our technology there, we can improve these systems with something that’s clearly much better. And I think it was a smart decision on their part, because hydraulic systems are a $750 billion market split between mobile and industrial hydraulics. So it’s a huge market.
opportunity and disrupting that market is actually easier than trying to invent a new small market that will take decades to grow.
Federico Ramallo (10:44.437) Wow, that’s amazing. I mean, I didn’t realize that it had still inside the belt, but that makes sense, right? Because you have a lot of tension, tension strength to support the forces, right?
Hiten Sonpal (11:01.107) Yes.
Federico Ramallo (11:03.969) So why has it taken this long for the industry to find other solutions for hydraulics?
Hiten Sonpal (11:18.761) Well, I think that it has to do with local minimas. Like if you think about battery powered cars, for example, Battery powered cars and internal combustion engine cars existed very early on. So folks are trying to make battery powered cars.
very, very early. But what happened is that the material science was not really advanced to that point. And so it was easier to build an internal combustion engine, and it was easy to then ride the curve of improvement in internal combustion engines. And now I think we’ve had a ceiling where you can put increasing amounts of effort for marginal gains in the efficiency of internal combustion engines. But in the meantime, the laptop market came about, and there was this demand for small batteries. So there was innovation.
Lead acid batteries obviously wouldn’t work, so folks came up with nickel metal hydride. Well, it was nickel cadmium at first, then nickel metal hydride, then lithium ion, then it was lithium polymer. Now it’s lithium ion phosphate and lithium titanium oxide. And so there’s all these battery chemistries that people have started putting money into. And so now electric cars.
are now feasible because the battery curve is advancing faster than the internal combustion engine curve. That has come to its maxima. Really, you cannot do much innovation there. And actually, the same thing happened with hydraulics. It used to be the case that people used to use cables and block and tackle pulleys for all of these large systems. And then hydraulic systems came about, and it created a huge shift in the industry.
giants like Caterpillar were born because they started building systems like hydraulics. And companies were able to ride the hydraulic curve for a long, time. But I think it’s come to the point where they’ve also hit a maxima. And these new belts that have just come out are an innovation that’s only been around now for 15 to 20 years. And people really haven’t thought about how to use these belts in a completely different way to achieve the same results of hydraulics. And as I mentioned, the materials that these belts are
Hiten Sonpal (13:24.415) made out of today are very, very non-exotic. So when you think about steel and you think about polyurethane, these are everyday contemporary materials. But when you think about the new materials that are coming down the pike, you’d think about things like Kevlar and…
carbon nanofiber and nanotubes. And so there is this innovation that’s happening in the material space where we can continue to ride the advancement in belts, which hydraulic systems can’t because they’ve reached a, as I said, a local maximum.
Federico Ramallo (13:58.678) Right, right. And also there was no need on that market to make the technological change, right? Because hydraulic was good enough, right? But I think you mentioned being able to build autonomous systems where now you have a new need that hydraulic won’t be able to suffice, right? So being able to build autonomous machines,
with belts, that makes it much more interesting and also has a benefit for the rest of the industry, right?
Hiten Sonpal (14:35.492) That is correct. most people sort of know this intuitively, but it takes some spelling out. When you see what’s happening with autonomous cars or autonomous vehicles, one of the things that needs to happen before they go autonomous is that they have to become drive-by-wire systems.
Certainly you can put money into a car that’s not drive-by-wire and you can add all the sensors and the complexity, but all of that just adds cost and then the company that’s making the autonomous capabilities has to put all of that investment in, but they don’t really know how to do all of that usually. It’s better if the cars that they buy are drive-by-wire. So if you look at the Waymos that are deployed, they’re all drive-by-wire systems and it’s very easy to build an autonomous system on top of that. And that’s what we do is we take a hydraulic system that’s, you know,
driven by liquids and hoses, which are very hard to control, hard to package and so on. And we changed that to a system that’s just driven by power and CAN bus. And once you have that drive by wire control, it’s very easy to build an autonomous system on top of that.
Federico Ramallo (15:43.105) Right. So you talk about that is three times faster and 90 % more energy efficient. Right. Can you tell us a little more where that gain actually comes from?
Hiten Sonpal (15:58.806) Yeah, so the speed and the efficiency games really come from the same underlying physics.
When you think about hydraulic systems, you’re trying to move fluid, which is typically viscous fluid, because if it’s not viscous, then it starts to leak. So you want something that’s thick with large molecules. And you’re trying to move this viscous fluid typically through small hoses and through small holes. And all of those areas experience that efficiency loss as you’re pushing the liquid. And the faster you go, the more eddy currents you have in the fluid, like with cars.
and driving faster, the resistance increases with the square of the velocity. And you have similar effects when you’re trying to move hydraulic fluid. So there is a practical maximum that you can get to. And also when you try to move things faster, you have to apply pressure. And also there’s a practical limit to how much pressure you can apply because it gets distributed across every seal and every interface. So those are.
the underlying physics that limit hydraulic systems from being fast. And then the intermolecular friction and the eddy currents create a lot of heat, which creates the inefficiency.
But then there’s also the drive train itself, which is the pumps and so on, which are also inherently inefficient because you’re trying to take electrical power and you’re trying to move fluids from it. So that transfer is not very efficient. So when you add up all of the losses, most hydraulic systems are typically 30 efficient. And our system,
Hiten Sonpal (17:31.381) is 75 % efficient. The actuator itself, from mechanical input in, rotary mechanical input in, to linear mechanical input out, is 90 % efficient. But when we add motors, which are typically 85 % efficient, the total efficiency of the actuator from the electricity in to the linear mechanical motion out ends up being 75%.
Federico Ramallo (17:57.451) Right, right, because you have less layers of conversions and you’re just transferring electricity from the battery or whatever source, energy source you have to the actuator, right?
Hiten Sonpal (18:11.644) Yes, and the key insight is that
We’re being very intentional about where the energy is being applied, right? So when you’re pulling, you’re being very strategic about where you’re pulling. When you’re pushing liquid, especially, you’re pushing everywhere, right? So that’s all force that’s experienced all over the machine, which ultimately will result in a maintenance problem. Because even places where you don’t want pressure, you’re ending up with pressure. typically only want pressure on the piston. You only want the piston to move back and forth, but you end up with pressure on
every hose interface. You end up with pressure on every seal and so all of those ultimately end up leaking. Whereas with a tension-based design, the only place that experiences any force is the piston and the anchor points of the belt on the other side.
Federico Ramallo (19:02.07) Right, right. And you also, with that, you also win reliability, right?
Hiten Sonpal (19:10.887) Yes, because a lot of the reliability issues come from the pressure in places where you don’t want them and also the heat that you end up generating. And in our case, we’re so efficient. just there’s not a lot of wasted heat. So there’s not a lot of energy going in places where you don’t need it. And so the wear and tear on the system is very low.
Federico Ramallo (19:35.362) So you don’t need.
Federico Ramallo (19:41.366) you don’t need to dissipate heat on the actuators because of how efficient they are
Hiten Sonpal (19:48.594) Yeah, sometimes for actuators that are very large and need a lot of power and run continuously, we would need to cool the motors. Because even at 85 % efficiency, if you’re dissipating 15 % in a motor, it’s ultimately just a question of how much power are you dissipating in that motor.
But that dissipation is very small compared to a huge radiator that you might have to build for a hydraulic system, where you’re trying to dissipate 75 % of the energy.
Federico Ramallo (20:24.704) Right, right, because the friction on the oil goes to heat the oil and then you have to cool it down. Otherwise, it breaks down.
Hiten Sonpal (20:33.189) Yes. Yes. Yes.
Federico Ramallo (20:38.4) Yeah, and with this system, you actually have heat on the motor, but not on the transmission, not on the belts.
Hiten Sonpal (20:49.317) That is correct. That part of the actuator, the belt part of the actuator is 90 % efficient. And the way it gets to the 90 % efficiency is we have ball bearings on the inside that are over 99 % efficient. And because ball bearings are really well understood and we know how to, we use automotive grade ball bearings. So.
The industry knows how to make them very reliable and very efficient and last for a very long time. And we’re just riding that particular curve. And so at any given bearing, you’re dissipating less than half a percent of energy.
Federico Ramallo (21:23.264) Right, right. That’s very interesting. And the other thing you, you know, when I was researching, I was reading that you can also recoup energy when you are winding down. That’s very interesting.
Hiten Sonpal (21:34.874) Yes. Yes.
Yes, this is true for most efficient mechanical systems. So if your mechanical systems have an efficiency more than 50%, you can typically back drive those actuators and recoup energy. And so for our system, which is driven by motors, it’s very easy to get the energy back because any time you back drive a motor,
you end up turning the motor into a generator and the electricity flows backwards into the battery. This is why electric cars, when you hit the brake system, essentially what they’re doing is they’re taking the motor and turning it into a brake and a generator and then charging the batteries back up. We do the same thing. One really interesting use case for us, because we’re 75 % efficient, so our round trip efficiency when we lift something and we bring it back down is 50%, which means that
Federico Ramallo (22:31.051) Wow.
Hiten Sonpal (22:32.313) Yeah, which means that if we’re using an application where we are only lowering heavy weights, imagine we’re unloading a container ship with a forklift that’s using our belt draw like technology. As long as the container weight is more than the end effector weight, the fork that we’re lifting up and down.
we can run 24-7 and not have to charge. Because as we lower, we get 50 % of the energy that we put in back into the battery. And that’s enough to lift up the empty end effector back up to the top. So we can run a perpetual unloading cycle. And the machine can run forever without charging. Now, only in the unloading case. In the loading case, also, we’re much more efficient.
But in the unloading case, we have this unique use case where we can run forever.
Federico Ramallo (23:37.494) Right, you’re leveraging the potential energy by loading down and your system is so efficient that you can actually leverage that because on a traditional system you will be needing to put energy to do that.
Hiten Sonpal (23:41.677) Yes. Yes.
Hiten Sonpal (23:48.397) That’s.
Hiten Sonpal (23:52.973) That’s right. Yeah, the energy that our systems would need were loaded when the ship was loaded. So the energy was put onto the ship. And when the ship showed up to the destination, at the destination, they extract all the energy to run the system.
Federico Ramallo (24:09.165) That’s very interesting, that’s amazing.
Hiten Sonpal (24:11.991) Yeah, it’s fun to think about, yeah.
Federico Ramallo (24:13.795) Right, right. the it it’s a I mean, it’s a revolution on the industry, right? Because this is something that no other product in the market could do.
Hiten Sonpal (24:29.407) That is correct.
Federico Ramallo (24:32.695) So you’re targeting industries that are notoriously slow to adopt new technology, right? So how do you get a new actuator into machines that have run the same way for decades?
Hiten Sonpal (24:47.253) So what I’ve done is instead of trying to build something and push it onto industry that, as you said, moves slowly and makes changes very slowly, what we’ve done is we’ve said, OK, we’re not going to make this an engineering problem. We’re going to make it a search problem, or we’re going to search for the right customer that has a pain point and they want to make a change now. And using that, we’ll build the proof points. And so we decided to target the oil and gas sector.
Oil and gas production isn’t going away anywhere. It’s one of the dirtiest sectors in terms of energy use, very inefficient.
and the systems run 24-7. So any efficiency gains, any reliability gains, any productivity gains are all multiplied, which means that we can charge more for our product, which also means that our customers are experiencing a pain point more so than say construction where the systems may be running eight hours a day. So, know, oil and gas production or any kind of production, food production, which would be another sector that we’re going to be looking at.
All of those pain points manifest themselves three times faster. And so there’s a lot of interest in a big pain point around fixing those systems. So that’s one sector we’re looking at. Another sector we’re looking at is the marine and maritime sector, where there’s a lot of interest in taking ports and making them net zero. There’s a lot of interest in getting hydraulic oil off of boats because people, whether it’s a luxury yacht or even a fishing boat, they don’t want to leave a
tail of oil behind. Especially in the fishing industry, they can’t afford to have a leak of hydraulic oil go into the tank and destroy that whole day’s catch. So there are these specific pain points in the maritime and marine space where we think that our customers are cost insensitive and they’re interested and really excited to adopt this new technology.
Federico Ramallo (26:48.993) Right, right. That’s very interesting. I haven’t thought about the marine industry, but yes, I mean, while you were talking about food industry, I was thinking about oil contamination to the food. But on the maritime, you have the same issue. And yours, go ahead.
Hiten Sonpal (27:07.867) Same thing.
And it’s not just food. It’s also upstream, which is not just food production. Upstream, we call it agriculture, when you’re harvesting the crops. There also you don’t want hydraulic oil into food crops. And Europe in particular is pushing hard on getting hydraulic systems out of the supply chain for food, because they don’t want any contamination.
Federico Ramallo (27:33.655) Right. And we also talk a little bit about that there’s a benefit for autonomous vehicles or applications. What use case do you see where the autonomous use case would be most useful?
Hiten Sonpal (27:59.217) I think there are any number of cases. There are a lot of companies that are looking at construction equipment and agricultural equipment and trying to make them autonomous.
The big challenge for a lot of these companies are when they start with the hydraulic system, they have to add sensors to get feedback. And the sensors they have to add have to be robust. But they have to be mounted on the outside of the systems. And they’re really, the outside of the systems are not designed for these delicate sensors and so on. And without the sensors, it’s very difficult to build a system that’s safe. For example, think of an excavator.
An operator that’s using an excavator knows that they put too much weight into the bucket. And if they move the bucket too far to the side, they’ll tip the machine. For an AI system to figure that out, you have to train it a lot, and then you have to test it, and then you have to build a safety case and convince yourself that it’s going to be right nearly 100 % of the time. Typically, when you’re trying to build a safe system like that, you’re looking for
one failure in a billion hours of operation. So it’s very hard to build a safety case like that. And so what we provide is a deterministic model where we can give you a digital twin and you can just do the math to say, if I move this bucket 30 degrees, because we know the forces, so we know the mass, if I move this bucket 30 degrees to the left, I’m going to tip this machine. So I’m not going to let the autonomous system do that. We’re going to put guardrails and let the autonomous system know you can’t go beyond this point. So we can build a safety case.
from the bottom and then the AI system on the top can do things that are maybe not entirely predictable 100 % of the time but they will never do something that’s unsafe because we’re stopping it from happening.
Federico Ramallo (29:49.57) Right, right. I think that what you’re talking about is this concept of intuition, of feedback intuition, where we human are very good at that. We adapt to a car, it becomes extension of our body, right? We learn how to drive it, right? And it becomes, we think less about the moving the gears and the machine.
Hiten Sonpal (30:03.918) Yes.
Federico Ramallo (30:18.851) and we just focus on, want to go there, right? But here the AI agents or autonomous agents don’t have that skill, right? And developing will take a lot of time and effort, right? And it’s interesting that you’re talking about a deterministic input that you’re providing it to a non-deterministic actor, right?
Hiten Sonpal (30:42.55) Yes. And to your point about intuition, like we also balance that intuition with an understanding of risk, right? So.
If a human operator is operating system, they might be willing to push the corners a little bit, but then if they see that their construction workers on the site nearby, they’ll dial back the safety envelope themselves and they’ll say, okay, maybe I’ll go a little slower. Maybe, right? It’s like they just start to create the space that they need to make better decisions themselves. Whereas with an AI system, like you have to encode all of that, right? And then how do you even justify it? Because if you start with the assumption that the AI system is working correctly, then why would you need to dial it back?
anywhere, whereas a human is like, I might make a mistake and so I’m going to back up, right? Sometimes when you’re driving a car and it’s raining, you’re driving slower. If you come to a hill and you can’t see on the other side, you slow down. You make all of these decisions adjusting for risk that sometimes I’m not sure that an AI system would make them unless it’s already been pre-programmed and then you have to simulate the risk model, which people aren’t doing right now. They’re just looking at, can I just get basic operation?
Federico Ramallo (31:51.001) Right. Right. I am. This reminds me of a side story, reminds me of my dad used to tell me how his mom teach him how to cook. And my grandmother would say, you know, well, you put some salt salt into the food. And my dad would say, how much, how much salt? And then she would grab some and put it. And it’s like, well, you know, enough. Right. Yeah. But what does enough mean? Right. Tell me how much.
Well, you you put a little bit until it’s enough, right? It would drive my dad crazy, but it was her cooking by intuition and him asking for explicit quantities because he didn’t have the experience of cooking, right?
Hiten Sonpal (32:38.507) That’s right. That’s right.
Federico Ramallo (32:40.836) And for this AI system providing, being able to provide explicit instructions to the mechanisms and not have to worry about that level of intuition, that provides a stronger foundation to automate the rest, right?
Hiten Sonpal (32:58.432) That’s right, because all of that intuition of how you want to describe it, those additional layers.
ultimately result in an AI system that’s more complicated and has more hidden layers. So if you, for example, wanted to start with a system where the AI was only using vision to infer where the actuators were, you’re probably spending a large percentage of the neural network just thinking about that. So you just have a lot more hidden layers, which means that proving that the model is safe is even harder. So the more you can simplify the AI model, the…
the number of layers you can reduce, the easier it gets to start to do proofs because then you won’t get emergent behaviors that you don’t understand.
Federico Ramallo (33:41.007) Right, right. You were talking about end of career switches before, right? And that remind me to the old printers, when you turn them off or you turn it off in the middle of a printing, right? When you turn it on again, you will start doing all these movements to figure out again the end of life, end of career switches, right? You will go one way and they will go the other way, right? Until you will be able to say, well, this is my position, right?
Hiten Sonpal (33:48.371) Yes.
Hiten Sonpal (34:05.001) Yes. Yes, yes.
Federico Ramallo (34:10.948) Yeah, we used to play with that when I was a kid. And eventually that evolved to a binary positioning system that is still not, you know, as precise as having the motor giving you with the whole sensors giving you the position.
Hiten Sonpal (34:11.647) That’s right.
Hiten Sonpal (34:37.457) That is correct.
Federico Ramallo (34:41.028) So I wanted to ask you about a little bit of your journey. You have been an engineer, a director, a founder, a president, and a CEO. What does each role teach you about being successful?
Hiten Sonpal (35:00.345) I think in a lot of ways all of the growth has been very similar. There are two things that have that happen simultaneously. When you’re an engineer, you’re responsible for the work that you’re doing and you have total control over the work that you’re doing.
And as you work your way up the organization and so on, your responsibility increases and you have to be ready to accept the responsibility, but also be ready to give up the control because you’re working with people. And so that whole process of growth is around trust, giving other people the ability to be successful, and then also being willing to accept that you’re responsible when they make mistakes.
And then it is your job to figure out how to be a better leader and fix the situation. So that can include better processes. That can include better talent. That can include better training. That can include better strategy. That can include a culture where people feel free to ask questions. So as a leader, you’re constantly trying to figure out how to increase the performance of your team so you can trust them and give up more and more control.
while at the same time being responsible for all of it. So that’s been the whole journey. It’s difficult at every step to let go, to let go, to let go, to let go, and still know that when things don’t work out, it’s ultimately your fault.
Federico Ramallo (36:37.486) That is so true. And it’s a struggle. I think that it’s a struggle with our own ego first and then with the people.
Hiten Sonpal (36:51.024) Yes.
Federico Ramallo (36:51.116) especially coming from a technical background, right? Where we’re used to, know, things are more predictable than people, right? You can build an equation and it’s more predictable than having a team of people, right? And, you know, that predictability, least for me, been, lack of predictability, at least for me, has been a challenge that I’ve been, you know, going through on my initial, you know, when I start taking more leadership.
Hiten Sonpal (37:22.926) Yeah, the other side of the unpredictability, which is definitely a downside, well, let me rephrase. There are two sides to the unpredictability coin. One is the disappointment, and then the other is the delight. So I’ll give you an example of something that happened recently where…
delighted by something that was unexpected. And before I get into that, I’ll say that, you know, when you’re as an engineer, when you’re putting parts together, the parts do what they’re supposed to do. When you’re building an organization with people, you can leave a lot of space and the people will just fill it. It’s magic, right? How things just work because people figure out how to talk to each other and interface and they’re intelligent agents instead of just parts that are, you know, sitting in a place. So it’s amazing to watch an organization come together. So recently we
won a contract with the Pentagon, a $3 million contract that we have to execute. And we have a very high bar for the people we hire. So hiring is taking us a while. And recently, we had a full team meeting to discuss what we could do with.
the people that we have. How are we going to make all of this work given that we have this contract, we have to fulfill it, we don’t have enough people. And so we did a brainstorming session and then at some point, you the team was talking, I was sitting, someone else was leading the facilitating the discussion, I was sitting there in the meeting and then I had to take a phone call and so I left and then I came back and the team was like, okay, we’re ready to give you an update of what we’ve decided. We’ve decided while you were gone, we have reorganized the company and
cancel a bunch of projects. I said, what? So then the team sat down and they explained to me what happened. They said, OK, we’re working on this first generation design, but we also have the second generation design that is showing so much promise. What we’d like to do is to stop working on the first generation design because the second generation design is so much better. It costs less. It’s manufacturable. And we can take the second generation design, put it across all of our projects, and leverage the infrastructure, leverage the software design.
Hiten Sonpal (39:34.696) And that way we’ll just be more efficient as a company. I was like, okay, that kind of makes sense What it meant is that some projects the schedule
would increase because we moved to the second generation design, but it would be a better product for the customer. So I saw how you win some, you lose some, but overall the company is more efficient. And then the second thing they wanted me to do is they said, we want to reorganize the company so that the cylinders team, the team that’s making cylinders, is the engine and it has internal and external customers. So that the work that we’re doing for the Pentagon, that team is actually a customer of the cylinders team. And that way the cylinders team can start thinking
about all the cylinders and say I can make this cylinder for this one internal customer and that one external customer and I can build that cylinder for those two external customers and so we can start optimizing at a macro level and that also made sense to me and you know that you’ve been successful as a leader when your team feels comfortable taking that level of ownership.
and running with it, that you’ve created this space for them to do something that’s magical. And now they own it. They own the decisions that this is not me saying we’re going to do this. This is them saying this is what the business needs to do for us to be successful. And that is a flip of a switch that every leader should experience when they aren’t running the business, but their team is running the business and the team is using them strategically as a chess piece. One of the leadership blogs that
read recently that resonated with me is a really good leader isn’t operating their team like they’re on the leader’s chessboard, but really the team is operating their business using the leader as a piece on the chessboard. So for example, if a team is having a delay with a customer, then they might call me in and say, Hitan.
Hiten Sonpal (41:35.687) We promised this date to the customer. We’re not able to meet it.
Will you please call the customer and deliver the bad news? And we have to figure out what things we might offer them in return. We might say, hey, we’re sorry it’s late. We’ll throw you in this thing for free. Or we’re adding this one other feature that we think you like. Or even though this is late, our next delivery is going to be on time because we have figured out these bugs. So can you please do that? Someone else might pull me in and say, we need to accelerate the development of this. I need you to approve a spend of several hundred thousand dollars.
to find a way to say yes and also find the money in the budget to make that happen. So that’s when things start working because the team is now using me to make them successful as opposed to me trying to push them around.
Federico Ramallo (42:22.957) Right, right. That’s amazing. I used to have a team where they would ask me, what do you want us to do? Right. So they would delegate the decisions to me. And it’s hard when you don’t see all the details to make decisions. mean, and I would make decisions based on the information that I had. And then they would say, well, this is wrong, but you approve it. So, you know, which kind of makes sense, but
Hiten Sonpal (42:35.069) Yeah.
Hiten Sonpal (42:43.944) Yeah.
Federico Ramallo (42:52.357) It took me a while to figure out how to change that attitude because at the end of the day it was their attitude but it was the result of my leadership.
And it took me a while to learn how to build that environment of trust where they could speak up and say, hey, these are the issues, this is how we’re going to prove it, right? And to your story, I learned how to be the last one to speak. Because if I speak an opinion in that meeting, right? I mean, you were talking about leaving the meeting for having a phone call, right?
If you say something at the beginning of the meeting, then your comment kind of leads the way for everybody’s opinion, right? And as a doer, hard because we want to do, we want to say things, right? So, you know, that’s what I say. It’s a negotiation with our ego first, right? And then everything else, right?
Hiten Sonpal (43:55.257) Yes, brilliant input. It is first and foremost an negotiation with ourselves. Am I ready? Can I let go? And can I still accept what is happening as my responsibility?
Federico Ramallo (44:08.621) Right. Right. And the you were talking about a budget, a budget of millions of dollars. But I think that the stakes, you know, for every level of corporation, right. Any any founder from small to big, you know, the stakes are high all the time. Right. Whether it’s a thousand dollars budget or a million dollar budget. Right.
So it feels like a lot of money for every founder throughout the way. So one of things that I’ve been trying to do more is to detach myself from the situation emotionally. So that way I can look at it without that emotional component. So what would be the best decision?
So anyway, I wanted to mention that because I mean, for the audience, that much money could sounds a lot, right? But it’s always a lot no matter where you are on the journey.
Hiten Sonpal (45:21.283) That is correct.
Federico Ramallo (45:25.305) So,
You’re also pursuing a master in machine learning, right? While running a company, that’s amazing. How does that decision came up to be and why have you decided to study machine learning?
Hiten Sonpal (45:33.188) Yes.
Hiten Sonpal (45:45.987) Yeah, so I have a computer engineering degree and so I understand electronics, understand software development, but what’s happened over time is that the software complexity has expanded substantially and so a lot of my knowledge and understanding has shrunk just because the field has gotten to be so big. And so one of the key areas was what’s happening with these statistical methods that are creating so much new value.
and I wanted to figure out what that was and I felt like no matter where I went or what I did, that knowledge would be helpful. Even if it was just to have conversations with other companies or even to talk to my customers that were trying to use these systems. More recently, I’ve had to pause the masters that I’m pursuing simply because the work is taking so much time. But I do intend to circle back and finish it as soon as I can.
Federico Ramallo (46:40.687) Right, Yeah, it’s a challenge of being a founder, being a CEO that you have so many things in parallel happening that you have to be there, right? It’s hard to have the consistency of time, right? Yeah, I had the same issue. I was unable to finish my college.
because I was pulled into work and everything was interesting. So more interesting than school, right? So wasn’t able to finish. Yeah, and then I got into leadership roles and I was thinking, well, maybe that’s when I would have more time, but the amount of interruptions you have throughout the day increases and it’s harder to own your time, right?
Hiten Sonpal (47:33.227) Yes.
Hiten Sonpal (47:37.196) That’s right. In leadership roles, oftentimes, you know, your time is not your time. It’s the company’s time and it’s the other people’s time. It’s your report’s time. They need your time. And so sometimes your own time becomes such a luxury.
Federico Ramallo (47:51.463) Right, right. Yes, yes. you know, I’ve been in situations where I will be firefighting and, you know, whatever you’re doing, you have to stop it to fix an issue that it’s life or death for the company or for the team or for the project. So, yeah, and, you know, circling back to a conversation about leadership, I learned…
how to be able to build teams that run autonomously so they don’t need interventions. And I would call them interventions that way I was able to speak with the team to say, the more I intervene, the more our dynamic is not working. So let’s figure out how to change that. Because I want you to be successful by being autonomous.
Hiten Sonpal (48:49.3) Yeah, I also have those conversations with my team usually, which is if you want more autonomy, which a lot of people do, they want more space to operate.
I actually have a specific communication style that I prefer that gives me confidence in giving people autonomy. So I have people on my team that are very, capable, but sometimes they don’t create confidence in the way they communicate. And so I have to remind them, like, if you want me to give you autonomy, please come with, here’s a problem, here’s what I think needs to be done, and here’s where I need your help. And if you don’t need my help, that’s totally fine too. But you’re just telling me, here’s a problem, and I’m working on it, or I have a plan, and then just give me confidence to let you go. And if I have input to provide, I’ll provide
provide input, then you can own it. But if you just come to me and say, here’s a problem and have no opinion about it, then I’ll start asking questions. And before you know it, I have the ball and I’ve taken control. And I don’t actually want to do that. I want you to run with it. Steve Jobs had this really interesting quote. He said, at Apple, we don’t.
hire people to tell them what to do. He meant, of course, his executive team. We hire people so they can tell us what to do. And so one great example of that is when he hired Tony Fidel. Tony Fidel then brought about the iPod, was something that he was working on the side, but he pushed Apple into making the iPod, and then that turned into eventually the iPhone. And so that’s one example of where you go hire the people who tell you what to do.
Johnny Ive to help him figure out the industrial design and they came up with the MacBook Air and all these amazing designs and he hired Tim Cook to tell him about supply chain so it wasn’t that Steve was hiring people and telling them what to do he was he was hiring people who were telling him what needed to be done
Federico Ramallo (50:42.406) Right, right, and that’s difficult. mean, that’s an evolution of being able to let go and trust people, right? Definitely, yes. One of the things that I teach the people that I hire is about what I call proactive communication. This means, you know,
because we have engineering teams working in Mexico for companies in the US. So we have the cultural difference, sometimes the time zone difference, although we’re central, so we can cover the West and the East Coast. what I talk about with them is proactive communication for me means tell your intentions. What are you going to do? You read requirements, you’re going to build software. Before you build it, express what you’re going to do. I am going to walk.
towards that direction, right? Because we don’t want to be interrupting the manager all the time. We also don’t want to make assumptions and be on the wrong, right? So we validate our assumptions, we put them in writing on a ticket or whatever it is. If, you know, sometimes a chat is a good place to put it, but basically this, describe where we’re going. And then if we are on the wrong direction,
I’m sure somebody will come up and say, wait, you’re doing, you completely misunderstood the point, right? So let’s talk about it, right? We save people’s, you know, people’s time, right? But we’re also try to reduce interruptions for the manager because if we interrupt, if everybody interrupt, then, you know, the manager is one guy, right? Or one woman, right? One person. So we, we, we describe what we’re going to do. Then we do it and then we validate again, right?
Hiten Sonpal (52:16.163) Yes.
Federico Ramallo (52:38.458) This is what we’ve done. This is what we’re going to do, right? Or I am going to do. And that saves so much time and friction for everybody, right? And it’s amazing when you see that working and you see how it changes not only the perception of the work, but the actual work, right? And the impact that we can have on these companies. I think that’s amazing. And for me,
The most important reward is to see these engineers flourish and being able to grow and take more responsibilities, make more money, being able to buy a new car or a house or whatever it is, right? And have a family. That for me, it’s the more rewarding thing of my work.
Hiten Sonpal (53:33.067) I think what you said makes a ton of sense because when you think about the conversations that are happening between managers and individual contributors, they’re task-based conversations. And when you think about the conversations that need to be had between managers and leaders, they’re direction-based conversations, right? And so what you described is, where are you going? And then you can collaborate on that level as opposed to, I’m digging a hole here. That doesn’t tell me anything. It’s like, hey, we’re building this road and it’s
part of building this road, we need to dig a hole here. Then in that context, you can say, it make sense to dig a hole here? If someone just randomly says, I’m digging a hole here, it’s like, you really don’t know. That might be useful to an individual contributor who’s told dig a hole there, and they can go dig the hole. It’s good enough for them. But the leader-manager conversation is very different than the manager-individual contributor conversation.
Federico Ramallo (54:23.684) Right. Right. And then the second layer that we’ve been working towards is this idea of understanding why are you building what you’re building? Who is helping? Right. What’s in it for the user? In our case, it’s a user because it’s for companies, right? But I mean, what is the first purpose of the engineer, right? To build something, right? I build a bridge. It’s the most beautiful bridge, right?
Hiten Sonpal (54:34.029) Yes.
Federico Ramallo (54:52.966) is great, but nobody uses it. So what’s point? So the second layer of purpose that I found is when we’re building, it’s changing people’s lives. And then that reward, it’s so much more powerful than building something that it’s mechanically perfect, but nobody uses it.
Hiten Sonpal (55:06.005) Yes.
Hiten Sonpal (55:20.011) That’s right. It isn’t, you may have built something new, but you really haven’t created anything of new value unless there’s someone on the other side that says this is useful to me.
Federico Ramallo (55:30.5) Right, right. And in software it’s harder to see because it’s zeros and ones, right? But we try to convey that to software engineers, right? And…
We have so many parts that it’s hard to see it until we start broadening up their vision. And you can see it on their face when they see it.
Hiten Sonpal (55:58.507) That’s right.
Yeah, and so whether it’s software or it’s hardware,
Federico Ramallo (56:02.225) So yeah, go ahead.
Hiten Sonpal (56:06.311) finding ways to bring the engineer closer to the customer always sort of closes that loop where as a leader, you don’t have to be telling them what to do, right? They see the customer’s pain point. They go to the customer side. They’re like, I see why I’m doing what I’m doing. And the customer doesn’t care about these three things, even though they’re so cool to do from an engineering perspective. So in some ways I found that that actually helps with the leadership problem rather than saying we’re going this way. just take them to the customer side and then the engineers come back and say, we should go this way.
say, course, let’s do that.
Federico Ramallo (56:36.73) You
Right, right, because now you have a team that is actively engaged on fixing the user’s problem.
Hiten Sonpal (56:47.134) Thanks.
Federico Ramallo (56:48.209) Yeah, and it’s easy to kind of, I mean, I love the technology. I love to get into, as we were talking before, talking about the gears and the belts and, but, and I can keep talking about that because I find it very interesting, but also talking about the impact to the people that the products are being used. I think that for the engineering team becomes much more important.
because it builds a purpose, higher purpose.
Hiten Sonpal (57:23.024) And that’s where people come to work to be inspired, right? If you can give them that higher purpose, the output you get, the productivity you get is incredible. And if you give people a higher purpose and they’re not inspired by that, then unfortunately, they’re often the wrong people for your organization, right? They deserve to be somewhere where they can be inspired.
themselves, like something that where they feel like I’m not doing this for a paycheck. I’m doing this because it’s also creating an innate sense of self-worth because I’m doing something that’s important, important to other people, important to me. It gives me more satisfaction than only the paycheck I get.
Federico Ramallo (58:06.183) Right, because then the money is important, but the purpose becomes much more important. I work with my team on providing them a mission and a purpose, right? Because you can have values, but the values can be on the wall, and if you’re not applying them on a daily basis, what’s the point, right? But I find that the purpose is much more significant. And then, yeah, I mean…
Money becomes important, what I do is I do performance evaluation every year or so to track their improvements, rather than them begging for a raise. I try to proactively schedule it so we can talk about it and have those opportunities.
Hiten Sonpal (58:58.086) Sounds amazing. I already want to work for you.
Federico Ramallo (59:01.127) So I wanted to ask you about AI and robotics. The revolution is accelerating and what do you think people are still trying to get about AI and robotics? What are they getting wrong? And where do think it’s getting ahead?
Hiten Sonpal (59:27.534) Yeah, I’ll say that there are several things that are where people are getting ahead of things, which are really impressive. So for example, when you think about autonomous cars, right, you think about what Waymo is doing. They have a really good architecture. They are sensing the data properly, not like Tesla is doing with just vision systems. They’re actually using LIDAR and able to get a real physics model of the system. And then what they’re doing is they’re building an increasing set of
corner cases and solving them and simulation and building a huge data model and they keep training their retraining their systems. So the systems keep getting better and better and better. And they also understand that the system is not going to be perfect in all cases. As long as the system is safe, that’s fine. If you have issues with navigation or if you have issues with decision making, they can always kick it up to an operator that could come in and say, don’t go down that path because there’s a party over there that’s happening with that street is closed. Come back down and turn left here or something like that. they’re, they’re,
how to blend the two, where getting the right data from the ground and then also augmenting it with humans where necessary. So I think that part is going to continue and as they add more and more miles, their cars are going to get safer and safer and safer because they’ve started with the right architecture and they’ve admitted to themselves it’s not going to be perfect, which is great.
Things that people are getting wrong. Well, one other thing about LLMs, and maybe I should talk about this, is LLMs are doing awesome, but all the applications we’re using them now, or most of the applications where they make a huge difference, are places where a mistake is not a problem. if I want to build a website, and I ask the LLM to build a website for me, or an LLM-based system, the website, if it’s 80 % good, it’s a great start for me. I can tweak it, I can enhance it, I can remove the things that don’t make sense, but I already got to 80 %
that, you know. But then if you try to take it and put it into a system that has to be accurate 100 % of the time, well, that’s not going to work. So I think there’s an asymptote there where just like with autonomous vehicles, at some point, some human curation is going to be required to make sure that the output is great. Either you need the curation on the output side, or you need the curation on the input side to make sure that everything that’s going into the LLM is accurate and makes sense. And then there’s enough of it to build an interesting model.
Hiten Sonpal (01:01:50.233) Things that I think folks are getting wrong or underestimating, well getting wrong, I humanoid robots indoors to me doesn’t make a lot of sense, especially in consumer applications. People have this vision that there’ll be a humanoid robot in every home and I think that’s a tough, tough market. First of all, because it’s very cost sensitive, right? Consumers can’t afford to pay that much. I mean the most expensive thing somebody has is a house, then their car.
Right? And then appliances and then a laptop and pretty soon you’re at the $500 price point. So if you try to think of inserting something in that stack that’s close to the cost of a car, it had better deliver as much value as a car. Right? And when you think about how much value a phone delivers and when you think the car is, right? A phone for a thousand dollars and a car for $20,000, you put something in the middle. It’s got to deliver that kind of value. And when you think about a human art robot, it’s going to be expensive. And then second is being put into a very complicated
complicated environment. Homes are about the most complicated environment you can put things in because there’s so much variety. There’s so many obstacles. Every house is different. And then some homes are old and not built according to the right standards. So you have thresholds where people can trip, robots can trip, and then humanoid robots in general that have legs aren’t really stable between steps. And so you try to combine an environment that’s complicated with an architecture that’s not stable most of the time around baby
and elderly people in steps where can roll down the, where can, a robot can roll down the steps, land on a separate floor or fall over a balcony.
That is a very, very difficult problem. And so people who are thinking that they’re going to get a humanoid robot in their home, that is at least 10 years out. Because even the safety standards for those systems are three years out. And they have to get revised multiple times. And then in the industrial sector, humanoid robots indoors, again, now you control the environment. You can control a warehouse. Why would you need legs? Because they’re already flat, because you’ll have to get humans in there anyway. And you have other machines that are wheeled, that are moving around.
Hiten Sonpal (01:03:59.063) You’re not going to use robots to carry 20 ton items. You’re going to have things with wheels that move those things around. So when it comes to factories, any company that builds a humanoid robot with legs to solve a problem will immediately get, and if they’re successful, they will immediately get replaced by a company that builds the same robot with wheels. That’s a lot cheaper. So I see that humanoid robots indoors, especially in homes, but even in factories, don’t make a lot of sense. Now, when you get outdoors, completely different pictures.
Because now wheel machines are difficult if you have a yard or gravel or places where there are hills and so on and so forth There a human robot makes a lot of sense and making them as large as possible So you can move as much as possible also makes sense. You wouldn’t want to make them human size They may be human shaped but you’d make them really really big and when you do that I see a place for rises belt hydraulic actuators that you’re not going to be able to put a standard motor on the shoulder of a very large human
because the weight scales as a cube and so it’s going to get very heavy. So you want to start looking at how humans do muscles. You’re to start thinking about things pulling as opposed to trying to build a motor with very very strong magnets packed into a small space which would be very expensive. And also those motors are like you know just the motors that are only like 75 % efficient so you try to drive too much power into one of those joints with the gearboxes and everything the harmonic gears super inefficient.
It makes sense for our technology to get involved. So I’m excited about humanoid outdoors, but not as excited about them indoors. And then finally about humanoid robots, folks think that they’ll be able to solve manipulation in the same way that they’ve been able to solve locomotion, which is moving around. But actually, walking is a much simpler problem than manipulating. Like when you think about
putting your hands in your pocket and being able to pull out keys instead of your phone or your wallet just by touch. Think about what is required to get to human level dexterity. That is a huge journey. That does not mean that you can’t have robots with arms that add value, but that curve that we have to get above is a much harder problem than actually the leg problem, which is in a lot of ways the easier problem.
Federico Ramallo (01:06:21.968) Right, right. mean, you were talking about that early when you were talking about sensitivity, right? Robots don’t have sensitivity. That has been an issue. And the more strength you put into it, the more dangerous it is, right? They don’t have the dexterity. That’s a very complex problem to solve.
Hiten Sonpal (01:06:37.786) Yes.
Hiten Sonpal (01:06:43.045) Yes.
Federico Ramallo (01:06:45.212) Yeah. And the, I think that the robots, know, indoor robots for houses looks cool, but they’re, you know, I agree that the level of complexity for that is so high, right? I like what Amazon has done with the warehouses where it’s all robotics. So you don’t know when one robot or the ecosystem ends, right? It’s a whole thing, right? And they were able to feed out,
all the complexity around it, but it’s a controlled environment. The closest I’ve seen for houses is a dedicated room. It’s actually a house, it’s kind of a hotel where robots would stay inside there and they would do the laundry. So they control the whole room, basically. And they fold clothing, they wash it and everything. And then they deliver it through a window.
But they can, and they’re not too late, they’re willed, right? Because it’s the most efficient motion, right, for robots, at least for now,
Hiten Sonpal (01:07:53.346) and cheaper.
Federico Ramallo (01:07:55.067) And cheaper, yes, it’s more efficient and cheaper to build, right? Right, right. And for outdoors, I’ve seen wheel with legs. Basically, you have wheels to move around, but then legs to kind of run over an obstacle or just, they can even jump, which for wow effect sounds great, but for actual application, I don’t know how much useful is that now, right?
So yeah, I think that for that type of environments, it’s becoming interesting, right? Like I’ve seen robots doing farming or flying over it, right? And I think that’s a great use case for AI and robotics now, yeah.
Hiten Sonpal (01:08:47.278) Yeah, a four-legged system with wheels is very, stable.
Federico Ramallo (01:08:54.76) So I wanted to ask you about what advice would you give to your younger self?
Hiten Sonpal (01:09:04.536) think it’s actually very, easy. would say take more risks, fail more often. Because I think that so many of us…
don’t realize what we’re actually capable of. And it isn’t until we’re put into a difficult situation that we rise to the occasion. And there are many difficult situations where if we don’t rise to the occasion, no harm. You can always recover. For example, if you take on a task that’s something that you haven’t done before, oftentimes if you don’t succeed,
It’s no big deal, you’ll learn some things. But if you succeed, boy, you can get on to the next one and the next one.
I mean, every founder or every person starting a company goes through this, right? They start a company and they try to make it successful. And during that whole time, they’re learning skills because they’re put into this situation where they’re constantly learning. But folks don’t have to take it that drastic. At any point in their career, they should be looking for more responsibility and they should ask for it. And they should take on the responsibility even if they can’t get the pay because the pay will follow if they have the skills. So something that worked well for me is, you know, I
I would be doing the work of the next level up before I got the title and the pay and I was advancing in that time faster than almost all of my peers because I wasn’t trying to negotiate Give me this salary. Give me this title. I just was like give me more responsibility and more problems Let me take more off your plate to my manager or whoever it was. What more can I take on? How can I do it? And then similarly I was going to my direct reports and saying
Hiten Sonpal (01:10:50.123) do you want to take on more because I have more work that I can handle? Let me give you some assignments that are choice that will help you develop. And so you find that there are ambitious reports that you work with and then you give them more and you take on more and your career advances and at least for the ones of your reports who are ambitious, their career advances with you. So it doesn’t have to be jumping into starting a whole new company. It can be anything from just finding ways to take on more.
Federico Ramallo (01:11:18.278) Right, I completely agree on that. was many times was doing the work that was not on my title. And the way that I explained this to my team is the crushing weight of the results is so heavy that you either get promoted or you get a new job offer. Right? Those are the two possible outcomes, right?
Hiten Sonpal (01:11:46.003) Yes, that’s right.
Federico Ramallo (01:11:48.808) Yeah, and I say not in a way to not taking hostage the company or it’s not extortion, it’s the reality. mean, you’re doing additional work, you’re having additional responsibility, you’re gaining the experience, so somebody’s going to notice, right, whether in the company or outside.
Hiten Sonpal (01:12:06.045) Yes. Yeah. And if they don’t notice in the company, then you just apply for the jobs that require you to have that experience. And you show up there and you’re like, have all this experience, you know, and then they’re like, okay. And then at that point, you negotiate because you already have a job. At that point, you negotiate and say, well, this is what the fair market value is. so within a year of you taking on a new problem, either your company pays you or somebody else pays you. Right.
Federico Ramallo (01:12:14.248) Right.
Federico Ramallo (01:12:33.085) Hiten, I truly appreciate you being here today. We are running out of time. I appreciate you agreed to stay a little bit longer and do a little longer interview. I wanted to ask you one final question. What is the future for RISE Robotics? And if you have any final remarks before we wrap it up.
Hiten Sonpal (01:12:58.266) Excellent. So I want to talk about four things. I’ll talk about our three business areas, and I’ll talk about fundraising. So one of our business areas is around working with the Pentagon in the United States. And so there,
Up until this point, before I joined, the company was only working with the Air Force. After I joined, I asked the team to take on other customers at the Pentagon. So we recently landed a contract with the Army. We executed it and they asked us to submit a proposal for a follow-on contract. And this year, we’re looking to get into the Navy. We found a couple of champions in the Navy and they have problems, pain points that we want to go solve to help them improve the quality of life of people that work there and defend the country.
take risks and have basically committed to putting their lives on the line. So that part of the business is growing well and as I mentioned to you before we won a three million dollar contract extension with the Air Force and that’s going to keep us busy for the next 18 months on that area of the business.
On the cylinder side of the business, that’s our core technology which we sell to OEMs, anyone that wants to make a hydraulic free equipment. There, we landed our first commercial sale in the second half of last year. It was to the oil and gas sector, so we’re pursuing that opportunity. That opportunity has an $11 billion TAM. That’s a vertical that’s going to keep us very, very busy, but adjacent to that, we’re also looking at marine, maritime, and food production very similar.
So we’re going to be patient about taking that on. Like I said, the oil and gas sector is going to keep us busy for a long time. And it’s also something that our employees enjoy cleaning up. So it’s going to be a win-win for all of us. Our third business area is Lift Gates. So we have a partnership with the company where we white label, call Anthony Lift Gates. We love them as a partner. We white label their product as our own lift gate. And we just finished our second pilot with a major beverage company. So we’re excited about that.
Hiten Sonpal (01:14:56.527) doing a third pilot and taking that into production in the second half of the year. So those are the three areas of our business and we’re moving along on all three of them. And the last area of course is that because we’re a startup we have to keep the business funded. Certainly the contract we have with the Air Force will keep us funded but we want to go faster than that so we’ll be raising money. Last year what we did was we took our institutional round and we opened it up to the public under a rule that was passed by the Obama
administration called regulation crowdfunding, which allows startups to raise up to $5 million in 12 months. And when we opened our round, had, took us about a year, but we had $5.7 million worth of interest. It was way more than we could take. We’re only allowed to take five. And we were the number one regulation crowdfunded campaign of last year. And so this year we have to, we’re opening up a new round so that we can take the overflow $700,000 worth of
investments, put it into the new round. And we still have interest and demand from a lot of other people. We have a wait list that people have signed up for. So we’ll be raising money again. And we’re giving an opportunity to people who believe in the technology to get in early. These days what’s happening is that a lot of companies like SpaceX
all the money has been made before it goes public. And so the people who are buying it are just providing liquidity to the early investors. They’re not getting in early enough to create value. And so what we’re able to do is both raise money for the company, but also create opportunities for investors who wouldn’t have access to something this early.
So I’m excited about that. So that’s what’s happening this year. if folks are interested in checking us out, they can look at our website, riserobotics.com. And if they’re interested in what the investment looks like and how to do that, they can look at invest.riserobotics.com. It’ll take them to our campaign page, tell them our business goals, our expected financial growth over the next five years, all our product lines, and what we’re doing and our recent successes.
Federico Ramallo (01:17:07.465) Amazing. And we’re going to leave the links at the description below so people can go and check it out. Hiten, thank you much for joining us today. It’s been an honor.
Hiten Sonpal (01:17:18.508) Federico, this was amazing. Thank you so much for your insightful questions. Really appreciate it.
Federico Ramallo (01:17:23.977) Thank you.
Hiten Sonpal (01:17:25.814) Bye bye.