EP05: OSINT, Space and the Evolution of Intelligence Capability | What’s New in New Defence Podcast

[MUSIC PLAYING] Greetings, everyone, and welcome to the Digital Defense Ecosystem Podcast, where we explore the global security landscape and examine finished companies who bring solutions to global security problems. My name is Valtteri Vvoortosova, and I'm the professor of practice for National Security and Security Policy at Tambra University. In this episode, we take a look at open source intelligence and its impact to national security. Joining me in the discussions are Vellipek Kivimaki from the Finnish Security and Intelligence Service and Iarko Andela, the CEO of Guba Space. Vellipek, welcome to the show. Pleasure to be here. Vellipek, you have a very interesting background. You've worked with Nokia, Microsoft, and the Finnish Defense Forces before joining the security and intelligence service, also known as Super. How would you comment on the utility of these experiences in your current job? So I would say it's useful to have that diverse background from different places, and especially having that private sector and public experience together, I think it's interesting. So it's obviously being on the government side of things are not necessarily as smooth always as possible. So it's good to bring that perspective with the private sector as well, which also helps us develop our ways of working on the public sector. So I definitely feel that's valuable. And across all of my career, then anyway, technology has always been there. So I am working in R&D in Nokia and Microsoft developing new products and then in defense forces of working on technology for sites, so figuring out how does that technology then transfer into new solutions in the defense sector and what the lies ahead. And we have very similar issues then on the current job as well in the security and intelligence service as well. So it's technology is always there. Yeah. How would you rate the experiences? I mean, is there something that the public sector kind of specially learned from the private sector or should the private sector look to the public sector in some cases as well? I guess there's different things. So in some ways, I guess the agility is something that we have on the private sector that I hope to bring on the public side as well. But then especially in the security space, I think it's important that we bring some of those learnings and special things that we have both in the military side as well as the civilian side that bring that knowledge into the defense ecosystem as well. When you think about companies doing collaboration with other companies or universities and whatnot. So they are certain security threats that might be associated with those. So it's good to also be able to bring that understanding into the private sector as well. And sometimes those are overlooked, to be honest, I think. And it's difficult because if you look, for example, look at universities and the funding is not always steady. So they have to look for alternative ways of funding and where does the money come from once or what's attached to it and things of this nature are important to understand. Right. So open source intelligence has really picked up with the war in Ukraine. When a lot of people have done, especially geospatial analysis on about the events in the battlefield, how do you see that development from your perspective? Well, my perspective into the conflict in Ukraine is goes to 2014 already. So I've been following the war since then from open sources. So it's going on 11 years now. And just if I go back those 10 years where we were, so we have some emerging capabilities in the commercial space sector, I would say that there was already some well-established player like Digital Global Max, later Max R. We have planet labs, for example, with their constellations coming online. But then as we've gone to this age now, it's not just the old electro-optical sensors that we have, but we have SAR, we have emerging hyper spectral capabilities, RF detection, and things of the nature. So just a range of sensors that's up there is quite different nowadays. And what's remarkable is that these are commercial capabilities. So theoretically, anybody could buy them. Of course, there's price tag attached to it, but still that these things that were ones just sort of the sole purview of government that they have the capability to launch satellites into space, put sensors up there. We now have a commercial spatial launch capability, commercial capability to develop these very exotic sensors. Now put a constellation of satellites up there, not just one or two or three, but hundreds. So it's remarkable this whole picture has changed, and it's also brought new capabilities into analyzing wars like Ukraine now, and especially since the beginning of 2022, what we saw that these commercial SAR providers were actually providing images already before the breakout of the war that's showing that, hey, the Russian troops are very close to the Ukrainian order. They're massing troops here. And these commercial providers were putting it out there, along with the intelligence that was released then before the war. So that was a fascinating change to see, that just amount and types of information that's available nowadays for open source work is remarkable. That is mind-blowing, even. When you consider these two trends, first of all, there's the technological evolution or even revolution. There's a lot more data in other words. And then there's the conflict in Ukraine when there's a lot more people practicing open source. Intelligence. So these two developments combined, would you say that the community has evolved or grown, has it become more professional, or how would you rate that? Well, it's definitely grown since 2014. So we have multiple different kinds of open source groups now working on this problem. Some in Finland as well, for example, a Blackbird group is one example here that is very closely following what's happening on the front lines. You based on social media, a patriez satellite imagery, etc. So it's definitely become a bigger field. So the tent has gotten bigger and osin so that we have more players under there. And I think one key thing that started to separate the different actors in that space is the professionalization as well. So we have some that are definitely getting more experience in this field and recognition in what they do as well. So I think that started to set these different groups and players apart in that space. Yeah. You mentioned the satellites and the different types of sensors that are up there. Often when we talk about the technological disruption that's going on, we talk about artificial intelligence and quantum. Do you see these technologies playing a bigger role in open source intelligence? I do think that AI is already starting to play a role. And if we look at large language models and computer vision systems, for example, there's been some ongoing development. I will say over the past five years already and different kinds of computer vision systems that could be exploited for open source intelligence work. For example, automatic analysis, video stream, very similar kind of a problem set. As if we think, for example, a drone footage, an automatic processing of drone footage, very similar thing with the open source footage as well. That similar kinds of things that could be done there. Some interesting experimental work on the open source side has been, for example, analyzing material coming out of Syria to detect different kinds of munitions in videos. So that's already there. And now with the large language models, then it's interesting that how do these AI systems then build to assist in the open source work? So whether it's finding sources, finding information online, like more agent kind of functionality that they're doing their own thing, trying to map what's out there. But also then even as we've done the collection, regardless of how we did it, then the analysis of what we have and sort of finding the needle in the haystack type of thing. So I think those could be incredibly valuable in the open source space. I guess quantum, we're still a little bit further away with that. But clearly it's going to have an impact on how we do AI, how we do different kind of modeling, optimization problems, et cetera. So I think definitely that will increasingly, as those technologies become mature, they will also assist us in sort of wrangling those masses of data that we have as well. Yeah. Well, one outcome of these new technologies is that you can produce fake data and well, deep fakes. And you can even make my image and video of myself and call my mother. And she wouldn't know. How do you see that problem in the open source intelligence world, but also in the intelligence world as well? How can we trust what we see anymore? It's becoming an increasing problem that there's been this whole public discussion about the pollution of the internet with this synthetic content that is being done. So whether it's AI-generated blog text or AI-generated videos, et cetera. If you look at some of the latest developments in video, like with the latest SORA models and sort of how convincing that video is that it's starting to fool people on a routine basis. So there are videos, generated videos, posted on Instagram and elsewhere that people are falling for it. So it's starting to become a problem. And especially if it's possible to weaponize these models in the way that they are generating to see full or harmful material that has the potential to impact then like real people in real events then we were sort of facing a major issue. And that's where provenance becomes interesting, interesting and important that how How can we be sure that this actually originated from a real device or a real sensor? And then the question becomes, is there some kind of water marking or how do we vouch for the authenticity of something that's coming especially for race to something important like criminal digital criminal evidence or war crimes or something of that nature? Then how that's a key issue that how can we be sure and how do we validate the authenticity of the material? How about blockchain technologies? Do they play a role in in in securing authenticity at all? I guess it's a broader problem set that it's it could be it kind of depends that where we start solving this. So my my perspective is that we should start solving at the sensor already and start start start from there and how do we ensure that at that point in time we're already putting putting something in there that helps us authenticate it or does it become later on in the in the chain for example if we're dealing with media houses that routinely deal with lots of lots of imagery are they the ones that are watermarked the images or the test ends of the blockchain or whatnot. I think it depends where we start solving this problem then what the specific technological solution will be. Yeah. I can only imagine since there's a whole new masses of new data you mentioned the internet is being polluted with synthetic data and and you're in a you're under a time constraint to produce produce results and analysis. How do you deal with that? Is that do you have these student technologies? Can they help you with these situations or are they only causing more problems? So I would say that it's part part of where the trade craft and sort of our value as an internal service for example comes in that we're we're able to establish we have quite a bit of information already about how things are. We understand what kind of deceitful information there there might be we're perhaps certain technical needs to do certain technical certain kind of validation etc. Forensics are developing all the time already. So I'm less concerned and that's maybe about the intelligence services or security services than their capability to deal with this this fake fake information. I'm more concerned by think about the general how these different internet services are getting getting affected and polluted and how does that influence influence the greater greater public. I think that's where the bigger problems lie right now. Agreed. For example elderly people have problems currently. I mean today and in the face with these new solutions that that problem is this going to explode. And I think the other other problem there is that we're losing trust as well because we're saying that even if we see something we're saying oh it's probably AI it's how can I be sure. So we've stopped caring as well. I think that's dangerous as well. That is agreed excellent point. So open source intelligence is it's it's a very powerful tool. Is it becoming a critical tool in your toolbox or is it even becoming a primary tool in many cases? So I would say open source intelligence historically has been a primary tool for intelligence services because going back to the 1940s even it was estimated that 80% of everything that intelligence service needs comes from open source. And if we think about sort of outwards looking looking interest of course there is quite a bit of information about the world around us already in open source. And it's important for the intelligence services as well to be able to use that information to their benefit as well because it's the special capabilities that we have they're also limited so we need to be sure that we're focusing those on the right questions as well. And utilizing that open source as best as we can to support that and also to direct those those cars resources to all in the right places. Right. And Rebecca you are also part of the Bellingcat group back in the day and and that is a very influential group in the open source intelligence community. How do you see the collaboration between these types of communities and and and official intelligence services? So I don't think there can be collaboration per se between these kinds of groups so they exist in their own space. And if I think about groups like Bellingcat and others they're often driven by by public interest. So they are putting putting things out there for the for the public benefit. And the intelligence services on the other hand they're very much customer driven in the sense that they there are certain requirements for the customers on what they should report on them. That's where that's where the focus is. So I think there are two separate things that we shouldn't sort of mix mix them together. But that's not to say that that intelligence services could not benefit from the kinds of work that these groups do. Like they're putting things out of the out of the open and we can learn from them as well just from from the information that they put out but also about the methods what they use and how they're able to use it. And if you look at Bellingcat. So I'm not not personally involved with Bellingcat anymore. I haven't been haven't been for quite a while. But if you look at what what they're putting out there on their GitHub for example. So they're putting putting these applications that they've they've developed now on how do you how do you analyze satellite imagery for example and how do you how do you find a shadowfully vessels on satellite imagery and these kinds of things. So certainly it's interesting to look at and how they're approaching these different different problems and learn from that. And in some cases there's code you can just take it take out and start using it as well. Brilliant excellent. So previously we discussed satellites and the new sensors and and how how that is revolutionizing in some ways the amounts of data and the quality of data that we can we can get. Which of these technologies or or sensors or computational capabilities are you personally most excited about what where do you see the most utility or or or so I guess there's I'm interested quite a quite a few things but just the optical coverage that we have and the emerging capabilities there the basically 24/7 coverage that we have now now I think that's that's pretty interesting. Overall we have of course circuit abilities some development finaled as well which are which are we we can see from the war in Ukraine already that they have they have tremendous value value in that that's over use use case but then out of the emerging ones hyper spectral is very interesting to me as well and just understanding how how that that data source can be exploited because the approach that hyper spectral then takes into imaging the world is quite different from the traditional sensors and it brings different kinds of analytical possibilities as well I think we need to approach it a little bit more like a data problem as well so I think those are exciting to be that what we can do with those. Yeah so so if you're an aspiring open source intelligence analysts and and and you're looking to train yourself in in in this domain how how would you go about it or ask in another way how would you train an open source intelligence and an analyst if you were to train one so when I train analyst though okay I also I also lecture at Johns Hopkins University as well I do basic so for open source then what I like to emphasize is when when you want to do it professionally is the process that how do you approach the problem when making making sure that you are clear about the questions that you're trying to answer as well with open source it's very easy to overcollect it's very easy to go down there from kinds of rabbit holes and sort of do do things but never a nest some never really start reaching the goal so being clear about what is the question you're trying to ask bounding yourself as well set a clear clear target this is what I'm trying to do this what I'm trying to answer bounded also in time that this is when I'm going to produce the answer I think that's the first thing that you need to learn because it's very easy to start start otherwise going in all kinds of directions so I think that discipline is one thing that that you need when you start start approaching these things and then the second thing is especially if you're talking about NGO a type of activity where you're looking for for example for war crimes war war crime evidence or something like that then different kinds of things like how do you retain evidence and proof of custody things like that those those actually start becoming really important that you're able to show that this material came from this place I collected it here like like this so that it's still admissible as evidence later on in the chain so that's another another thing so it's it's not just about like hacking away finding finding things breaking breaking things and sort of finding finding interesting things but it's it's being specific specific mindful about the problem that you're trying to trying to address and making sure that the process that you're following is good I think that's great advice for doing basic research as well at NEDU universities so so all the listeners who who are doing research I think should learn from that comments I really like that being focused and disciplined is is one of the key things so and that's when it comes to the sort of methods that how do I do things that there's so much out of out there on the open side now now that you can learn from these different groups that are putting out there up there their pores explaining how they do things there's podcasts and books and everything about the different methods that you can use but one thing that's good to also understand is that if you learn a specific method for a collection today it may not work tomorrow because there's so much change that's happening all the time with the different social media services for example that they're they're changing API's breaking things without that and you don't necessarily even know realize something that's happened until your tool breaks so just it's it's a lot of effort to keep up as well and keep up your up your skills and methods if that's that's the word you're taking right so if if I'm a student of open-source intelligence do you see it more as a as a study of journalism or or data science or is open-source intelligence its own study module or a field of study I don't think it's completely separated from everything else so It can be inspired by things that are done in journalism. And if we look at, for example, what Bellencann has done, it has a collaborate with journalism outfits and produce the kinds of things that would be considered open source in the others is not in the other context. Some of the problems in open source are data problems. So that skill set is also relevant in this. So it really depends what you're trying to do on what you're objective with what the open source work is. And when it comes to analytical skills, then there's very similar to scientific scientific, how science is done, scientific thinking, as well as intelligence analysis, those same kinds of things, things and thinking skills you need in the open source work as well. So I don't think it's completely separate. It's own thing, but it's sort of drawing and bow. It's crossing the boundaries in many different directions as well. And I guess it also depends on your role. What, right, what you're playing there. Right. And it's like, are you working with social media sources or satellite imagery or what are you working with? I think that's also what defines where the inspiration comes from. Yeah. Well, speaking of satellite imagery, and we previously discussed that it's now possible to have 24/7 coverage of the world with relatively cheap means. So Finland, for example, can now start to build data sets of the globe and events that are taken place. But that also then means that everything is becoming visible. And this puts new questions about accountability and transparency on governments. How do you see that development? So I think definitely, like we talked about earlier, those these commercial capabilities that pretty much anybody can use. So I think having those eyes in the skies used by anybody, by non-governmental groups, journalists, et cetera, of course, it's making everything more visible. So it's more difficult to hide what's happening on the ground, and especially in the war zone kind of thing, where it's a very dynamic environment, and things are happening fast. Then clearly having that visibility there in those moments, then it potentially captures things that somebody might want to hide. So it's definitely transparency is something that is increasing along with these technologies, which also means that then is driving accountability as well. So people are more likely to be accountable for their actions. And we're seeing that already with the war crimes and prosecution based on social media evidence, for example. So similarly, in some places we see now, people focusing on use of satellite energy to find mask graves, for example. So these kinds of things are happening already, and it definitely is driving accountability. Right. But this concludes our time together. Thank you for the excellent discussion. Thank you, it's been a pleasure. [MUSIC PLAYING] Enrico, welcome to the show. Thanks very much. Thanks for having me. Enrico, I understand that Guba Space is all about hyperspectral imaging in space. And you also have a constellation of capabilities up there. Can you please elaborate on that? What is it actually? Sure. I think we're doing quite a unique thing on a global scale, actually. We combine hyperspectral imaging, which is kind of a new way of looking at things from seeing things that you can't see with optical or SAR technologies or other methods. But then also a key part of our system is a large constellation, what we call an always-on constellation, which means that we are measuring everything on Earth more or less continuously. So it is not like a traditional satellite where the customer orders an imaging passing from a specific spot. But we are covering everything anyways. And we will be informing our customers automatically about all kinds of threats or disasters or what should they do to mitigate these upcoming threats. So we are more like an highly reliable early warning system rather than just getting information from a specific spot in addition to other methods. Well, that is super interesting. How many satellites do you actually need to enable that? And where are you now on that? What does your roadmap look like? Yeah. So it depends a little bit on where you are in the world. So the more north or south you are, the more frequently we can pass a certain spot. And in equator a little bit less. But if you think of the Finnish border or Finland or the hike of Finland, we need about eight satellites to be on a daily revisit rate. So a measurement once per day throughout that whole range. So the entire Finnish eastern border once per day with eight satellites. But then if you think of the daily globally, so we need about 36 to 40 satellites. But then there needs to be a little bit of resilience towards clouds. And also there are stuff that happens faster than once per day. So we are actually aiming to a constellation of 100 satellites, which then provides to the three times a day measurement, even in the equator, and way more here in the high north and also can tackle them way faster events and also cloud coverage at the same time. We are now first two satellites up. So we're in the early phase. But the first ones are up and operational. And we have the next batch of six in preparations where we also do a little bit of upgrades to the satellites. So those will go up at the end of 26, early 27. So basically in early 27, we already have a daily revisiter in the high north. Excellent. That sounds super interesting. Now I got asked when you say that you're launching satellites. Where do you launch from it? Is that US, European based capability? Yeah, currently it is US, so SpaceX. That's more or less the only option. Obviously we are following what's happening in the launcher scheme here in Europe. However, it's not there yet. There are a few small launchers that are preparing and have done some tests. But it's not good enough yet for actual commercial services. We definitely hope that there will be more options. Just not only because the option is in the US mainly, but to have just more options because small satellite businesses growing globally is also quite booked. So even though it would not be geopolitically the sensitive or anything, everything stands on one company. And from European perspective, so very into perspective, obviously it would be great to have more launchers here. Yeah, well, who knows? Maybe there's a lot of talk that will be increasing, so we shall see. Sure. And obviously we're talking to those companies that are developing these services here in Europe. Yeah. So you did mention the geopolitical situation, and it is a reality that the US and China are dominating the space domain currently. How do you find Finland's capability to jump in there and be a player in that area? Yeah, I think Finland has shown that a small country can be quite a significant player, relatively speaking in certain fields. For example, Finland is really strong in remote sensing. So Earth observation from space. And we have eyesight, we have Cuba space, and then we have other companies that are also doing a lot of stuff in the new space. But like I sized the leader in their market. If you look at like high-perspectral in the world, Cuba space is right there in the top. And actually we are the more or less the only, or we are the only high-perspectral company in Europe that has operational capability. And we are the only one who's in the Copernicus program, for example. And these are coming from country with like 5 million people. So I think that's a really good use of resources. And the government has been, of course, backing us up as companies. But it has not been a part of a national strategy. It has come up from all to university originally. And many things have happened so that co-incidences in a way combined with the entrepreneurial spirit of certain people that we have these companies. And if you look at Sweden or Norway, other Nordic countries, bold countries, they don't really have this kind of earth observation capability. Sweden and Norway are putting a lot of effort in the launch capability. So different countries have different specialties. And Finland, yeah, it's really strong in earth observation. You mentioned the government's whole support and the business mentality of few individuals. But from a start-up perspective, how-- how this is a leading question, actually. How easy has it been from like a financing and getting investment type of journey? Yeah, that's a very interesting question. I think that we-- we're following the trends of technology, VC investments. So it is, when times are riskier, then they need, the investors need, in a way, more proof that the markets are there, as you have traction and all that. And then you have the generic problem of the Finnish VC landscape that we lack, the A series lead investors, like more or less completely. - We do, yeah. - So we have the seed investors. So you get started. And also, I must say that governmental support has been really good that we, for example, have support from business Finland, quite significantly on top of those investments, but then when you go past a certain point, then Earth observation, space-based Earth observation is rather tricky, because you actually, you first need to get the capacity up. And then you need to start, we're selling the data or services or whatever you do with the satellites like in our business model. And then you can kind of show that, that hey, now we have these solid contracts. So you kind of need to finance the satellite first, and the development of the customerships before you really have solid contracts. So that's kind of a problem for many of the investors that don't want to take that market risk. Currently, obviously, this, like the possibilities in defense, which are like vastly growing, and kind of the interest that we are also getting is like mitigating some of that trouble. But yeah, so it's an interesting dilemma that currently the VC, especially like Deep Tech VC world, is quite cautious, but on the other hand, defense is pulling quite a lot. So we are like somewhere in the middle of that. And if the company would have gone to this path, maybe wait before COVID, it would have been a completely different story. But now after the 21-22, when this investment market started crashing, it is not that easy anymore. - Right. - But yeah, I mean, we've seen now in Finland also, like bigger seed rounds, a round coming up also in Deep Tech and space. So that is changing, hopefully, for the good. - Yeah, agreed, agreed. Now you mentioned that you have to prove the process in a way that and send in the data. And let's focus on that data for a little bit. So from what I understand, you're looking from above and you send the picture, if you will. Is that the data or do you also analyze the data to get actionable intelligence out of it? - Yeah, yeah. That's a good question. So high-per spectral imaging, if you start from that, so it's imaging, yeah. We get images, but it is actually like a measurement technology. So each pixel is kind of like a spectrometer. And that's a device that is used to identify and quantify what things are made of. So if somebody would take a picture of me with a high-per spectral camera, you could figure out like, what's the material of my jacket and my shirt? If it is 100% cotton or if there's like 5% like some elastic material in it, or whether some sort of a dark sport on my skin is maybe an early sign of melanoma. So it really goes into kind of a molecular level. So that's kind of what we do. That's the specialty. So we produce what we call hypercube. So it's like image cubes. So there's a third dimension. We have the image, normal to the image, but then the third dimension is the spectral dimension. - Okay, right. - So it's way more information rich, but the information is not straightforward to understand. You cannot see that from the image, like that there's a cotton field somewhere. So you need to develop analytics and algorithms. And obviously it's mostly AI based now. It's like machine learning, but also new methods. And that's what we also do quite a bit. Like half of our company engineering power is already analytics and AI, even though we build satellites and we build the cameras and all this high-west stuff, but that's kind of the direction. So you get a lot of data, which is not straightforward to analyze, but you need those algorithms to produce this robust automated insights to the customer. So for example, we want to provide to our customer not hyperspecial images that they would need to stare at. They would need to hire like an army of PhDs to kind of try to figure out what's in this strange hyperspecial image that we don't understand, but we want to crunch that data ourselves and provide information to the customers, whether it's a warning signal that hey, we have detected a relevant man-made change in this area, you know, please do something about it. So that we have detected a dark vessel over here and its coordinates are this, and we actually have discovered this vessel earlier based on the spectral signature of the vessel. And so it can be the information that doesn't even need to have an image. It can be just a warning signal with coordinates or the type of warning, something like this. Or in the civilian side, you know, 'cause we'll use company, it can be like a yield forecast of wheat in a country updated every day, like 90%, 90%, 91%. So then, you know, the local governments or farmers or traders know what the expected crop is and when it's going to happen. And that's all coming from the same basic raw image. So if we image an area that has fields, crop fields, it has forests, it has water bodies, it has enemy activities, we don't need to take, like an image for each of those customers, for each of those cases, it's a single hypercube. And then we just crunch out of it different streams of information based on the analytics that we're developing. - That is super interesting. So in a way, you're providing intelligence as a service. - That's right, that's our main goal. And one of the reason why we do it is that the more raw data, the more source data that you have to, you know, for the training of the algorithms, the better off you are. And we, 'cause we have our own constellation, you know, we have kind of an infinite source of raw data. - Right. - So think of like, you know, maritime domain awareness, type of applications, like, you know, is there a dark vessel somewhere? - Yeah. - So each of our satellites will see like hundreds of legal vessels every day. So we can teach the spectral library, like continuously to become like really robust. And it like continues and continues, like forever. - That's great. - And then you have a like a really precise information on first detecting very reliably the ships figuring out whether they are dark or not. And like even like which individual ships they are. And where they have been previously. So actually backtracking some of the dark vessels on their route. - Yeah. - And nobody can, if somebody would buy like a few of those hypercutes from us and you know, try to create a model basically, it will always be like inferior. So actually like the level of services and the quality of the service is going to be way better if we do it. So that's like one of the main reasons why we are kind of covering the full value chain. - Oh, that makes so much sense. And then this is really, really exciting capability. It does come with a security question though as well. Because obviously as you gather that type of data and a lot of it, somebody else might be interested in it as well. So I imagine that your security concerns are quite large. - Yeah, absolutely. So we gather a lot of interesting information. We have quite tight security protocols in addition to the like the technical encryption keys that we have to prevent our satellites being hacked or our data pipeline being hacked. So there's lots of like shields and also backup systems. If something would happen, then we also have a way to actually like, you know, shut down, take back. And if you think of our customers, you know, it can be a defense customer, it can be a border control customer, it can be like, you know, police or ministry of interior of some country. But also civilian. So, you know, it's a very, very, very valid point. Yeah, so definitely like the defense requirements for defense are quite different, because you don't get to know all the details yourself even. Yes. In the commercial part, it's all about shielding that company's data, that their competitors can't use it or anything like this. So even with the civilian companies, we have to be really careful. So when we have been building up the constellation and the satellites, we have built in already quite a lot of security in the system, in the pipeline. Right. Now, this commercial utilizations, they're also very, very interesting and you already mentioned analyzing the crop yields. For example, but I can imagine that forestry might be interested, forestry management. What in your view is the most interesting use case from a civilian commercial sector? That's an interesting question. It's a short question, but it might be a long answer. But yeah, I mean, there are now markets that we kind of know that we can help these people and they are interested in this crop yields and it's not being done well currently. So definitely like agriculture overall, that's like a huge market. But the customers there, I mean, if we detect which kinds of crops are growing where, and this is like specific information for hyperspeical imaging. If you use like an optical satellite image and you have a field, which is yellow, you don't know what it is. If it's ochre or wheat or barley or whatever, but with hyperspeical imaging, you can actually detect the difference in the species. So you can map out what's growing where. Obviously the farmers themselves know what's growing on the field. So that kind of information is not necessarily interesting for them. They may be interested more on the, you know, what's the distribution of nitrogen in their field. So good thing you can figure out like when you put the, you know, fertilizer or not, sort of compression agriculture. But then this like a larger scale information interests, you know, governmental entities that are interested in food security in their country. Being organizations that are interested in, you know, buying and selling crops, insurance companies that are interested in whether the, you know, the crops that they ensure are doing well. And if something happens, they want to know it right away. And then how much crop was lost and out of what reason so that they can do the repayment like immediately, saves them actually a ton of money. I can imagine. And also like, of course, their customers are happy if they can do it. And it's all coming from the same kind of basic, basic information. But there's also like a bunch of areas that are not, not served now because actually if you take any earth observation service, commercial service, it's pretty expensive. So if you buy an image somewhere, it's like, you know, from hundreds of euros per image to, you know, 10,000 or more, depends a little bit on the modality. What it's like a bad optical image or like a very high resolution, something or a SAR image. But it's like, you know, it's typically in a thousand if you go and task a satellite and get an image. And obviously like a farmer or somebody, they can't like do that, especially if they don't know how to get any information out of the image. So like these insurance companies, as an example, they want to go to like the African and Southeast Asian markets. But the farms there or the farmers don't have that kind of liquidity that they could afford like a Western priced crop insurance, which can be like, you know, 2000 euros per year. So they can't pay for that. But if it's 20 euros per year, then suddenly, you know, this huge market, you know, market that was not there before, you know, opens up. Yeah, yeah. And so when we have this, well, I mentioned this always on constellation and everything is automated. And we concentrate on our hyperspectral imaging actually to one specific parameter, which is the quality aspect of the data so that it can be automated. That doesn't need like manual checking. If everything's automated and everything is running smoothly, then we can reach the level in the course that actually enables the insurance companies to go to those markets. So if you, if you can take the price down of Earth observation data, you release the access. So you have updating information all around and not only per request somewhere. So you have this like time series, like, you know, daily time series everywhere, which is a really big value in itself. And then you combine the kind of the hyperspectral part in our case, then you can break into these new markets that didn't exist before. So that is like, you know, where we see that a lot of the growth will happen. So yeah, it's definitely what kind of sets us apart from all like hyperspectral competitors who are still working on this way, traditional tasking mode as we call it that. Customer orders an image and then they, you know, task their satellite and provide the image and then the customer needs to like crunch whatever stuff they have out of that image. So it can take, you know, two weeks to a profit year for them to get the actual information. Yeah, that's it. It's not like, you know, it's more like an academic business model rather than even a business business model yet. You know, it makes sense. But this always on concept is super interesting and, you know, the mind goes to a place where you imagine that, you know, you can see and analyze basically, well, a lot of things. Yeah. It has the potential to change governmental accountability and transparency as well. How do you see that in your work? Yeah, absolutely. I think people have not been yet that prepared that this kind of a capability can happen. Yeah. You know, I can imagine. Yeah. So, so, but some are definitely like very interested about it. You know, we also have like on the business side when talk about things like supply chain management, ESG reporting, you know, and this new like for a deforestation regulation by EU to follow up all these things that's happening globally, even though you like have an European company, but what's happening in there like cocoa plants somewhere, Southeast Asia, to have that accountability from a third party, you know, saying like, what's what's really been done on those fields, those specific fields. Then, you know, they don't need to order like satellite asking from us to, you know, do that. They just inform like these are the areas that we're interested in and we're interested in about this kind of stuff. And then we start in feeding them the information with like daily updates and they can then add them to their reports to their customers or use them ourselves. So, a little bit of the same thing with governmental entities that want to monitor like their country, like, you know, farming, do we pay subsidies to the right farmers? Right, right. Like I think like 2% of farms in Europe are checked. And like, maybe? Yeah. Okay. Anyway, so about 98% is like, we're not checking. We're not actually checking. Maybe they use some sort of information for that as well. And, but you know, there's lots of room to improve like how to optimize that and we can help, of course. Super interesting. Ericka, you mentioned that you have two satellites up in the air and very ambitious plans to increase the constellation. But I understand that even with the two satellites that you have, you're getting massive amounts of data that can be utilized. Is that correct? Yeah, that's true. So, we take, you know, with the first satellite, we've taken 10,000, 15,000 images already. Second one is now operative as well. So that that capacity is going up. So yeah, I mean, the customers don't need to wait or we don't need to wait, you know, for a larger constellation to be also commercially operative. All right. So now we are making like the first like major agreements. We have already like a 5 million agreement with the Cobericus program, for example, running on data. So we are operative. With two satellites, we can't like cover the world every day yet. But we do cover the world. So we can already provide like strategic information on different kinds of changes anywhere, anywhere on Earth. So yeah, definitely it's where you don't need to kind of expect us that only when we have 100 satellites then we do the stuff. But now we are already there. And in a couple of years, like I mentioned earlier, daily here in Finland and very soon daily in the rest of the world. So super relevant now and even more super relevant. Yeah, the cable village is just like increasing continuous fee. Exactly. It will never, even though we would not like launch more satellite at some point, the system will still become smarter or exactly because we gather the data. data, crunch it, gather the time series everywhere. So it will be like an ever evolving system, but obviously we need to ramp up the hardware as well. - Perfect, thank you. - This concludes our time together. A huge thank you to our guests for the great discussions we had. And a huge thank you to you, the listener as well. Please make sure to join us again in the Digital Defense ecosystem podcast, where we explore the global security landscape and examine finished companies who produce solutions to global security problems. I'll see you then. (upbeat music)