Powering the future of humanoid robotics-transcript
Dale Tutt: Welcome to the Industry Forward Podcast, the show where we examine industry and technology trends with the help of various experts from around Siemens and beyond. My name is Dale Tutt, and I am the Global Vice President of Industry Strategies at Siemens Digital Industry Software, and I’m your host for today’s episode. On past episodes, we’ve been talking a lot about humanoid robots and what’s been driving them. And why now? And one of the key constraints that we talk about, or maybe the key enabling technologies that we’ve been talking about for humanoid robots is the fact that you have to have semiconductors and the software that is used to drive the humanoids and the electronics.
Another aspect that we haven’t talked about as much yet is batteries. And one of the key factors about humanoid robots is the fact that they can move about freely in your factory. You don’t want to have power cords stringing all around the factory floor, so you’ve got to have good batteries and they have to be able to recharge fast. The battery technology can become one of the most critical constraints and enablers for humanoid robots. If you don’t have good battery performance, this can just result in either short usage time or long downtimes while the batteries are being recharged, and so the assets are now setting idle. We want to dive into this today. And so we’re here to discuss these constraints and how to navigate them is our own expert from Siemens, Puneet Sinha. Hello, Puneet. Why don’t you introduce yourself to our listeners?
Puneet Sinha: Hey, Dale. I’m glad to be here to everyone. I am Global Head of Battery for Siemens Software. And as part of this role, I’m responsible for Siemens software solution development, go-to-market, and partnership when it comes to battery ecosystem.
Dale Tutt: Well, welcome to the show, Puneet, and I’m really, I’m glad to have you here. I appreciate your time today, and I’m looking forward to the discussion. So as I mentioned, you know, we’ve talked a lot about humanoid robots and the importance of the humanoid form. A lot of times people ask, Why do we want to use humanoid robots as opposed to, I would say, normal robots? And part of that is that the factories and the products that we’re using the robots to assemble have been designed around the human form. And so you want to be able to take advantage of that form in these environments without having to redesign your products in your factory. Part of the human form is the ability to move about freely. And so I think that’s critical for the behavior of the humanoid robot if you’re going to truly be mimicking the operation of the person.
So Tell me what your thoughts are. As I mentioned earlier, you gotta have the battery, you gotta have a good battery that allows a humanoid robot to walk around and go two hours before your break time, so to speak, like the human form does. But you know, so the battery really is a critical aspect. So what are your thoughts about the importance of batteries as it relates to the humanoid robots?
Puneet Sinha: You said it very well, Dale, that when it comes to humanoids, There are a few key enabling technologies. Of course, semiconductor, the other one is software, and then, of course, the third part is battery. The way I look at it, battery is the source of fuel, source of energy for humanoids. And especially when you look for these humanoids to work in industrial environment to bring the level of productivity that is expected from these humanoids They need to be working non-stop. And that’s where a key constraint starts to come.
What is the power source that can allow these humanoids to work efficiently around the clock while making sure they’re mobile and able to do the things that they are supposed to do in these safety-critical environments? If you look at the humanoids of today, yes, they are powered by batteries. The kind of batteries that are used there are typical lithium ion batteries, which are the kind of batteries we have in many other devices today. But the challenge comes to the operating time. Some of the humanoid companies who are already putting their humanoids in the factory environment, they are reporting that their humanoids are working somewhere between two to four hours before they need to be recharged. Two to four hours. I mean, as good as it sounds, it is really limiting. for a factory environment, that becomes a very critical criteria for what is the right battery that allow these humanoids to do their work round the clock or for a really long period of time before getting back to recharge.
And this is where the whole industry, and when I say whole industry, there are two parts of the equation. One is, of course, the battery suppliers, but also the humanoid companies who are building these humanoids, how they need to collaborate to make sure the right energy source, the right battery technology is made available to them that can be integrated easily into these systems, ensure the operation of the humanoids while making they are safe, but at the same time deliver the right charging mechanisms so these humanoids can be charged as fast as possible.
Dale Tutt: I have a few follow-ups on that. And I made a bit of a joke about, the humans taking break, but that is kind of the reality that, in most factory environments, actually most office environments, doesn’t matter, people will typically work for a couple hours and then they’ll take a 15 minute break and or they’ll take a half an hour lunch. So is it really that critical if the humanoid robot can go 6 or 8 hours? Or is this a situation where if they, especially if they’re working alongside humans, you know, can they go, can you go to a charging station and plug in?
I was watching Star Trek the other day and the Borg and the recharging, regenerating chambers, like kind of imagine these humanoids walking up and plugging in somewhere. Realistically speaking, what is a sweet spot? Because battery size does matter. And if it gets too big, then it’s heavy and it’s extra weight. And if it’s too small, then you have the problem as you’re describing. You know, what is a sweet spot is it a reality that you can, they can work for two hours and they can recharge for 20 minutes? What are the economics for these companies as you’re thinking about that?
Puneet Sinha: Yes, as you said, I think it comes down to how to optimize how much energy or how big of a battery you’re putting in a humanoid while making sure the humanoid is not just bogged down by the weight and the volume of the battery. But at the same time, it is, not going for frequent recharge. So there is optimum point, and of course, it depends on different applications. But the way I look at it, if you see the humanoids are supposed to work alongside with humans in the near future, that means the humanoid should be able to make at least the energy profiles of humans. At least. At least, right?
So if you look at that, seemed to me that having at least a battery pack that can make this humanoid work for five to six hours before it is going for a charge, perhaps ideally eight hours, but between six to eight hours can be a sweet spot for a lot of those industrial applications. Why I’m always insisting on that because, hey, the humanoids, if you look at the, think of a future, and humanoids are part of the part of our society helping, working alongside with us in different environments. Different application may allow humanoids to have different amount of energy, right?
So if there is a humanoid at home, perhaps it doesn’t need to have that much of energy or that much of how many hours it is working versus if it is in a factory where humanoids are being relied upon to do some safety critical things where the throughputs matter. Just to add one more point, if you look at Let’s say eight hours for the sake of simplicity. Today we are around two hours. One way to look at it is how can batteries go from two hours of operation time on a given charge while they are being fed in the same amount of volume, in the same volume and weight profile, how the batteries can have four times more energy?
And the other very important part of the equation also is how long does it take to charge them? and how does the charging profile going to ensure these humanoids are in the operation for a long period of time. So the charging time as well as the, you know, how long they’re going to operate on given charge, both of these are equal part of the equation that needs to be solved through the collaboration of battery suppliers, but also the humanoid companies based on which application they’re working for and how the whole humanoids is being designed.
Dale Tutt: Okay, awesome. we could probably go on and on about that. I was thinking about like all the different battery usage profiles you can work on, but obviously that’s part of that optimization. But part of that optimization as well is how is a human-owned robot consuming the power? You know, there’s really, I guess there’s a couple things that are actually going to be consuming the power. One is the motion of the robot itself, the arms moving, the legs moving, but then the sensors and the electronics.
And so obviously, the electronics can be a big consumer of the power, especially when you start thinking about AI and the discussions that we have around the power needed for some of the AI applications that are on these chips. So, we kind of always joke about the amount of computing power today versus what we had in the past. in 2026, there’s the same amount of computing power in a smartphone as the rockets that went to the moon in the 1960s, and they had massive boxes all over the rocket. And with all the miniaturization of electronics, it has also changed some of the power consumption.
So, and we’re going to continue to see this, I guess this trade-off. You have the miniaturization, but you have this higher demand on the electronics for the physical AI and just, you know, for running electronics as well as all the sensors. I think people sometimes will forget that it’s not just the motion, you got to have all the different sensors that are also consuming power.
What are you seeing in the battery industry as it relates to how things like the miniaturization of electronics might affect the humanoid robots? Maybe what’s going on now, but also where things are heading with technology and the energy density and the improvements that are coming.
Puneet Sinha: It’s a very interesting question, Dale. And the way I look at it, there are two parts to the question. One is how battery companies are looking at humanoids and their energy profile and their speed at which humanoids are growing and the kind of functionalities companies are looking to put in there. And then mapping that to what kind of battery technology is needed to match that kind of growing sophistication of autonomous behavior these humanoids are looking to do. And this actually brings me back to CES show earlier this year. I remember when we both were there.
One day I was taking a tour and there was a booth close to our Siemens booth, and these guys were talking about batteries for humanoids and a lot of robots. And I had a conversation with one of the executives there, and that’s pretty much the same question that I asked him that, hey, what is keeping you guys up? Because in certain ways, it’s a battery. Battery companies have a lot of experience in powering all kinds of vehicles by now, from electric vehicles to the consumer electronics every which were. So why this humanoid adds a new angle, or does it add a new angle of complexity?
And that executive of this company came back to autonomous power and the sensors and the compute engine these companies are going to put in there. And the implication of that on battery energy, is how it needs to be cooled. It’s a life cycle. And this actually is very similar to if you look at autonomous vehicles. If you look at autonomous electric vehicles, I mean, there are a lot of electric vehicles on the market for 15 plus years now. But by adding autonomous compute power, at least in the beginning days of when the autonomous vehicles were being engineered and designed, many of our audience can recall there were pictures of the whole trunk full of computers. That’s how the autonomous vehicles started, right?
And what that was the amount of power needed to do the sensing and computer inference was tremendously high, like a multi-kilowatt hour. And if those computers are running all the time, in the early days of autonomous electric vehicle development, there was a lot of power that battery needs to provide just to power these power-hungry computers. And of course, over time, as companies became more sophisticated, the advancement in semiconductor technology, the compute, some of that power demand has come down. And the way humanoid technology is today, it actually reminds very much to the early days of autonomous electric vehicles. There’s a lot of focus on, hey, bringing the right functionality to those humanoids. That means having all sensors, but not redundancy of those sensors, compute power.
But the disadvantage of that is on the power consumption, because some of these can be more than a kilowatt hour, more than a kilowatt of power consumption just because of this AI compute, which is a lot, especially when you look at a typical humanoid. When I look at it, when I’ve talked to some of the humanoid companies, And I asked them, Hey, give me some kind of a power distribution. So they said for a humanoid, for an industry application, when it is idling or standing, it’s like 100 watt. If it is walking, maybe doing some work, maybe 500 to a kilowatt. And if it is some heavy-duty work, it can go to like two to three kilowatt. But then just because of the AI compute and how much power that AI compute is going to have, it can be a wide range of distribution from 100 watt to more than a couple of kilowatt.
What that means is depends on the AI computer architecture of humanoid, the kind of sensor they’re using, how they’re computing, whether they are computing on board, whether they’re leveraging some of the edge AI computational power. You are looking at somewhere between 5% to 40% of power of battery just to power this thing.
And that has wide implications because 5% versus almost 40% of power going to power just the computer or the autonomous operation of humanoid is very wild from a battery company perspective in terms of what kind of battery they need to provide.
And this is the reason that increasingly we are seeing companies are investing heavily on solid-state battery, which is, you can say, a next-generation battery compared to where today’s battery technology is. And that is because it offers a lot more higher energy density. That means in this given volume, you can put almost three times more energy density, energy in the same volume, which is important.
But again, as I said, just doing a new chemistry, bringing a new design of cell and chemistry alone is not good enough. because it also depends on a close collaboration, understanding the power consumption profile of today’s humanoids and the future of those companies and having that collaboration with the battery companies is very critical. And this is what I’m increasingly seeing in the industry of humanoid companies and battery companies sitting down and drawing this on a piece of paper to understand where they are today and where they’re going to go next.
Dale Tutt: Okay, cool. Well, that’s It’s kind of interesting about solid-state battery. I think you said it was like three times the energy density. So that’s actually pretty compelling when you think about like, if you build a humanoid robot today, basically, I’m assuming you can put it in the same size pack with a new battery technology and kind of upgrade. Today, most everything is lithium ion. I mean, it’s what’s in automotive and everything else. And one of the things I learned recently was that if you have a vehicle that was built on an electric vehicle that was built on a lead-acid battery, that you can’t just drop in lithium-ion batteries in there because they have different discharge characteristics than the lead-acid batteries do. And so, like the motors, we’re talking about this in the context of a golf cart, that the motors, if you just drop in the lithium ion batteries, the lithium ion batteries will just kind of keep discharging at maximum voltage when you’re climbing a hill, whereas the lead acid batteries, the voltage will sag and you don’t put as much power.
So you end up burning the other electronics up when you have a lithium, if you just do a flat out replacement of the batteries. And for our listeners, if you think about this, you know, maybe many of them are used to lithium ion batteries now. And you were mentioning the solid-state batteries as the next step. So maybe can you share a little of the differences between what these are, what the technologies are? And a lot of times people think solid-state, what does that mean? so what are some of the differences? And then maybe think about like the considerations. Is it as easy as just, somewhere down the road, you just swap the batteries out or does that have some sort of an impact on maybe the power electronics that are inside the humanoid robot?
Puneet Sinha: Great question, Dale. So again, let me take a step back and. Give a brief overview of what is solid-state battery technology and how different it is from lithium-ion. Before I talk about differences, I want to talk about what is similar between the two. The good thing is there is a lot of similarity between lithium-ion battery of today and if we talk about solid-state battery of future.
The similarity is both of them use lithium as an ion that will be transporting between the two sides to generate electricity. Today’s lithium-ion battery that is in our cars, in electric vehicles, which is in our consumer electronics devices or powering energy storage system for grid. What they have is a liquid electrolyte. So the electrolyte is the medium that transport lithium from one side of the battery to other side to create the flow of ion that creates electricity. The benefit of that electrolyte is, hey, I mean, when ions are flowing through that electrolyte, they flow fast. That means that can generate a lot of power. The negative part of those electrolyte is that they are flammable.
If those batteries, for any reason, get into overheated conditions or in a condition where they are forced to operate beyond their typical operating window in terms of temperature or voltage, it can lead to fire or gases coming out and, you know, some safety events. So that’s today’s technology. And in order, of course, in order to avoid that happening, we have a lot of controls on those batteries so that it doesn’t go out of their temperature window and the operating window.
And these battery needs to be cooled so that if I go to the solid-state battery, then the biggest change is the electrolyte is going away from this liquid electrolyte to a solid electrolyte, as the name of the technology suggests, you know, of course, and there are different versions to that, you know, ideal situation is a solid-state electrolyte. And what it does is these solid-state electrolyte, they are non-flammable, so more robust in terms of their thermal robustness, their behavior.
They are less likely to catch fire. The other thing that substantially opens up the solid-state battery energy density and the behavior is use of lithium metal. So lithium metal is very, is a kind of a dangerous material to some degree because explosive behaviors in that. And then if you are using that to take lithium in and out of it, the way it behaves that if you have a lithium metal sheet, the way it behaves, it starts to create dendrites, those needle-like structures, the way the lithium deposits have come out. And if those dendrites grow, it can poke the structure of internal structure of a battery and short circuit the battery, which is absolutely detrimental. That is the reason in today’s battery, nobody can use lithium metal. Although everybody wants it, because if you use lithium metal as one side of the battery, it improves the energy density by orders and orders of magnitude.
So the big reason of why companies are looking for solid-state batteries, because if they, the premises, if they’re using the solid-state electrolyte, it suppresses this dendrite formation characteristic of lithium metal because of the things that are happening at the interface and then the microstructure electrochemistry. And that is a massive enabler for starting to bring lithium metal as one side of the battery. And that just opens up the amount of energy you can allow to pack in. And that is the goal that every company who is investing in solar cell battery wants to get to, that how can they have the right choice of electrolyte that ensures that dendrites don’t grow so that they can get to lithium metal and improve the energy density while making sure that now because the conducting medium is a solid, lithium is going to transport through a solid at a much, much slower rate than a liquid. So it still wants to provide enough conductivity so that there is a power that can come out of it.
I’ll pause here, but the point is, if I summarize, solid-state battery unlocks the opportunity of bringing lithium metal inside battery safely and ensuring that a lot of the safety-related issues that liquid electrolytes have to deal with, they are not there. And that opens up how much energy you can pack, how you can operate it, how much cooling it requires or the lag thereof. So to some degree, start to simplify the overall architecture of the battery pack.
Dale: Thanks for joining us on the Industry Forward podcast. And a special thank you to Puneet Sinha for sharing his insights on one of the most important enabling technologies behind the next generation of humanoid robots: batteries. From energy density and charging strategies to solid-state innovations, it’s clear that the future of robotics depends on advances far beyond software and semiconductors alone.
If you enjoyed this conversation, be sure to subscribe and join us next time as we continue exploring the technologies shaping the future of industry. Until then, I’m Dale Tutt. Thanks for listening.
Siemens Digital Industries Software helps organizations of all sizes digitally transform using software, hardware and services from the Siemens Xcelerator business platform. Siemens’ software and the comprehensive digital twin enable companies to optimize their design, engineering and manufacturing processes to turn today’s ideas into the sustainable products of the future. From chips to entire systems, from product to process, across all industries. Siemens Digital Industries Software – Accelerating transformation.
