How Skydio Ships Flying Robots With Just 20 Product Managers

0:00We've all probably heard of drones. We actually believe that we're on a journey from toys to tools to infrastructure. Alden Jones, VP of product at Skydio.

0:08Carlos: You are literally competing head-to-head with Google, the labs in terms of talent. Yeah, and on the Mechies, the mechanical engineers, electrical engineers, you're competing with... I mean, we're in the Bay Area to get the best talent, but we're definitely competing for it. When can we expect fully autonomous bits of robots flying? It is very near term. I think we're... Do you oversee both the software and the hardware components? Yep, and then we have software leaders on the autonomy side, SVP of engineering overall, also work with the hardware engineering and manufacturing leaders. In Ukraine, they're doing a lot more, unfortunately, really large-scale battle, and so today a Ukrainian soldier will not move on the battlefield without air support from a drone.

0:50Carlos: I'm curious to know what is your background in order to make you want to combine both worlds, because it's also like not something easy.

1:26Carlos: emergency calls a month and the metric for when nobody gets dispatched. 20 product managers, no travel budget, go get the signal. A$3.5 billion commitment funded by revenue, not debt. Let's get into it. Alden, welcome to the product podcast. Thank you for having me. Thank you for having me. I'm technically the host of the podcast, but you're hosting me at your office. Yeah, but you're in charge now. We'll see. And I had to come because what you guys are building is incredible, right?

1:55Carlos: It's opportunity to see. And I hope that we can take some of that magic and explain what you guys are really building here. But maybe let's start with the beginning. Why am I using the term robot, flying robot, instead of just drone? Sure. So I think we've all probably heard of drones and we think about quadcopters and many people probably think about them as toys. Many people who work with them maybe think about them as tools. We actually have believed that we're on a journey from toys to tools to infrastructure. And when you think about robots being used to do really important jobs, there are so many different things that we are building that isn't just the flying quadcopter.

2:35So we're building different kinds of platforms. We're building robotic base stations to host those platforms. We're building robotic arms. So in order to deliver the full solution, we don't just build quadcopters that people are probably used to. We're now building a suite of robots to solve problems. Give me a sense of the scale of your business today. So we're just under a thousand people, which includes manufacturing, which we also do here in California, and a super multidisciplinary team of experts. So we do everything in-house. We do everything from software, hardware, logistics, supply chain, manufacturing, chip-down design.

3:10and most of the team in engineering is here in California, but we actually have engineering offices around the world to capture certain disciplines like autonomy and cameras, some defense tech, and then go-to-market is all over the place.

3:24Carlos: And you guys cover the full stack, you're vertically integrated. Yep, vertically integrated, build the hardware, ship the embedded software, ship and build all the cloud software, all of it. And you as the product executive, how do you actually fit into this picture? Yeah. So day-to-day, the way I think about my team is we have sort of the core product platform team. So these are the folks who are looking after hardware, the core software experience, core autonomy, embedded work, communications or connectivity, as we call it, and product operations. And so they're supporting building a lot of the tech stack, essentially, that our business is built on.

4:01And then we also have vertically integrated PMs who, they aren't quite GMs, but they are really focused on delivering bespoke value for a particular use case for a customer. And so we do that across several different industries.

4:14Carlos: So you oversee both the software and the hardware component of the robots? Yep. And then we have software leaders on the autonomy side, SVP of engineering overall, and then also work with the hardware engineering and manufacturing leaders. So typically host CPOs that work in software products only, like Netflix, Airbnb, many others. Yeah. This is quite unique. And I'm curious to know what is your background in order to make you want to combine both worlds, because it's also not something easy. Yeah, it's a great question. So maybe just a spoiler alert, I'm not an engineer. I have a history degree.

4:51And I was really motivated at a young age to join the military. Come from a family who's done service, and I really wanted to try and give back. And so that's what I did earlier in my career. And I think in that, I got to see how technology needs to be used to succeed in really critical situations. and as I got back into the workforce, you know, I had some kind of fortune 500, I call it leadership jobs. They were kind of boring stuff. It's what you do when you get out of the army, but I eventually found my way to an innovation role at a telco company where I got to start to experiment with drones.

5:23And I think I had an opportunity there to really teach myself about technology and like got to, you know, start to work a little bit in software, start to try and code a little bit, learn about drones, learn about physics. And ultimately that, that led me to Skydio. And then And I actually started at Skydio as the customer success leader. So we had a team of zero. We built all the functions. So everything from pre-sale solution engineering to post-sale support. But this is just an incredibly technical company. And so I got to build really great relationships and learn a lot with engineering leaders and hardware software.

5:59As we grew the business and grew the company and bumped into all the problems, I got to learn about how this tech stack works and really dive deep to explain what was going on. and then ultimately I got the opportunity. I will also say that we have fantastic technical founders. So Adam's our CEO and co-founder. They're fantastic, brilliant engineers. Adam and I work quite a lot together. I would say that he, visionary-wise, absolutely hardware vision and probably a little bit more executing on the hardware side with my team and on the software side, probably a little bit more leadership there.

6:30Carlos: Cool. Now I want to dive a little deeper into the actual product. Yeah. You are giving me, I wouldn't even call it a demo. that was real. You were flying multiple drones, actually, multiple robots, sorry, actually at the same time, right? So what is the actual use case or main use cases that you are covering with your flying robots? Yeah. So maybe just to zoom out a little bit, you mentioned this before, we call ourselves a flying robot solutions company. And so at the end of the day, we are trying to solve business problems or critical problems for four kind of main use cases. So one is response.

7:01We call that drone as a first responder. You can think about that matches pretty nicely with public safety. ISR, intelligence surveillance and reconnaissance, which is a term that is used in the military essentially for gathering information on the battlefield. And so we provide use cases, essentially camera drones to collect intelligence and work with militaries. We have all departments under the Department of War in 29 Allied Nations. So very proud of the work that we do there. We work on security. So this is, imagine large, fixed sites that our customers have. So think of a data center or a dam or a power plan.

7:33They want to secure that today. Customers usually have roving patrols of human beings walking or in cars 24 hours a day. It's expensive. It's inaccurate. There's a lot of false alarms. And then inspection. There are trillions of dollars of infrastructure around the world.

7:48Carlos: And this is what I did in my previous role. Digitizing a portfolio of assets allows businesses to be so much more efficient. And I have a couple of stories that I want to tell, but I want to talk about DFR just real quick, real quick. So just census scale on DFR. So there's around 250 million to 300 million. It depends on how you count it. Public safety or 911 calls in the United States every year. Today, Skydio and Jess, about 3.5, 3.8 million 911 calls every month. We respond to 55 ,000 911 calls every month. And around the country, 25 million Americans live within two miles of a Skydio DFR dock.

8:27And that is up from 12 million, I think, just like eight months ago. So things are really growing. And the use cases there are essentially helping public safety get the best information as fast as possible to make the right public safety choices as they're helping the community.

8:45Carlos: I want to double click on that because I was asking you this off camera, the traditional alternative today is what? Yeah. So in public safety, there's two alternatives. The one that probably everyone is thinking about, you immediately reached this conclusion, which is right, is the helicopter. It's an interesting proxy. I mean, it's a flying camera. It's just really expensive. It's$3 ,000 an hour to operate. Going right of a helicopter is somewhere between$10 and$25 million, depending on how much money you want to spend. So it's a very expensive tool. and you just cannot respond to the density of calls that you have in any given city.

9:21And so one of the things that we do with DFR when we sit with our customers early in the pre-stage or the pre-sales process, we actually sit with them with their data and we map all of their 911 calls for the year. They give us parameters on which kind of calls and priorities they care about, and then we advise them on where they can best put docking stations so that they can most quickly respond to those things. The other alternative, which I know you've had, I've listened to a few of your guests. I know some of your listeners have done this, but if you go on a police ride along, which you can actually request from your community, if you ever want to do that, you can go to the police station and ask, is just the radio.

9:53They'll still have what's called a mobile data terminal. It's a little Toughbook kind of computer in their car, and they'll get a one-line dispatch, you know, suspect with knife threatening, whatever. That's all they know. No images, nothing visual, and we're such visual creatures. and maybe a radio call saying like, get there, or go off your this job and go do that. There's a lot of nuanced department to department. There's I think like 18 ,000 different public safety agencies in the United States. So it's a wide tapestry, but generally that's the use case, or excuse me, that's the workflow. And so it's just imperfect information in a really high stress scenario.

10:28Carlos: Let's talk about success metrics, right? Because here you are talking about literally responding to emergencies in some cases, right? So how do you measure success? We actually have something in the product. I should have shown you this. I didn't. Oh man, what a mess. Maybe later. Yeah, afterwards. We call it the DFR outcomes dashboard. So we measure a couple of critical things. We measure response. So did the drone get there first? That's like a big important thing. We measure what the response time was. We talk about this a little bit off camera. And if you go to any city council meeting, any mayor, any police chief, they are really interested in what is my response time.

11:04There's a lot of budget and funding and a lot of contributing factors to this, but it's as a community, as Americans or as anybody, where we expect public safety to show up when we call and how fast you show up matters. We also track if the drone provided useful information. So like, did it actually help the officers on scene make a better choice? And then last, we actually track, this is a fun one for maybe B-Schoolers, we track how often an officer was not dispatched because the drone got there so fast, realized that there was nothing going on, it was a nuisance call, and it was a waste of public city resources to send a cop and so they don't go.

11:44And the best part about all of this is the customers report this themselves in the product and then their leaders get to see all these metrics.

11:52Carlos: And this is the main segment for your business? This is the largest part of our business. The other three segments are, yeah. The other two, I would group three and four kind of into one. So two is defense. Yep. What are some of the main use cases there? Defense is really interesting. If you think about a military unit, it's like a mini city. So they have a base. It has a border or maybe even a mini country. It's got a border. There's houses where people live. There's a grocery store. There's police. There's also people going to work every day. There's commuting. There's traffic. There's training areas.

12:27So that's like on base. And that's probably not what most people think of when we were talking about the military application, but it's an important thing. And it's something that we definitely serve our customers. And that's mostly the security use case. On the ISR side, tactical ISR is what you imagine. I think all of us are reading the news today and we know what's happening in Ukraine. And it's a very unfortunate, tragic situation. But essentially, the goal is to have the best information possible to keep yourself safe and to complete whatever objectives that you are assigned. So, I mean, I'm happy to get personal.

13:01Like, I'll think about my own deployment. So, you know, my job, I wasn't anything fancy. I was a truck driver. I was a truck driving officer. So I had 20 trucks, folks that worked for me. We would bring supplies back and forth in Iraq. And so we might be driving from base A to base B to bring them, you know, medical supplies. I'll give you one example as a mission. And there might be something in the road. And, you know, we would stop and we kind of look through our binoculars and we'd like look around and we just maybe we get out of our truck and kind of like go poke at it with a stick. You know, like that's OK.

13:32Well, instead of doing that, just send a drone, find out exactly what it is. And then you get to make a decision. In Ukraine, they're doing a lot more, unfortunately, like really large scale battle. And so today, a Ukrainian soldier will not move on the battlefield without air support from a drone. And they're doing really sophisticated stuff, unfortunately born out of a need. So they will recon the objective. They're using drones to see where the opposing forces are. They're using drones, what's called in a kinetic way. So that's like to deliver battlefield lethal effects. We don't do that. We're camera drones only.

14:08But our job is to help those soldiers identify essentially where the people who are going to hurt them are and then use that to make decisions on the battlefield.

14:17Carlos: And similarly to the first segment that you mentioned, the traditional alternative maybe was no alternative, right? Air support is something that was extremely expensive. Air support is really expensive. Like a Predator drone, we've probably all heard that in the news, is millions of dollars. You know, special forces or units that are in really hot areas might get some air support, but you just can't build or buy enough of them. And so our belief is that, you know, thousands of camera drones on the battlefield giving commanders much better information to make better decisions is a way to democratize the technology down from, my battalion was about 750 people.

14:51We never had air support. And we have units today who at the squad level, which is 10 people, they have a drone.

14:58Carlos: And again, like thinking about success metrics, I can, first of all, think about life saved here based on the example that you provided. Sure. Is there any other way to measure outcome success? You know, it's really interesting. I could rattle off a bunch of flight metrics. We're about to cross our 5 million flights, which is really exciting. All those metrics that we give, none of it includes our tactical ISR users because all of that is completely disconnected, offline, air-gapped. So we actually don't get as much feedback. There's different kinds of feedback methods with national security customers.

15:29But yes, live saved is the metric. or sometimes grim, but this is, you know, when we ask our soldiers to do something, sometimes it can be that like they accomplished their mission, which might be to go move to a piece of terrain or capture something or go fight.

15:44Carlos: And then moving us to the next bucket or segment, right? I was just at the World Cup final. I was born in Spain. So I have to say. Amazing. Congratulations. It's so cool. Thank you. We were rooting for you. I saw one of your drones there. Yeah. Right. So maybe that's a good way to exemplify how you guys are covering also the other type of commercial applications? Yeah. So we talked about this a little bit. Public safety agencies are using the drones there, but they're doing what we would call a physical security use case. And so a stadium is a great example. There's a big stadium. There's tens of thousands of people.

16:18You want to protect them. You want to make sure everyone's safe. You want to do perimeter sweeps. You want to know what's going on. You want to be able to get eyes on a hotspot. So the two ways that we think about security applications is kind of dual use. So one is autonomous patrol. We essentially tell the drone it's all happening autonomously to just fly around. And the way it can happen autonomously is we're also running AI communicating back to models to essentially look through the images that we're

16:46Carlos: getting and either onboard the drone with certain models or off board with other models, like identify if there are things that the operator decide, or excuse me, the customer decided that they cared about. And so that's kind of the patrol mode. Today, that's happening with guards who are walking around or people driving cars like we talked about, which isn't the greatest job. Or response. I can't mention the customer's name, but we're working with venue kind of companies as well, but there's also data center companies, also logistics companies. But you can imagine they have an alarm. And today, a guard has to physically go and go to where that alarm is or try and see it in the camera and clear the alarm.

17:26We pulled these stats. We were talking with some customers recently. About somewhere between 90 % and 95 % of all alarms in physical security in the United States are false alarms. So it's just a huge waste of time. Like, why wouldn't you just send a drone? It'll probably get there in 50 seconds or less, depending on where it is on its route. And then you immediately know, like, do you need to do something or do you not? And now that we're building indoor and outdoor drones, you know, we're really excited for this application to start to scale.

17:53Carlos: I like that. And I think that example applies to all the segments in terms of confirming false positives. Yeah. And like not only accomplishing a certain mission, but like knowing when there's no mission. Totally. You don't have to do anything. And like, why not send the marginal cost of the 3 % of battery is like nothing. Yeah. So it's like, it's not quite software, zero marginal cost, but it's almost nothing. It's a couple of electrons. Why wouldn't you do that? Let's actually talk about the software because you showed me some of the behind the scenes. And it seems like one human can handle multiple flying worlds at the same time.

18:24Yeah, which is very exciting. I want to talk about two things here. So one, just to nerd out on the tech. So today we have customers who are flying one pilot, four drones. We did two to one. It depends on the use case, like depends on how much attention you need to pay to every drone. But you can fly, you can autonomously command the drone to go do things. One of the things that makes Skydio great is our autonomy. And what we mean by that is a combination of deep learned models that are on the drone that are looking. We have cameras that look in all directions. And so we are both building a three-dimensional model of the world to identify and digitally create objects that we can go around.

19:01And we're identifying things to like know where we are and make smart decisions like the sky.

19:06Carlos: So like a Minecraft. It's very, yeah. So the 3D model is very Minecraft. You just need to get enough to know that you're, and especially if you're flying 45 miles an hour, like you're not going to resolve fine details, but you just need to make sure you don't bump into things. But one example of like where the deep learn models is being used on the system is we might be, we have customers who will fly into a parking garage to try and clear an incident. And the drone can automatically just find its way out. And when it knows that it sees the sky, it can be like, okay, I see this guy, I can go do what I need to do.

19:36Carlos: So that kind of stuff. The reason why autonomy is important is the only way that you can truly operate multi-drone is you have to trust that the system can handle itself, handle itself also when there are off nominal situations. Like let's say the city loses power. You need to trust that all those drones can fly home by themselves, get them safely, land by themselves. And so 1 to 4 is an amazing application. That's just the beginning. This is going to continue to scale. And one of the things that I, the second thing I wanted to talk about is for a long time in the drone industry, there was somewhat of an adversarial relationship with the FAA.

20:11People, like you'd go into the forums and people would be complaining.

20:14Carlos: So, what is FAA? This is the Federal Aviation Administration. Thank you. And we have taken just a completely different approach. We believe it is our duty and our mission as a vendor and as a manufacturer to earn the trust of our regulators. And so the way that we do that is we built autonomy. We've hired the best folks in aviation safety in the world and regulation and understand it deeply. And we work with the FAA every day to make sure that we are earning their trust. And in doing so, we have been given all of these groundbreaking waivers before everybody else because the tech can support what we want to do and they believe in us and we prove it to them.

20:54I guess they didn't believe right away. And we're also working collaboratively with them. So you said that one to four is the beginning?

21:02Carlos: when can we expect fully autonomous fleets of robots flying? I don't want to give away too many things about our user conference, but people should tune in. It's in September. It's not part of the question. It is very near term. I think in the industry, people imagined a world where we would talk about this and talk about it and nothing happened and nothing happened. And the scale that we are reaching in autonomous flight that's happening every day for the majority of all of our flights, even if it's a response like it's most of the time flying their autonomous being flying back there's some time when they're on this on the objective but we've now earned the trust that this system can do what it needs to and so like this is this is not 10 years out this is not a that's not what this is anymore it's happening now and obviously as a product that i'm sure you're looking for patterns and ways to group certain use cases but at the same time you guys are covering a lot of ground.

21:57Yeah.

21:57Carlos: Quite literally, right? So how are you thinking about funding the operations and making sure that you can continue building across the entire stack? So it's interesting you ask that. Like we've been working with these customer bases for quite a long time, but I would say that, you know, for the first, the company was founded in 2014. So for the first half of the company's life, there was just a lot of sort of R &D and innovation teams. And we had this novel thing. We had autonomy and we had this obstacle avoidance and people could see where this was going, but we weren't delivering value. And as a company, we then really got close to our customers and started to deliver real value.

22:37And I think what changed for us is X10, which is the workhorse quadcopter platform, the dock, and the remote operations software. When those things came together, it fundamentally changed the paradigm. And so lots of folks wanted to work with us and DFR really took off. And from an investment perspective, the way that we think about it is we actually prioritize that pretty ruthlessly when we were on that path. It was the number one company goal that we were going to make DFR successful. And we didn't invest as heavily in some of the other areas for a period of time. But once DFR took off the way it has, and we've been so pleased with the results that our customers are having and we get to benefit from them, we say internally, we've now earned the right to go invest in all these other areas.

23:21We've always had a large military business and we're extremely proud of that. It's a little bit more lumpy and some of the other businesses are less lumpy.

23:29Carlos: This kind of reminds me of what we're seeing with a lot of the LLMs, with all these labs, right? They were in the dark for many years, recent money, no product. Suddenly, product, product market feed, they're growing. At the same time, the CapEx investment is massive. They're growing revenue-wise. They're not profitable at all, right? So as I think about hardware investments, I don't know if data centers is an area of investment, but other areas, like how are you thinking about finding the operations and also making sure that you are, I think you had a stat that was like a certain level of commitment in terms of funding.

24:04Our CEO announced this stuff a few weeks or months ago. So Skydio is committed to 3.5 billion of total investment over the course of the next five years. And there's like really two flavors of this. Like this is all of the work that Skydio is going to do, all of the drones that we're going to produce. And there's a lot of investment in essentially the US and allied supply chain. So in order to build the tens of thousands, hundreds of thousands of drones that we're going to be shipping, we want really secure supply chains and we want a lot more investment in the U.S. and our allied nations. And so there's been a lot of work.

24:37And that's part of our commitment is U.S. manufacturing backed by, you know, allied and partnered supply chains. And so a lot of the investment is going to go there, plus all of the other activities that we're going to be doing.

24:49Carlos: So that's a pretty big number. Yeah. With a B. Yeah. Yeah. So as you look at the competitive landscape or the marketing in general, what is your strategic positioning and that gives you the confidence to know that you are going to be able to raise those funds so you can actually deploy them? Well, the great thing is we're not taking debt. We're not trying to go raise a big round to do that. That's just the revenue of our business driving how we're going to go make those investments, which is really exciting. I think our last round, we only raised like$100 million because that's not what we're trying to do.

25:21We're not funding this via debt. We could decide to continue to do that if we wanted to go after even more markets, and we might decide to do that. But that investment particular is just revenue funded, which is really exciting. In terms of the market landscape, yeah, I think there's a couple of things that make us special. We are, and I don't put myself in this category, but the company is a world-class engineering company. I am so impressed with the talent that we have here, and we have such a high bar for talent and a culture for how to go get the

25:49Carlos: best engineering talents. Where do they come from? All over. Clearly on the autonomy side, you can imagine labs, PhD programs, autonomous driving, probably some of the more traditional things. But on a software engineering side, we have folks from all different walks of life. So you're literally competing head to head with Google, the labs in terms of talent. Yeah. And then on the Mechies, the mechanical engineers, electrical engineers are competing with... I mean, we're in the Bay Area to get the best talent, but we're definitely competing for it. So in terms of them, you mentioned before the team, like how you are actually designing it.

26:26Carlos: You have people who are owning different segments, not quite GMs in terms of owning the number, right? So give me more details, because I think this is quite unique, right? Like most of the product teams that I've interviewed, they have software. So maybe their taxonomy is slightly different. I just want to learn more about it. Yeah. I think to zoom out a little bit, this org structure comes from a philosophy. So our philosophy at Skydio is customer first. And I was listening to a bunch of the folks that you have interviewed. I think like one thing that is true from all successful companies is that they are customer obsessed and they're customer first.

27:02It might sound different and how they do that might be different by company, but it's very clear that that is the most important thing. Like, are you delivering value for your customers? And we have a very mission-oriented culture here in terms of solving real problems for customers. So that's the start. And then when we look at the team, we know that we have deep, deep investments that we are making from an engineering side to do the things that we do. So just to give you like maybe one example of it, X10 is an amazing platform, but our customers in the military and public safety and security and in inspection, they need to fly at night.

27:38And they love the fact that we have 360 degree obstacle avoidance and autonomy, but it only worked in the day. And that's 12 hours of robots not doing what they need to do. And so we invented Night Sense, which is this active illumination, invisible to the naked eye in IR, so that you can do all of the autonomy functions at night. That is deep investment across a wide range of teams, hardware, software, autonomy. And so we did not want to disrupt the engineering and invention engine of Skydio. So you have this platform at the bottom. And then at the same time, we are incredibly customer focused.

28:17And we believe that if you don't really know what the customer needs, then you're not doing a good job as product, as engineering leaders. I say this to everybody who joins my team at Skydio. Our team's not very big. We're like 20 product managers across that whole stack. So we run pretty lean. We like it that way. But when you join Skydio as a product manager, I call them product leaders. There is no travel budget. If you need signal, go get it. I don't care where it is. Like the thing about physical AI is we love data. I host something called product data day, where we all get together and we all look at stuff and we poke at each other's data, but you have to go see the robot in the real world and see how they're using it.

28:54Otherwise, How do you know? So you need that deep embedded with the customer focus from the product and engineering teams. The product leader is paired with an engineering leader and an engineering team and they go do what they need to do. And then we have what's called strikes. So a strike is you might have a deep investment. We have a feature called Pathfinder. So we take maps of the world, digital surface maps. We know where buildings and trees are. and we autonomously will plan a route for the drone that is the most efficient way to get there without bumping into anything. That was a deep feature that we built.

29:28We built it very expressly for DFR to minimize the time on site or the time to get to site. And so several core members of autonomy and embedded engineering went and we call it a strike. They embedded with the DFR team and they just worked with that team. They sit at the same place, they eat lunch together, they do everything for six months until it's done and they go back to platform.

29:49Carlos: That sounds like what other companies would say, forward deployed engineers. Almost. Yeah, yeah, yeah. Diverse somewhere, yeah. So it's engineering plus product as part of the same org with an R &D platform. Yeah, and design. I can't forget my design partners. What do you think about design here? What do you mean by that? So when we think about design, I think there is so much raw engineering that goes into like, how do we deliver something that works? It's a really hard problem. The product managers are spending a ton of their time thinking about like, is this the right thing that I need to prioritize?

30:30We have so much demand for what we're building that every day is like an agonizing choice between like, where do I invest here or here? And design is really important for us to make sure that what our customers are experiencing is intuitive to use, feel safe for them to use. If you think about like more and more autonomous actions from a flying robot, a customer needs to feel that this thing is going to do what they ask. And when it takes its own action, is doing so in a safe way and showing them how it's going to work. I don't know if you have any, if you've done any self-driving with a Tesla or whatever, but like, it's great that you know where it's going to drive.

31:13You need to have that intuitive user experience to make sure that they trust the robot.

31:19Carlos: So let me use the example of Tesla here because in a way, I can see some similarities. They are building hardware plus software. The software can be upgraded very frequently. How long is your hardware cycle? So what we call a major program is about two years. We've done one, I think, R10, which is our indoor drone. we might have done in like 13 months or 18 months, maybe 18 months, but generally about two years. And we make big jumps. We also do revisions on that hardware to make reliability improvements, to make other product improvements for capabilities. And we have attachments and accessories and stuff like that.

32:00So that's shorter, but I'd say generally it's about two years. Well, thank you for the masterclass

32:04Carlos: on hardware with software. I don't know about that. In a flying robot context, This is definitely a very unique episode. We've never had anyone who's been doing something like this before, so it's been a pleasure. Yeah, absolutely. Thank you so much for letting me on the show. It's been great.