In this 10th episode of the second season of the Space Capital Podcast, we’re discussing the geospatial market, which is expected to grow from $63 to 148 billion over the next five years. Joining Chad Anderson, is his fellow Managing Partner at Space Capital, Justus Kilian to discuss the massive opportunity of the geospatial market, that's growing quickly and has proven to be recession proof through the markets of the last year.
In the GEOINT Playbook we will walk through the Space Capital thesis, which analyzes geospatial intelligence (GEOINT). This framework helps us connect the dots from the origin and constraints of the geospatial stack to the larger role that space based technology plays in an ecosystem that intersects with the modern tech industry and serves customers across a wide variety of markets.
The difference between GPS and GON. GPS is the navigation, right? It's positioning, navigation and timing. The timing element's very important, but most people think about it in terms of the dot on the map, and which way am I going, and waypoint and wayfinding, right? Geospatial intelligence is the map.
Welcome to the Space Capital Podcast. I'm your host, Chad Anderson, founder and managing partner at Space Capital, a seed stage venture capital firm investing in the space economy. We're actively investing out of our third fund with a hundred million under management. You can find us on social media @SpaceCapital. In this podcast, we explore what's happening at the cutting edge of the entrepreneurial space age, and speak to the founders and innovators at the forefront.
Welcome to the Space Capital Podcast, where typically we speak to the founders we backed, but today we're going to do something a little different. Joining me in the studio today is my fellow managing partner at Space Capital, Justus Kilian, and we're going to dive into the geospatial market, which is expected to grow from $63 to 148 billion over the next five years. This is a massive opportunity that's growing quickly and has proven to be recession proof through the markets of the last year. So needless to say, it's an area that we're highly focused on at Space Capital. At the end of last year, we published the GEOINT Playbook, which is our investment thesis on the future of geospatial intelligence and our outlook on digitizing the physical world. Justus, it's great to be talking with you today about one of our favorite topics at Space Capital. To kick us off, maybe it would be helpful to give our audience a bit of background. Why would a VC firm that trades on our expertise be publishing our insights like this?
We spend a lot of time understanding the way that technology fits together. We research, discuss, debate, invest, learn, and as a thesis driven investor, we can monitor and evaluate how our thesis develops and evolves. We use that research to then inform how we invest, how we're trying to shape the direction of a market. A key part of what we do in our process is putting that thesis out there and helping others see and connect the dots and see how the pieces fit together. So for us, that really started with creating a framework to understand how space technology, which feels so abstract and far away and very limited in its scope. Apollo Landings and International Space Station actually exists in a commercial aspect that touches all aspects of our lives and every major industry. So that research, that deep dive effort for us started with the GPS Playbook that we wrote several years ago and provided a really important framework to help us understand how you go from very narrow use cases for space technology into wide scale adoption. So that set us in motion to today.
We kicked things off with the GPS Playbook. We've written a couple of other playbooks. We've got three key technology stacks that we focus on, GPS, geospatial intelligence and satellite communications. And to frame this up and kind of put it into context, we're tracking investment in the private markets. There's been $270 billion invested into 1,800 unique space companies over the last 10 years. And most of this has gone to satellites. 90% of it's gone to satellites, 9% of it's gone to launch, 1% of it's gone to some of these other emerging areas, but satellites is really where all the action is.
Within satellites, there is infrastructure, distribution and applications. It's something that we discovered when we wrote the GPS playbook. It's a really helpful framework to think about space technology, that the infrastructure is... For example, the satellites, the GPS satellites that Lockheed built, and then the distribution. So that was built by the government for the government and military purposes, and it was really limited in its reach until commercial GPS receivers were developed by Trimble, Magellan, Garmin and others that harnessed that really valuable data, those signals that were coming off of those GPS satellites, and made them accessible to the tech community who then built a seemingly infinite number of applications.
Starting with turn by turn navigation, companies like TomTom in the nineties. Yeah.
So most of the investment has gone to satellites and most of that has gone to applications, and that's where most of the value comes from as well, right? We see the Commerce Department came out with a report right around the time that we put out the GPS Playbook that said that in the US alone, there was a trillion and a half dollars of economic value generated by GPS, and we know that it's generated some of the largest venture outcomes we've seen. This way of looking at the world, infrastructure, distribution, applications has proven very useful, not just in GPS but also in the other space technology stacks.
And in GEOINT, it was really helpful for us also, when thinking about SpaceX removed the barriers to entry in 2009 when they launched their first customer. We started to see all these new constellations of large quantities of small satellites being launched and generating an unprecedented amount of new data from orbit, really timely data about the surface of our planet, movements on the surface of our planet. And then that data was then getting picked up and used in applications as well.
So GPS was a dot on a map. It gave us a location at a point in time. The underlying information, that base map, was created within the geospatial world. And there had been a number of big established players that have been working in that, both on the data collection side, so think Maxar and Digital Globe. And then on the processing base maps structuring of that information, think Mapbox. And these two worlds have not played very nicely together. The Maxar has existed largely to serve governments. Mapbox existed to serve more of a tech community and more of a developer community. And these universes, space-based assets, a very narrow understanding of Earth observation, didn't play nicely into a much broader tech stack. So we've been observing what's been happening in Earth observation with the launch of many satellites.
We've seen some of the first movers in this area, whether it be Skybox Imaging, which our partner Tom Ingersoll helped build and take to exit, or some of our first investments in companies like Planet and ICI that were early movers built not only the hardware, but the full stack in terms of the ground stations, the processing capabilities, and then ultimately the analytics on top of that. Being early movers, they had to build these full stack solutions. It had kept the industry focused on pretty high value customers, low volume, continuing to serve governments and a handful of very large industry players. And it kept geospatial intelligence relatively small.
As we've been watching patterns in terms of our own investments and what we're seeing and what we're talking to with leading research institutions and big tech players and where they're seeing all the pieces come together, we started to see that framework of infrastructure, distribution, application, the specialization was also showing up within geospatial intelligence. And our thesis, I mean we've been working on this report in particular for four years probably in a variety of different shapes, and it's evolved a lot. So the framework gave us a way to see the layers in which the evolution is happening, and I think come away with some pretty unique and powerful insights that connect that world of tech and applications to that world of sensors and data and the hardware and aerospace community.
I think this was one of the big takeaways from the GPS Playbook is that you've got really two communities there, right? You've got the space community, which is the ones who are building the satellites and the sensors and the space infrastructure, and then you've got the applications. These are being built by the application developers, the software engineers in Silicon Valley, the ones that maybe don't know anything about satellites or how they work or they don't go to the same conferences. They're completely separate groups of people with seemingly very different interests. And I think if you talk to either side, you would get a mixed reaction as to... If you go and talk to an Uber software engineer, do you think that you work for a space company? I think it depends on who you talk to. I mean, I think that there are some people who are there that are focused on the GPS signal and enhancing that and making it better. And they would definitely say yes. Many of them probably don't.
On the space side, you get the same sort of reaction. So these communities don't ever really overlap. I think that that's really one of the areas where we've done really well to help bridge that gap at Space Capital. We've tried to do that in these reports and tried to draw those connections between the space infrastructure and how it's actually being used and how it gets into the hands of customers and how it's becoming applications.
And it wasn't sufficient for us to just have this thesis, see history repeat itself, see observations in deal flow. I mean, we went out and talked and worked with Nvidia, Amazon and AWS, USC and professors at Stanford and validated this hypothesis with them as well. Do you see this different specialization and layers? And I mean, it was really helpful, and that's a part of the research that matters. It's not us just seeing the pattern recognition, seeing companies that follow that pattern recognition, but industry practitioners, key tech players that are building out verticals within their offering, researchers that are seeing talent and training their next generation of talent to fit into these verticals. It was very confirming of that initial thesis.
In the era of SpaceX where you've got access to orbit and we've got new entrepreneurs coming in and innovating and experimenting with different platforms and different sensors and things, all of that's really interesting, but that's just one part of the puzzle, that the Earth observation piece is one input into a much larger geospatial market. I think that's one of the key takeaways from this research that we put out is, again, connecting those dots and bridging the gap between the Earth observation and the geospatial market at large.
Another important insight that came out of that, we often approach space as a lens by which we look back at Earth and see how data and technology is flowing down into different industries. So satellites was a natural starting point on the infrastructure side for us. But very quickly, you realize that satellite data alone is insufficient. So oftentimes when people are talking about the innovation that's happened within Earth observation, it's gone from optical to SAR to infrared to hyperspectral and multi-spectral, all these different types of sensors, which is great, and they're more abundant and there's commercial providers, and that is true. They're very complex. They offer a very rich layer of information, but satellites offer a certain set of benefits. They're inherently global. They're farther away from Earth, so their resolution is going to be less. Their cost point relative to some other data collection methods is lower given the scale that they operate at.
But in itself, satellites are only one entry point into that critical infrastructure and the platforms by which you can capture data. This goes to the beginning. I mean the NGA has been a pioneer in tipping and queuing process where they use satellites and broad SWAT detection to detect something that's interesting or relevant or important, and then queue up another process, whether it be aerial or drone or even much more localized even now to the point of cell phones, to be able to get much more focused, timely, precise information about what they observe at a broad scale. And the industry, geospatial intelligence moving beyond just sort of a narrow earth observation definition, that is also becoming possible. You have high altitude balloon platforms. You have aerial data that's being captured. You have drones that soon are going to be able to travel beyond line of sight. You have a proliferation of handheld and mobile car base data collection methods.
From the surface of the Earth all the way to lower earth orbit, you have a tremendous amount of insight that's being captured and creating different geospatial perspectives at each layer. And those sensors exist at each of those different layers. It's a much more holistic way to see how all the pieces fit together, and it's learning from the practitioners, literally the people that created the intelligence, how they do this process. And it's now becoming available at commercial partners in centralized data repositories where you can get access to this information. It's not one data provider. It's many on many different platforms and many different sensor types.
Yeah. So that's the infrastructure is that we've got everything from Earth orbit and we're looking at things from a global scale, from a very unique vantage point in space, all the way down to ground sensors, which is giving us super high resolution, very specific information. The power is unlocked when you fuse all of that data together and you start to get unique insights because each one of those different sensors has trade-offs, right? It's better at something. And the more specific you get, the less global you get and vice versa. So having all of that at your fingertips, putting it all together in a way, fusing the data together is all very interesting. And in the report, we go through some of the key things that have unlocked this, right? I mean, it's not just sensor platforms, it's also compute GPUs.
Moving from that infrastructure side, all these different platforms collecting data, the exponential growth in information that's coming off of these platforms creates a challenge to users, adopters. There's barriers to actually making this information usable, manageable. So companies like Digital Globe were processing petabytes of data on premise using tools that they'd built in-house, moving that infrastructure and that data capture into the cloud through what Amazon and Microsoft are now being able to do, to be able to downlink directly into the cloud, allows you to bring not only modern CPU and TPU and GPU capabilities into the computation side, but also storage benefits and a whole ecosystem of modern technology tools, whether it be AIML or as-a-service business models that allow you to make better use of your data. Well, first process it, interpret it and create value from it at scale, which had been very, very difficult in the past. Now we have a whole new set of APIs and developer toolkits that allow us to fuse that information, get much more precise and more sophisticated.
We're seeing scientific discoveries every week on the sort of benefits of sensor fusion and new deep learning techniques to be able to extract more and more value out of this information. So it's going from a very verticalized complex tech stack where academics, PhDs, they were the only ones that could access this or very well resourced companies to now a Python developer that has certain access to libraries that they can bring in this information, plug into an API, spin up an instance on AWS, and do some pretty incredible calculations and come away with some very powerful standardization and insights that could be integrated into a product or an application. So that's the shift that we're seeing. Particularly when we move from infrastructure to distribution, it's these as-a-service scalable solutions that allow you to process more, interpret and go much farther with the data, with much less background knowledge about the satellites, the sensors and where it came from and how it came.
And a lot of that's enabled, like you said, through cloud services and cloud infrastructure, these big tech giants coming onto the scene, enabling that, collecting that information through their ground station as a surface. Now we're fusing it with all this other data. We're making a lot more use out of it. We have a lot of advanced cloud compute capabilities that we're now applying to this. Satellites were pretty late to the game with regards to cloud. It wasn't until, I think, 2017 the digital globe moved over to the cloud versus on-prem servers. So this has had as big of an impact as low-cost launch has without a doubt.
So now that we've got this, we've got all this data down, and we're now between some of our portfolio companies, I mean we've had SkyWatch on the platform. Right now, it's a matter of removing the complexity from the system. So instead of having these verticalized solutions, whereas a customer, you would have to go out and talk to each one of these customers. You'd have to, one, know the landscape, figure out who provides what, reach out to them specifically. In a lot of cases, they don't even have a website. You have to call them up on the phone. You have to negotiate your own pricing. They have minimum order requirements that are really big. They have legal contracts and things that you have to negotiate with each one of these providers. These are high barriers to entry.
Over the last few years, SkyWatch and some of the big tech companies have removed a lot of that complexity from the ecosystem, which is great because they're making it much easier to access this information. They are aggregating all the data onto a single platform, fusing all of that data together and making it really easily accessible through an API. And we're starting to see the first applications leveraging access to this new data now.
Within distribution, we have seen two very interesting advancements that are very subtle beyond cloud, beyond processing, storage, compute, downlink. The first that I think is worth talking about is SkyWatch. They are able to centralize and aggregate a lot of the data suppliers, Earth observation, but also aerial like we talked about, structure this data, ensure common file formats, make sure that it's clean and orthorectified and there's no cloud coverage in the imagery. And making that data then accessible based on a geofence location that allows for high volume but very low value customers to start to engage with this data. Now that may seem trivial, but the entire Earth observation data capture and infrastructure is not set up for that. They're focused entirely on large value, low volume customer segments. So this essentially API-based marketplace is opening up entirely new customer segments and centralizing access to information. I mean, that's a huge step forward.
So one is on the supply side, the achievements that they've made. The other is actually on the demand side, and it's pretty clearly known that geospatial intelligence can be used within mobility, it can be used within agriculture, logistics. These are the big known common use cases. But developers experimenting with this data, you can actually then start to see what sensor types and what resolutions and what altitude or elevation is most in demand in terms of what the market wants to build their necessary solutions. That's a fundamental shift in how infrastructure gets built historically. That capability has been scientific imagination saying, "Hey, we can build these satellites. We have this really awesome sensor. This is the resolution, so we're going to put it up there and just make this data available and people will build things on it and they'll figure it out."
That was fine for free publicly available data. It allowed an early geospatial market to experiment and create these small consulting type services where people are developing that, but it didn't create high quality, repeatable, scalable data that can be used for commercial purposes across very specific industry verticals. And it doesn't fill in the gaps. It doesn't focus on the customer needs and then build the sensors and align with actually a demand driven innovation cycle. And so we're starting to get to that point. As you understand where innovation and where these new customers are starting to demand sensor types and data, you can actually build solutions for them. SkyWatch has built a solution similar to their EarthCache. They now have TerraStream, which allows them to help align a two-sided marketplace that actually makes the market more efficient, creates and captures data that people actually want and are willing to pay for. That's a big step forward.
Given the time it takes to build satellites and sensors and put them into orbit and the cost, I mean, it's a bit pretty difficult proposition to propose that, "Hey, I'm going to launch these satellites. If we build it, they will come." That sort of technology push has been dominating the category up until very, very recently. There is an infinite number of use cases, and that's the beauty of it is when you remove the barriers to entry and you just allow the innovation to come in and you allow an infinite number of people to develop an infinite number of use cases, the people who are closest to the customer develop solutions for that customer.
That's when you start to understand what customers actually want out of this data. It might not be a better sensor, it might not be better resolution. They might not care or they might care a lot, but you don't know until you start to talk to the customer. That's the number one rule in startup is, if you've got an idea for something that someone wants, you go out and you talk to those customers and get some feedback. When you start to flow that upstream, it's really, really interesting to see where the gaps in the data are and where they can be filled.
I mean, it's very obvious when you look at the commercial providers within Earth observation, they're at 50 centimeter resolution and all optical data. That's where people have been competing, and now we're getting to a point where it's 30 centimeter, that's what people are competing for, and that's a little bit of SAR. We know based on these demand signals that there's interest for over 280 different types of data and sensor types. Right now, the market is providing 10 to 15 of those. So there's a huge gap in terms of what the applications need and want and what the engineers and the businesses are creating and providing. So they're all competing on razor thin margins for a select group of areas based on known use cases, and they don't understand this huge potential that exists out there. So that's what a marketplace can help unlock. It can create these new use cases, put the right types of sensors up there and build a lot of new opportunity.
The second major development that we've also invested in within that distribution layer goes to the training side of the AIML challenges and actually acquiring sufficient training data. The NGA has a great quote. I heard this a couple years ago at one of their annual conferences, and longest, most sophisticated user of geospatial data, they have said their biggest challenge is getting sufficient training data sets to actually make use and actually get value out of their observation data.
We've got too much of it now. That's the issue is we've unlocked the access, we've put it all in the cloud, now we have too much of it. How do we make use out of it?
So synthetic data is a really powerful tool to help you quickly create what you want to train for, do it at a very low cost way, define edge cases or unexpected scenarios that haven't been directly observed to bring into your model and rapidly accelerate, and I would say, increase the resiliency of what you're trying to detect for. And that cuts way beyond. When you get to this level, you're thinking more than just satellite imagery. This shows up in healthcare and it shows up in autonomous vehicles, and all of them are using computer vision and optical and different types of imagery to make critical decisions. The use of synthetic data to train and improve and increase the robustness of your models and your detection capabilities is incredibly valuable. I think you see that in that distribution layer, the horizontal nature of some of these tools, the building blocks that allow developers to do more and build better engineering tools, software capabilities, detection methods that bring to bear the ease of application development.
Now that we've unlocked this access, now that we have made these very dense data sets available, accessible in a format that makes sense to developers, that fits into their work streams and workflows, allows them to build end user applications on them. We're now starting to see the first of those come to market. This is an area where we've been watching very closely because this is all blue ocean stuff. We're basically witnessing the birth of location-based services, but we think that this could be much, much, much bigger than that $36 billion market.
The number of applications are seemingly infinite. We're seeing it built into insurance through Arbol and they're parametric insurance platform that is using satellite data to validate claims and have an objective measurement and an automatic payout of insurance premiums. We have an agriculture play with Regrow who is using satellite data to power their software solution. Long story short is that we have a few of these in our portfolio, but the potential here is much, much greater.
When do we start to see this really take off? It feels like 2023 could be the year where we really start to see the application developers get ahold of this data and start doing really interesting things with it. And one of the key reasons why I think that is because geospatial data is providing valuable insights to enterprises and governments, both of which are large customers that are willing to pay, and we're seeing them put their money where their mouth is. Through the market downturn in 2022, there has been a focus on business fundamentals and revenue, and government contracts make up a big piece of that. And in Q2, the National Reconnaissance Office, one of the big five US intelligence agencies, they made their largest ever purchase of satellite imagery directly from the providers.
But there is a lot of demand for this data because it's so important for enterprises to understand the market, whether it's going up or down. It's providing them information to be able to assess. What we like to say is that as the world becomes more dynamic and uncertain, these customers want more of this data. So because of that, it feels like in these tighter capital markets that there's going to be more opportunity. People are going to be looking for the opportunities where there is business fundamentals, where there is opportunity for revenue and near term revenue. And it feels like this is one of those areas where you can build a geospatial application and get to revenue pretty quickly.
I think there's a handful of drivers there that you highlight, which is in times of uncertainty, the value of data becomes greater. Companies are increasingly relying on data to make decisions. So I think the pandemic was a great example. You weren't able to send people out to every location to validate and create a sense of ground truth on what's happening. The adoption of satellite data and the willingness to experiment and utilize that capability became more important, not less important. So there was a greater willingness to experiment with new ways of doing things. Now, with the financial downturn, large companies are having to find ways to figure out how to streamline their operations and improve revenue.
The difference between GPS and GON. GPS is the navigation, right? It's positioning navigation and timing. The timing element's very important, but most people think about it in terms of the dot on the map, and which way am I going in waypoint and wayfinding, right? Geospatial intelligence is the map. This is essentially four or five years of research that we've tried to distill down into this report. So hopefully it was helpful. Hopefully you learned something. We certainly did by going through the process and by writing it up and by thinking through it, and by, as you mentioned, reaching out to some of the leaders who have been in this category and been in this field for a lot longer than we have. So hopefully it gives you an interesting lens through which to view the opportunity. We think it's massive. Let us know what you think. If you'd like to hear more from us, we can talk through some more of our research. Yeah, hopefully you enjoyed it.
Thanks for tuning into the Space Capital Podcast. If you enjoyed this episode, please leave us a review and subscribe to make sure you never miss an episode. And if you're interested in learning more about investing in the space economy, I invite you to visit our website spacecapital.com where you can get access to more industry leading insights and learn how you can join the entrepreneurial space age.