PARIS — Many New Space companies shy away from talking about costs as though it’s an inappropriate topic. Emiliano Kargieman, founder and chief executive of geospatial data provider Satellogic, thinks it’s not just appropriate, but essential.
Satellogic has eight 1-meter-resolution optical observation satellites in orbit and another 16, with 70-centimeter imagers, scheduled for orbit aboard China Great Wall Industry Corp. (CGWIC) Long March rockets by mid-2020.
The company has raised slightly more than $100 million, enough to reach cash-flow break-even based on the orders booked so far and those expected once the new satellites are in service. From there, it’s on to a 90-satellite constellation that Satellogic says will offer a combination of low-cost, high resolution and revisit frequency to unlock a much larger commercial market beyond the initial government customer set.
You are building your own satellites, as is Planet and Spire. Isn’t the supply chain now sufficiently developed to outsource that?
We consider ourselves to be the only completely vertically integrated high-resolution Earth observation company. Spire and Planet are essentially building cubesats in house. I would argue that for cubesats today you probably have the supply chain to do that. There might be not necessarily a lot of gains.
We chose to be vertically integrated for two reasons. One is economic. To reach the unit economics we need to transform this industry, we need to squeeze every cent out of the cost of satellite building and mass manufacturing. Vertical integration has allowed us a 10x lower cost than we would have if we had to buy pieces outside in the market.
The other part is that vertical integration gives us the flexibility to solve the issues where they need to be solved. This is particularly clear in the payload part, such as the camera, where we chose to solve issues at the adaptive optics level in our camera where we could have chosen to solve those issues by software later. We picked the one we think is more efficient.
Our camera technology and the adaptive optics we’re building gives another 8x efficiency advantage compared to any other satellite in our class. If you put both things together, you can solve the problems where it makes more sense. Plus there is a cost advantage when you integrate vertically and can really push to build the right technology for the mission.
I’m not saying it’s not worth paying 10x more for a great product on the market. I’m saying that product might be well-fitted for a variety of reasons of different missions and we just need it to do one specific thing. So when you’re building technology that’s good enough for what you’re doing, there’s a lot of efficiencies there.
So if you put both things together for us that gets us 80-100 times better cost-efficiency than anybody else trying to build high-resolution imagery satellites today.
You talk as much about cost as about performance.
Delivering low-cost data and both high resolution and high frequency to customers is the key to unlocking the market potential.
To do that, you need to put a lot of high-resolution imagery satellites out there. To justify that infrastructure investment you need to lower the cost to the point where the market can support it.
Reaching those unit economics has been the driving principle since the start of the company. We were finally able to prove that technology in orbit over the last few years. Now we’re starting to scale it up. The vertical integration and the fact that we were able to solve the problems at the right technology layer are really the things that give us the differentiation that will build our company.
There are multiple Earth observation constellations now being built by New Space startups. It’s hard to see how they all survive.
When you have a new wave of technology companies like this ,you get investors excited early on. But they don’t really have the capacities to discern what a good investment looks like compared to a bad investment. They don’t have the track record or the technology experience.
You see money going in a little bit indiscriminately. What you find is many of these companies of the first wave don’t survive. They don’t have the right product, they aren’t aiming at the right markets, they haven’t made the right decisions.
This is true in the EO business but it’s also in the wider satellite world. I’d be more concerned today about the failure of some of the larger LEO communications constellation and the effect that might have on the market than I would about any other player on the EO market, which has received comparatively little investment and little notice from investors so far.
If any of these big bets fail, that will pull investors down, but it’s also healthy. It means money will start to learn to follow the quality deals, and think more about the fundamentals. It’s also healthy as a discipline for companies in the industry to execute better, to focus on creating real value, not so much on selling big pie in the sky missions but to focus on delivering true value to customers.
So I’m not only not afraid of that, I’d welcome it in that sense.
What is the status of your orbital infrastructure and what’s the service life of your satellites?
We’ve got eight satellites in orbit now. We’re building 16 that we’ll be launching in the next eight months. We designed them for a three-year lifetime in orbit, we’ve been operating some of them for over five years now. So you could say we’re over-engineering a little. But 3-5 years is a reasonable lifetime for our business model. I’m more comfortable renewing our fleet and renewing our technology in orbit completely every three years than operating a longer-lived infrastructure.
There’s the capex-refresh rate that is built into our business model. In our case, because of unit economics, that capex refresh rate is significantly lower than anybody else’s.
That’s a strong argument for why we can support vertical integration. The satellite sector is still too small to really reap the advantages of the economies of scale that come with a diversified supply base. If you’re going out to buy reaction wheels for a satellite, there’s just not a large enough market that you’re going to have really a lot of economies of scale with a company that only does reaction wheels that can survive and innovate.
So for the time being, vertical integration is what makes more sense in this industry. It’s going to change. With more and more satellites being built, there’s going to be room for specialization and for successful satellite component companies. But we still don’t see that. We still see that vertical integration is really the way to go. It makes things more complicated from the execution standpoint, but the advantages are very significant.
So the ecosystem isn’t there for sustainable expertise that can maintain high quality in different components.
This is in flux. We’re constantly looking at opportunities where we can develop some of our suppliers into taking larger parts of our component or subsystem level stuff. We push more things to our suppliers, sometimes we take more things from our suppliers. We understand that’s an advantage for us. But the driver for how we make decisions about supply chain is mainly about cost, quality and reliability of that supply chain. When you have the right combination of these, it’s doesn’t matter if it’s us who is doing it or someone else. It’s just hitting those numbers.
Where are you on financing?
Since the beginning of the company we’ve raised a little bit north of $100 million and we’re well funded for this next stage of our business. We have not been required to make announcements on funding rounds, so we haven’t done that. But you should hear something from us before the end of the year, clarifying a bit of that. But it’s safe to say that we’ve raised a little bit north of $100 million over the last few years and we’re well-funded for what comes next for us.
Was the last big tranche the June 2017 Series B?
No, we closed some funding since.
How far does the cash you have now take you?
It will be enough to put our company at cash-flow break even and good enough to continue to execute on our plan — launching the next 16 satellites we’re building now, starting to serve customers and growing our dedicated satellite constellation business.
We’re seeing a lot of traction in the market. If it came to it, we could continue to grow the company organically based on cash-flow. We do expect that we might want to accelerate the deployment of the constellation of satellite and may be looking for additional funding next year at some point. But we’re in a solid position where the funding we have now is good enough to get to cash-flow break-even and to help us to continue to grow based on our own revenues. It’s a first, I guess for the EO industry, or at least for the new wave of EO companies.
A couple of the eight you’ve got now might be nearing end-of-life by then. But that will give you say 20 satellites to operate for a bit. Can you start revenue generation then?
We expect to end this year with around $80 million in bookings and a very healthy pipeline for our dedicated satellite constellation product.
We recently announced a $38 million contract to partner with a Chinese imagery analytics company, Abdas, to support them with monthly remaps of the province of Henan for their own analytics.
For us it was a validation of the dedicated satellite constellation business model and a great first step into the very interesting Chinese market.
We see a strong traction for a dedicated satellite constellation product in Latin America, in the Middle East, Southeast Asia, in Africa. Essentially it’s a product that allows a country, or a state government, to kick-start their own Earth observation program for a fraction of the cost it would take to buy a single Earth observation satellite. Its a turnkey solution with no upfront capital investment, without the technology or project risk.
We really see our revenue growing strongly based on this product over the next 24 months. The satellites we’re currently building and launching in the next eight months will allow us to serve all of the customers that we’re in conversations with for this period of time, so I think that’s a strong part of the outlook we have.
At the same time, we’ll grow this constellation, from the roughly 19 satellites we’ll be operating from mid-next year to the 90-satellite constellation that will give us weekly remaps of all of the surface of the planet at 1-meter resolution.
For us, this is the big strategic turning point. We’ll will be in a position to start opening this new data set for a wider market and really start going after some of the more mainstream applications for high resolution Earth observation that we believe have not yet been served because the cost, and frequency and quality of data is just not there yet.
Airbus has offered a similar service, so does the PanGeo Alliance a nation buys one satellite and then gets access to a full constellation.
In our case, the customer gets essentially complete control over a fleet of satellites on top of their area of interest. They don’t have to rebook capacity or go through a vetting processes or agree to any kind of shutter control.
They don’t have to compete for capacity in their area of interest with essentially anybody else. They get complete exclusivity of control over a number of satellites on top of their territory. They do their own scheduling and tasking and the data is end-to-end encrypted and downloaded into their own private cloud where they can do whatever they want. This capability today is unique. No other service provider in the world is in a position to provide this to customers.
The Abdas contract starts in 2020 and covers how many years?
It’s a six-year contract. What our customers tell us about the advantage of the service that we’re offering, it really gives them for the first time complete control, the privacy of tasking and processing the data and it comes at a small fraction of the cost of the direct access service of DigitalGlobe or Airbus, a significant selling point.
China is becoming more self-sufficient in EO data. How were you able to crack the market?
In this case it’s not at the national government level, it’s at the province level. It gives the province autonomy so the do not have to compete for resources with neighboring provinces or with anybody else in China.
The sheer volume of data that they’ll be able to collect is something that was not previously available to our partners there. Those were some of the important differentiators. They couldn’t get enough high-resolution data from existing satellites. Above all, they couldn’t get the guarantee that they would get the frequency and resolution of data that they needed. They can count on this data at the right frequency and at the right resolution. For services it’s really paramount.
When you’re competing for resources with other stakeholders, and that applies also to direct-access programs from Airbus and DigitalGlobe, having the guarantee that you’ll have the data you need at the frequency that you need is something people are willing pay for. And when the cost is right, it makes it easier to make the decision.
Is this kind of deal unique to the Chinese market?
I think you will find a similar thing happening elsewhere if you look at different government agencies, and different groups trying to access capacity in a given nation. It’s a similar conflict conflict — multiple groups trying to access the same resources and wanting guarantees that they will be able to access the resource. So there are other market opportunities besides China, maybe not marketed in quite the same way.
Looking out 3-4 years, you’ll have next-generation 30-cm Legion satellites from Maxar/DigitalGlobe DG, the 30-cm Pleiades NEO from Airbus and maybe others. A 1-meter ground sampling distance is a good place to be but will you need to go to sharper imagery fairly soon?
For the applications most people are interested in, 1-meter resolution has been the sweet spot. The cost of the infrastructure to get to high frequency and low-cost of distribution for the data is right.
We’re pushing into higher resolution in the next generation of satellites we’re building for next year, going down to the 70-cm range.
Sub-meter gives you an increase in capabilities at a relatively small price to pay in terms of swath and coverage. I don’t see us going into 30 cm or anything close to that in the next 3-4 years. Again, there’s a lot to be said for data acquisition and distribution at the right combination of frequency and cost.
We’ll build our first constellation that gives us the data at 1 meter and slightly submetric resolution over the next couple of years. Then we’ll continue to grow this constellation to get daily remaps of the planet in that resolution range. We think that’s going to be a game-changer for the commercial industry and the demand for high resolution will probably come a couple of years after that.
What’s the difference between the swath width at 1 meter and at 70cm?
If you use the same center it’s linear, so you lose 30% of the swath.
What’s the swath width of the satellites in orbit now?
It’s around 5 km for the 1-meter resolution satellites.
Do you do image resampling for your customers to produce sharper images or is that not something they want?
We capture 1 m resolution natively in every spectral band. We don’t have to do panchromatic sharpening.
Our spectral resolution it’s actually better than [DigitalGlobe’s] WorldView-3. That satellite offers 30-cm panchromatic images, but when you go through the spectral domain, it’s 1.2 m in resolution. We actually have a 1-meter native spectral resolution. When we talk about our 1-meter- resolution data, we’re not considering any super resolution or post-processing on the data.
Is that something customers care about?
The structure of what we call customers is changing a lot. If you’re thinking about distributing images for customers to look at with their own eyes, I think you’re correct. The new generation of customers for EO you’re talking mostly machine learning to derive some knowledge about what’s going on. To that extent, the information that exists in the raw data is more important than the tricks you can play in post-processing. That changes the equation.
Will government be the dominant market for commercial EO for the foreseeable future?
It’s really about the markets at any particular cost of distribution. The explosion of value-added services based on EO that came after Landsat Imagery because it was widely available and free gives you an idea of how markets can grow once you hit the right distribution cost.
High-resolution data continues to be extremely expensive. There is no way but to task satellites for the imagery targets you want, and tasking a satellite carries with with it a huge opportunity cost. You are aiming your satellite at a given place and not somewhere else.
As long as this is the case, you are going to be building your market case based on what applications support your data acquisition and distribution cost and not on what value you can add to a customer and how much they are willing to pay based on the value that you add.
Until we break out of that problem, the commercial market for high-resolution data will not grow to reach its full potential.
I believe we’re in a position to solve that problem for the first time. The constellation we’re building will be the first one to allow us to tap into mainstream applications for Earth observation based on high-resolution, high-frequency, low-cost of collection and distribution of data when we have enough satellites that customers don’t have to pay that opportunity cost.
What are the technical characteristics of the 16 70-cm satellites launching next year?
The satellites are 38.5 kg dry mass and go up to maybe 40-42 kilograms launch mass, depending on how much propellant we use.
And the orbit?
We fly them around 470 km.
You can get five years’ service life at 470 km?
W can, yes, keeping them there with their propulsion systems.
Who owns the IP for your satellites?
We own our own IP. We design our own adaptive optics and we design our own cameras and we have patented technology around our camera design.
In addition to governments, what are your biggest commercial verticals for the 19-satellite constellation to be in orbit around H2 2020?
We have some traction for the dedicated constellation product in the insurance and reinsurance industry, but that’s the only case on the commercial side for this particular product.
We have been working with commercial customers in the oil and gas industry for infrastructure monitoring, in the agricultural sector for supply chain management and in forestry for operational management and in some cases for cartography.
Given our constellation now and going into next year we will continue to work with those customers but on a limited capacity. To really serve them at the right price point and at scale we need to launch more satellites and that’s what we are doing.
You selected China’s CGWIC as your launch service provider. What criteria did you use? Price, schedule assurance, other factors?
It is 95% price. We are really, really focused on creating value for our customers. We believe the only way to achieve the potential for this industry is to achieve the right unit economics. We are all about unit economics.
Launch is a significant part of our capex expense. So cost is the main driver, by far, of our decision to operate with one launch provider or another, and then availability would be the second.
The combination of cost and availability is driving the decision for us and the truth is, availability is significantly more important when we start scaling the constellation. It becomes less of an issue when we have to replace satellites that are in orbit. Cost will always the be the most significant driving factor.
Your contract with CGWIC was all for Long March 6 or for a mix of vehicles?
It’s a mix of both Long March 6 and Long March 2D.
Are your investors going to be looking for exits in 6-8 years or are they more patient capital?
Our strategy that we picked when we thought about partnering with our investors has always revolved around thinking about the time horizon of the project. When I started the company back in 2010, I knew that to accomplish what I wanted to do with Satellogic it would take me around 15 years.
When you have the time horizon in your mind, you don’t go after venture capital. You don’t go after institutional investors on a seven year cycle. You try to surround yourself with investors who are a lot more patient than that.
In our case it’s a combination of strategic investors and family offices and wealthy individuals — groups that can think in a longer time horizon. That has been most of our investor base over the last few years, Tencent being one of our investors but there are also many others.
When we picked our partner that was a significant factor.
Who’s another noteworthy investor?
[Fondo] Pitanga in Brazil, a family office that has been backing us since our Series A and is taking a bigger and bigger position over the years. They are a strong supporter fo the company. Family offices in general, from agriculture and oil and gas industries.
You really had a 15-year time horizon when you started?
I did. I think we’re on the cusp of a really big transformation in this industry and with the potential of space, as a market, to help some of the big problems we have here on Earth. For me it was a career-defining moment. I decided I would spend the rest of my life building technology for this opportunity. I took a long-term view and I still have that view.
If it takes 20 years or 20 years instead of 15, that’s fine. I’m here for the ride.