Practical Logistics and Future Vision

You are now listening to the EU Startu ps Podcast . Hello and welcome to another episode of the EU Startups Podcast. As always, I'm your host, Dave Sunning Garcia, the news editor here at EU Startups. Today I chatted with Yane. He's the CEO of the airship developer, Kellu. Now, Kellu is out of Finland and has been around since 2018. They recently came to our attention due to their 15 million funding round just raised a couple months ago. Now, their airships are actually quite interesting because to those of us that are unin itiated into the history and technology, they might seem dangerous or even outdated. But Yanai was actually quite convincing in his arguments towards their uh aerial data use and even their defense uses when it comes to the current geopolitical tensions between the Ukraine and Russia. Now he was really able to talk about not only the defense applications, but also the environmental impact and even the practical information of how an airship works, how uh it's lifted up into the stratosphere, and the uh everyday running and maintaining of such technology and hardware. So I hope you're looking forward to that. Hello, Yanne, and welcome to the U Starbs podcast. It's a pleasure to have you on. Thank you very much. Yes, no, uh, I'm looking forward to our chat today. Uh it's it's not every day we talk to uh an autonomous air fleet airship uh developer. So it's really quite an interesting topic and I'm looking forward to it. And so before we get into the questions, I think it's probably best to let the audience know exactly what it is that I'm talking about and uh who you are . Sure. So my name is Janna Hietala. I'm the CEO at at Gillu. We operate persistent aerial autonomy . We've actually had a lot of discussion that should we call these things an airships? Should we call them LTA, lighter than air drones? Or what should we call them? It will give people, you know, different um images obviously. Well before they see actually what we what we do. But I think airships is a good description and we want to respect also the history when it comes to aviation and and use of airships as a platform. And still to this day it is a very much, you know, one of the most efficient ways to counteract gravity. We don't have anti gravity devices, but we're able to use lifting gas to make aviation Well that's that's an interesting place to start, I guess, and the history of it all because I hear airship and I imagine a big hydrogen bomb uh that uh goes up in flames. Uh you know, it it has this you talk about the history and it has to come up also that, you know, that there must be people who must kind of criticize the idea of it saying, D didimn,'t we discard this a hundred years ago due to safety worries ? Well, let's get a couple of facts straight in in that, especially. So airships were first, especially when when we think about the defense concept, and we are we're now operating both civilian side and defense side. So we need to go back to French Revolutionary War in the late 1700s where they were first used. They actually in a lot of ways sort of uh pred ate other t other modes of aviation, if you will. When it comes to Hindenburg and the perception of you know explosion, uh that wasn't even, you know, the the worst airship accident. And the worst ac airship accidents have actually happened to uh Akron and Macon, which were US Navy airships later than Hindenburg, and they were using helium as a lifting gas. So any aviation is dangerous and obviously Hindenburg was its tragic tragedy uh uh uh when it happened, but it was also you know one of the early stage when the TV cameras started to be uh available and it was very well documented and filmed other than other other other accidents. But when it comes to uh airships and safety, um it wasn't actually the the reason why airships were defunct or or not being used. It was the second world war where we needed the aluminium that was used in these massive structures that were longer than football field. And they it took a lot of time to build them, that they were defunct at, least in the large Zeppelin case. But Navy continued to use airships until 2017. So the last US Navy airship was decommissioned in twenty seventeen, we were founded in twenty eighteen. But when we take that into what we are doing, we don't have let's say other things common with the Zeppelins or Hindenburg, if you will, than let's say the concept of airships of using lifting gas to counteract gravity. One of the hard things with airships, and there are multiple really cool airship companies and startups, even today, but there are very few that are operational, very few that are commercially viable. So you have LTA Research, which is uh founded by Sir Gabe Bryn, if that name rings the bell, that's the Google's founder, with you know a lot of um capital to invest. They built now actually the modern version of the Zeppelin. So this is a rigid construction . And we've had Goodyear blimps that people kind of recognize. These are made by Zeppelin uh uh Luft tech Zeppelin NT, uh Zeppelin Luft Technique. Uh and they they still manufacture these and these these have a gondola and semi-rig id structure, and they're able to carry you know people for sightseeing and the different sports events. And then we have a smaller blimps, which are non-rigid construction, where you basically use the overpressure inside the gas envelope to provide rigidity to the platform. Because you cannot, you know, if you have a very um non-rigid structure and you provide propulsion, it doesn't tend to work. But nobody's been able to make airships mass manufacturable. So there's not a line of, you know, like Forbes T-motor line. That's never never uh never happened uh in in history where you could actually manufacture hundreds of thousands of airships. And that's mainly due to the size. Research is now being built that the Pathfinder one, it's the old airship docks in um the the um in San Jose uh at the Moffat Airfield. And these are literally where they built the old uh Navy airships, and these these are huge, hundreds of meters of long buildings, because the Pathfinder one is 130 meters long and the next version will be even larger. So the physics actually dictate your airships generally tend to be very large. They're more efficient in terms of the payload or the mass they can carry through that. And that tends to happen. So you have kind of this upwards mass mass circle that happens. We were able to make, and this was the hypothesis when why we started Kellloop, that could we benefit from that almost persistent flight capability or perpetual flight capability because the airships were able to fly around the world when we had the had the Zephylins, and they did. Could we use that to make an unmanned platform much smaller and by making that smaller we could make it more economically viable and mass manufacturable? But the world hadn't changed as much at now changed due to the Russia's attack to Ukraine, the uptake of autonomy into the security, and the fact that we do do need to have this mass manufacturing capability in order to have scalability. I would say that the only kind of recent history analogy with the similar transformation that what we are now doing has been done by SpaceX with Starlink . And the concept of having 10,000 satellites in the low Earth orbit providing communications is not really something that they taught as a first, but they were the first one to make it viable . So Kelly is in a you know starting from the hundred-year-old concept, actually looking at that, why it makes sense from the first principled perspective, being efficient counteracting gravity, understanding the bottlenecks, which we're able to solve to enable that to be mass manufacturable, and using modern technology of autonomy AI, hydrogen fuel cells have advanced massively, and all of these different components required to operate autonomy have miniaturized, become more available. But there were still hundreds of hard problems we had to sol ve. That's why we did it in a, you know, our version of a Skunkworks lab, which we'll talk about a little bit more. But it just required a lot of persistence to be able to create a new type of aerial vehicle that's regulatory approved, that's able to operate in high wind conditions, it's able to operate in freezing and arctic conditions that we have in Finland, and as well in a GNSS denied electronic warfare jammed condition s . Okay. Well, that that was quite the history lesson there. That was very expansive. We covered the French Revolution all the way to SpaceX. It could be its own TED Talk. So uh so you you know, you you have all of this information here and clear you're passionate about it. So I wanna hear a little bit about uh you personally, kind of what drew you into this world of airships, aerial data, because uh you know from from what I did which was a quick scan of uh of your LinkedIn and all these things, right? It kinda I couldn't really see a lot of uh a lot of a connection to that before you started with uh Kelly . Well I I I think it's it's not that sort of uh surprising why I am I am here uh in in this company. My background is in tech. I grew up in the early stage of you know, we didn't have internet in the in the early nineties, we had bulletin boards. But that was the precursor of internet to be able to connect, uh, to be able to have that sort of um the grow up in that kind of the phase of early early stage internet. Um and that enabled in a way uh look at maybe maybe to look at the world a little bit differently through that digital lens, um, how to do different things differently, how to access information, how to solve problems in that digital domain. And I always tell the story that I really haven't had a real job. I've just been entrepreneur all my life, uh, since you know very young age in that sense. And we were quite successful with the previous business, which was to build AI-enabled enterprise software. We bootstrapped that. I I lived in in London for a couple of years with my family, uh doing that and then exiting that business in 2021 um and retiring for a bit just to prove that you know I can also stop being entrepreneur, which lasted about a year. But I think the primary motivation at that stage, you know, to jump back, there was COVID, uh the the war started in Europe by Russia attacking Ukraine. Finland in general, we we are I I I'm from here, from eastern Finland. This is where we are based. We're about 60 kilometers from the Russian border. And Finland has had a long history in defence as well . And that was kind of the combination of you know, we need to do something about the situation uh in Europe's defense. We need to contribute to technology, and where Kellu had entered at that stage, so I I joined a few years later when it when it was originally founded, they had enabled essentially the technology to work. So the early stage feasibility, the hard grind of well, can we actually do it? They had evidence that well we can actually do it. And you know, I'm I'm very much interested in solving hard technology problems. And these days, um with Gelu , we've already invested and done more work into the software side than it is to the hardware side. A lot of the autonomy is about software and AI, which is my comfort zone. But I'm also excited about learning new new areas of technology when it comes to how C2 systems work, how radiolinks work, how different payloads and sensors work, and how do we mass manufacture at scale. But you know, we have a fantastic team. It's not me doing this, but we have a fantastic team that enables us to do what we do. Whoa, you're a great mouthpiece for the company. Clearly, you know a lot and you're passionate. So it's good to have you on. Uh and now that you've brought up this kind of uh defense angle, I I guess it's a good time to jump into that topic because it's an interesting side of what the company does. Uh I've found that since the Russian invasion of Ukraine, there's been such a uh a uh push for dual use technologies, right? Uh so technologies that have a devent defense element , but also have another element to them, whether it's data, analytics, uh what have you. So uh I I kind of want to hear about how you personally think about building a company and building an offering that has a defense tech angle to it, but also has a uh you know, a from your press materials I've read the the idea is to make a digital twin of the earth, you know, monitor for uh forest fires, that kind of thing. So how is that balance uh developing for both uses ? No, that's a that's a very good question. And we started purely from the civilian side. So that was about making making the autonomy, aerial autonomy scalable. So can we have airships that cover Finland, Europe and provide valuable data about the physical world for us to be able to solve hard problems? Hard problems could be environment, natural disasters, uh, you know, degrading critical infrastructure uh and and generally like the combination of early stage AI when it you know started from chat GPT, uh GPT two models to GPT three, it wasn't that hard to see at that point that, well there,'s gonna be a physical world dimension of the AI. It's it's now literally now emerging. And the biggest difficulty there is that, well, the infrastructure to collect the physical world data, satellites are not sufficient. They they are not able to provide granular level data on a centimeter or millimeter level about the physical world. So can we approach it from another dimension? But that's still the long term plan that we have, and that's the biggest impact that our technology is able to do in the you know sustaining habitation in earth not necessarily you know other companies will build uh humanity into multiplanetary species which is which is fantastic and inspiring, we are all about you know making uh life sustainable here, using that autonomous technology. But the priorities change and priorities, especially in Europe with defense, you know, being very underprepared , under resourced for tens of years in Europe, NATO, the security picture changed rapidly when Russia attacked Ukraine. And in a lot of ways, we are still playing catch -up . And the reason why we need dual use technology is that the traditional defense works in a 10-year cycle. When it was the, you know, first world War,ond Secld Wor War, it is not, you know, unknown that technology gets developed very fast during warfare. Really uh uh in in that scale. Um and now now it even 's it's in Ukraine um but there we feel very very you know similar as Finland felt in the second world war defending against Russia with underresource uh uh on uh under being under resourced um and that's that's what ukraine is really very much doing they're using everything they can to innovate to discover use of drones uh that people recognize but the the sense of urgency in other European countries hasn't really happened and that's what's being done, where we need to move from that ten year development cycle into the weeks and the months that Ukraine does in order to have a credible deterrence, the defense starts by having ability to defend , having a credible ability to defend, and that means when you have a credible ability to defend, that means that uh it's not worth it to attack. So that deterrence is missing and that deterrence is missing because the innovation in the drones and autonomy has happened when we look at Ukraine being able to use few hundred Euros strove to destroy a million million euro tank. That's an example of it. Recently we've seen in Iran where US attacked Iran being able to destroy billions worth of high value radars, surveillance planes . I don't think anybody you know reasonable made the calculation that that would happen. But the world has changed through the autonomy and mass manufacturing in a way that now actually these low-cost uh drones are able to take out at long-range uh very expensive capabilities that are not just military capabilities. These are critical infrastructure part, our communication, our power lines, our sort of power plants and others that would be seeing happening in Ukraine. And that kind of means that it's no longer a defense responsibility. It is the responsibility of the industry, society, what we Finland we not Finland has been done very well to actually create solutions that rebuild back that credibility terms. So we play a part in that. So so we've gone to uh NATO's Diane Accelerator, so the defence innovation accelerator of the NATO and that was designed especially for sourcing out of you know thousands of applied companies and solutions, the core innovations which NATO can use to solve these problems which the traditional defense cannot solve in time. And we kind of began one of the highlight companies of that going through phase one, where there was uh 2 600 uh 2300 companies that applied, 74 were selected. From those, the the 15 companies were down selected into phase two, and we became the first one to receive a rapid adoption service contract, which is a framework contract with 32 NATO allies, for that kind of direct procurement, they can solve source from us our technology capability. Because they identified that our capability fills the critical gap between satellites and you know ground based sensors and drones. Okay. And and I mean I I have to ask because we you know we we should also reference of course your fifteen million uh euro funding round that you raised back in April, which you know is indicative of obvious interest. But it does make me wonder about that if there wasn't this kind of geopolitical tension that you're talking about now, do you think Kelly's offering as just a technology would still be seen as valuable and as worthy of this kind of investment and uh research, or has it been pushed forward by in large part by its defense use ? Well, I think getting let's say a reflection of investment landscape is helpful there. Looking at from the government perspective, you know, the problems that were with with the environment, you know, EU regulation regarding restoration of environment that that had a multi-billion impact into Finland, for example. Understanding how to maintain our biodiversity in the forest, our critical infrastructure was still still a problem at that stage. So the investments and investors were looking at uh impact investing, environmental, you know, CO2, uh uh carbon capture, uh carbon markets, a lot of this capital is now shifted into defense. It's shifted because of the sense of urgency uh in a lot of ways that well now we need to have the technologies capable of actually defending and building the deterrence but also the customers have shifted the governments have shifted the investments that they're making that's what we see in Finland. You know Finland has had an economy that hasn't grown as as strong as other Nordics in the last 20 years. Uh so Finland is saving in everything else right now, but they're investing in the defense because we have to. So this is a reflection in a way that Kelly had a very high growth um and high high value opportunity to push into environmental use case and demonstrate that large large scale, but the the priorities have shifted. So now we need to deal with defence in Europe and we have a unique technology that thanks because of that civilian use case and that environmental you know uh uh problem sets that we were solving solving earlier, now we can uniquely uh help solve these defence use cases. But for us, this is an intermediary step going back into that infrastructure play using the autonomy in a larger scale. So in a lot of ways, this is accelerating the pace we are going to the market and how the technology is maturing. You can take the same analogy to Starlink, where, you know, I think in a lot of ways where Elon Musk started is by, you know, first looking from first principles perspective that well, we need to decrease the cost of kilogram or ton to orbit by having a launch capacity. And he was successful building that reusable launch capacity with the Falcon rockets. Now he needs to have a payload in order to accelerate the process. So he yeah, let's launch satellites. What could we do? Let's let's launch 10,000 starlings. And now that's a hugely profitable business because they basically uh revolu revolutioniz ed communications globally, how that's being done. So that's sort of similar analogy and Starlings are used also in defense and resilient communication that played a critical role in Ukraine's ability to defend itself. But it wasn't started originally defense purely in mind. It's it's sort of a journey where you have your main quests and then you have your side quests, but the long-term plan, like for them, uh is kind of the big picture. Uh for them it's the making humanity multiplanetary species. For us it's sustaining, you know, habitation here on earth. Well you you and you you've mentioned this uh kind of uh environment al aspect to what you do, which I think is important to get into as well. Uh, because from what I've read, there's it's it's uh there's no emissions involved, there's uh you know it's near silent. Yeah, how important is this environmental impact to what you do, especially when you're comparing it to uh drones, uh helicopters, planes, whatever that could be kind of doing similar work that you are, how important is it to the company to be making sure there's the environmental aspect to your offer well these are all features of the technology and you know if we would have thousands of drones which is you know there's companies with that vision, they're pretty annoying. They're very loud and they they you know look like a flying spiders in your backyard and it's you know it's it's not the people get annoyed. We get a very positive reception because it's it's non-threatening. It kind of you know floats in the air, literally that's what Kello means. Um and it doesn't make a sound. But you can in in in the same sentence, you can use these properties to describe our capabilities in defense. Being a very you know hard to hard to detect target in in a many ways because you're silent and you don't have emissions, being able to use hydrogen uh and hydrogen fuel cells, you can see actually large-scale drone operators that are using drones in Ukraine moving into hydrogen fuel cell technology for this exact reason. When you don't have a thermal fingerprint, because your gas oline engine is burning fuel making a very loud sound that is easy to detect your target in in multiple ways. So these these sort of properties that were important for us to be scalable are And what's uh what's the plan c ause uh I I'm not sure how many uh of the airships are now are operation, which if you have that number, that would be uh be great. But how how large do you plan to take this? Is there a kind of block at some point we'd say this is too much, or can this be scaled globally as large as it needs to be Yeah it's it's not the huge number. So so now now we are manufacturing Gen 1 airships and we are operating that just to give you an amount of uh you know indication of scale, uh if we would operate Belgium we would need five airships to cover whole Belgium. Uh we are now operating most likely the largest unmanned airspace in Europe. We cover about hundred thousand square kilometers of one third of Finland and then a large chunk of Sweden right now where we can operate an RSpace management is. So for Europe, we're launching the Gen 2 path on this year. So it's the same form factor, but like you know, uh NVIDIA going from A100 chips into B200 and B300 chips, they become better and powerful and you know enable different things. That's what's happening with the Gen 1 and Gen 2 platforms, what they're doing. So for Europe to cover whole Europe for that persistent aerial monitor ing use case to collection of the data and providing provision of the security, we need about five hundred airships. When we scale that to North America, South America, you know, take Australia, some of the friendly Asian countri es that we want to work with, we need about three to three thousand five hundred airships. If you put that in the historical context of airships, in fact, we operate today the largest airship fleet in the world, but there are not that many airships in the world as of now. But this would be ridiculous concept when you take that into the airship, uh uh airship world. But if you compare this amount of Starlings, this is less complex than a low Earth orbit satellite. It's less complex to operate and it's less complex to get into space. So it's not that unreasonable in that sense. Okay. Well, I imagine anybody who's listening to this is probably kind of screaming at themselves saying, When is he gonna ask about how it works practically? Like how it looks because I I imagine a lot of people kinda wanna hear about the idea of w what does this look like? Do do you have a a big warehouse and and you just kind of let them go up like a balloon? Uh how often are they maintained? You know, what what does a normal day kind of look like from launch to land? Sure. I I mean it's it's very very practical and it's uh in in some ways that the ground logistics is low heck and we wanted to keep it at that way as possible. The one one of the airships, its home is a sea container, so forty feet standard C container. And now you can imagine like a metal box uh having one person to be able to take it in and take it out. We do fly checks, we follow all the regulation that's required for unmanned operations. The airship literally is, you know, taken the start kind of the ship up surface, that give it a gentle notch and it will fly straight up because it doesn't understand gravity. It can ignore gravity in that sense. And then basically mission command and autonomy takes over. So we can operate a fleet of airships from single location from a different country. And that was always built in into the concept that well, most drones, even today, will use radio links. They will use radio controls and radio links, and that's limiting their fact fact that how far they're able to go from the operator. If you take Radio Horizon, it's everything between from ten kilometers, maybe the best military data links are able to get to uh eighty kilometers or hundred kilometers. But you would need to have a very large mass to have a that kind of radio line of sight. Um plus this was always built in because the flight time is so long, the endurance and ability to move so far is is is there and we want to decrease the amount of people hours per flight hour. So in every day and every week we would operate multiple airships from multiple bases for operations for that large game with the intention that the primary mode is that it's always in the air. So you know it's a it's a little bit of you know ridicule of maritime ships, but you know, they figured out that if ship is in the bottom of the ocean, something has gone wrong. And if you take the analogy to the airships, it should be the same way with this this mode of aviation. To make it really persistent, um, it needs to be a h large faction factor of the time means it needs to be in the air in operations. So that's what we are driving towards. So we operate as much as we can and as as far as we can. And that is the hard truth of aviation in general, that when you're creating a new aircraft and aircraft type. That's why it takes so long and t takes so much money. Is that you need to need to operate it a lot in order to discover all the edge cases, all the problems that when do you need the maintenance? There's no field guide for airships or unmanned airships that will you need to maintain it every thousand hours. You need to figure it out how often you need to do it. You need to look at the wear and tear. And the only way to do that is just to put them into the air all the time for for specific missions to really hard environmental conditions, including freezing conditions here in Finland. And now also in this defense domain, they need to be under uh electronic warfare and jamming and spoofing, which we have for free here in Finland. So Russia they're trying to deter Ukrainian strikes into Russian ports. That basically bleeds into Finland. It's disturbing aviation. It's d disturbing everything with GNSS receiver that you rely on to get your position. Your your you know uh DoorDash or whatever uh uh you know foods food delivery company you use they, rely on G PS to track that where they need to go and where they're going. So in here in Eastern Finland, we have a perfect test environment with really tough conditions for this kind of thing, including the GNSS jamming and spoofing. And that's enabled us to go fast. Because we can test every day. We can fly every day in these harsh conditions. And now if we operate in in let's say Mediterranean, that's actually quite easy. Okay. Well we're we're we're nearing the 30 minute mark, which is usually as long as I try to make these, just because if not people will start getting angsty. So uh and uh yeah, I really could keep talking about this. I think a not a lot of natural questions come up because it's not something a lot of people know anything about. So I I I'm sure anybody listening to us also has 10 or 20 questions that they wish they could ask you. And and uh I I I guess I would encourage them to also kind of look into the company because it is interesting. But with uh with that in mind, I think what I want to wrap on is the idea that uh the company started out of a shit, which is a great story really for any company. Uh and and now we're looking at fifteen million raised. I want to hear and that was in if I have my dates right, that was found in twenty eighteen, the company. Yeah. So so we're talking about eight years ago. So where do you see the company eight years from now? Uh you know, do you see it continuing on raising Well, just out of the origin story in a way that you know great companies are found in garages quite often but you know the best ones are in sheds. But for for us, you know, if you're making bold claims to be able to revolutionize aerial monitoring, you need to have evidence. And for us that was really the skunk's work skunkwork story of kind of being in a in a place where nobody can see us. We can we can fail in peace, we can just experiment and try different things and in stealth mode in the early stage. But you know, we we appreciate that history and we have a very large uh photo of that early, early days in our factory right now. But looking looking sort of eight years ahead, uh, we've accumulated a lot of knowledge exper,iential knowledge on LTA autonomy, the lighter than air technology. And you know, we can go, we can scale the technology. There's a lot of work work in in in you know making that 3 500 air shapes operate every day, all day, and you know, providing that data and data logistics. So a lot of software problem sets. We are now already starting to invest into physical AI. So we launched as part of the round, we announced our AR lab s. And because we're able to collect unique data set about physical world, and we have already done that for many years, that is the unique challenge for many of the AR labs, that they don't have any differentiating data available. So we have something that nobody else literally has right now to be able to do foundational research into physical AI understanding and going from let's say let's let's predict pixels into understanding that world why these the actions and things are happening whether they are in regards to the environment whether they're regards to physical let's say critical infrastructure. And geospatial works in in many ways solving one problem at a time. So you need to create a treaty reconstruction that's metrically accurate and, then you measure certain things of the physical world, you sort of do an analysis based on that out of digital twin, and then you've solved one problem. But the physical world has millions of problems. And through AI, we're looking to solve all of these problems. Similarly, as you know, foundation language models are solving a lot of different problems. It's it's a source of very vast knowledge that you didn't have access previously. You had to do a lot of work specializ ation. Now that's enabling a lot of th different things at scale. So so when we when we think about thinking about eight years ahead, that is really the what we are focusing on is that we have built the infrastructure that provides that unique data set and we are custodian of that data, using it responsibly for solving really hard problems regarding physical world using AI. We are still looking to partner in that domain as well. You know, we already worked with with Nvidia, we're using their hardware as well. They're doing a fantastic job in in terms of building that infrastructure for AI. And now we basically produce petabyte scale data sets. And this is this is quite unique problem that we don't think that we're gonna be only one working on this. There's gonna be multiple different AI labs. You know, A AMI, the Jan Le Kuhn's company, uh was founded for physical AI. They raised I think one billion just for for seed funding. Uh but we are in the same sandbox, but we are solving it from the different side by building that infrastructure. We're gonna we're gonna connect in the middle at some Okay. Well, thank you very much for the chat, Yan. It's been a pleasure to have you. Uh as I said before, it's not something that I've really ever talked about with anybody. So uh I think it's a very interesting topic and uh here to you startups will definitely be keeping an eye on you. So thank you very much for joining.