Today I’m sharing part two of my interview with Gabe Dominocielo, co-founder of Umbra Space. If you missed part one, where we talked about things like how Umbra differentiates itself from other Synthetic Aperture Radar providers, using SBIR’s to grind out early tech development, and much more, I highly recommend starting there first.

In this newsletter we talk about other ways to bootstrap a deep tech startup, getting to market later than competitors, contrarian approaches to the Earth observation market, and more.

Now let’s dive into part 2.

Sponsored by…

The SpaceCom conference is a catalyst for advancing every space project–no matter the mission. Use code ‘SPACEDOTBIZ’ to get 20% off any conference package to SpaceCom, the 2nd annual Space Mobility conference, and the Global Spaceport Alliance Summit. Get your tickets here!

So you said there are kind of three ways to build slowly and grind it out on government contracts. You mentioned the SBIR way. Were there two other ways you had in mind?

You can be a subcontractor to a prime. And the prime doesn't have to be Lockheed Martin. It could be like Deep Space Systems or some other small contractor. And you just go and work hourly, and then you develop your own product or your own dime. So you build your business the way that every successful aerospace company has. Every widget, every little screw, everything on anything. Blue Canyon was bootstrapped this way, all of these companies before this craze of Silicon Valley. I mean, we landed people on the Moon without venture capital.

So to take the other side of that argument, what do you think are the appropriate conditions for a startup raising a large funding round within the first couple years, let's say? What would it take to ever make that make sense?

Growth. Look, obviously I didn't bootstrap the whole company. We did grind it out for two years. We did ultimately raise money, but in an ideal world, I would've raised no money and would've been General Atomics. So the answer is “How much of the company do you want to own?” So if you are doing a widget, which many of these space companies are like whatever, a component company, do you need VC money for that? You should have the ability to see if a customer will buy that product without fundraising.

If you want to go and scale some huge constellation, you definitely need outside capital. But if you're doing a widget, you don't need to go to space. You don't have to pay for launches. You can go and get customer money from DoD and then DoD can help you support commercial customers with more money. The thing is, it just has to be slow. So the alternative is how much of the company do you want to own? And if it's like, I need an exit so I can get a million dollars and buy a Porsche, great. Then take as much money as possible, SPAC, and have a great time, but at that point you are not building a business over long periods of time.

How do you think about being last to market when you take that slower route?

We’re the last. Look I’ve had all these companies who’ve been warming up the market with hundreds of millions of cash spent in educating customers. Then we came in and we undercut everybody’s prices and sold a better product.

It’s not software. With software, first mover advantage with a bunch of early adopters and network effects is a huge advantage. If you have a billion people using Facebook, they’re probably not going to switch over.

If you’re in a market with a thousand customers who all hate each other and they’re competing, and one of those customers suddenly is like “I’ve got a competitive advantage by paying one-tenth the price for a ten times better product”, first mover advantage does not matter because everyone wants to save 90% of their money and will switch to that product.

Earlier you mentioned partnering with a prime. Umbra is partnering with Maxar. First of all, I want to make sure I understand that relationship correctly. Does Umbra provide satellite data to DoD through the Maxar platform? How does that work?

So Maxar is able to sell our data to almost anyone, except of course some bad actors, neither Maxar or Umbra sell to or will do business with bad actors. But really there’s more to this question. So we've worked with a prime in some way for almost our entire life of the company. And those relationships are not public, but Maxar is our most public partnership.

Our relationship with Maxar is like dating. So we've been talking for a long time. We have really, really clear communication. We have a ton of mutual respect. The way that I saw the partnership is that it was inevitable because we modeled our unit economics off things that Walter Scott [Maxar CTO] said.

Walter Scott did a remarkable video on Worldview Legion [included below] which clearly explains optical satellite unit economics

So we really learned a lot from Maxar and had a tremendous amount of respect for them. They also realized that we have a capability that essentially no one else has been able to prove and they have the best capability, so we're kind of stuck with each other, but it's okay, we are in love.

So you don't have a concern that Maxar is going to say at some point, “We just want to do this ourselves”

I'm sure they could, but they are too smart to do it. I would not wish building a synthetic aperture radar satellite on my worst enemy. It is really, really, really difficult. And then on top of that, competing with Umbra technology is so difficult literally no one has done it.

With Maxar, it works way better to complement each other than to compete. I want them to win, now they have their gigantic sales force selling our data. They're making money. We’re making money. There are no losers, there's no gotcha. I think there's a lot of perception that the primes will just come in and crush you and try and do it.

But also, when we did the deal with Maxar, our valuations were not that different. This was before their acquisition by Advent. At the time, they were worth under two billion dollars. And we had just closed our financing at over close to a billion. So it wasn’t like there was a totally lopsided power dynamic. Plus we also like each other.

How do you think about the fact that you don't necessarily own that relationship to the DoD customer because it's through a prime?

So obviously, Maxar does not own our relationship with DoD. We still have our own relationship with DoD and with the Intelligence Community. DoD/IC and Umbra need to have our own special relationship and direct clear communication. So they have to understand that Umbra is trustworthy and reliable. So the important stuff, we do prime with the US government.

But if Maxar has had a relationship with a USG program manager or some copper mine for 12 years and wants to sell some of our data to supplement their cloudy data, great. Remember, I'm selling Nike's, I don't care. It's a unit that produces money. Distributors sell Nike's. Fabulous.

So your team has placed a large emphasis on commercial customers. But in a recent TechCrunch article, you stated that 70% of your revenue comes from government customers, and I've heard for years people saying that the use case for remote sensing data is ready to break out. Why do you think that you're able to unlock a commercial customer that has been so difficult for other remote sensing companies to unlock over the last 10 or 15 years?

I was quoting the overall market, not Umbra’s sales. Joe Morrison [Umbra VP of Commercial Experience] and I were going back and forth on what the actual overall market breakdown of commercial customers actually is. My belief is that it’s like 75% defense/government, but I don’t actually know.

The largest customer is the US government because the data cost is so high. The next biggest is foreign governments, obviously allied governments. Then you have oil and gas, which is gigantic. That industry is the size of many countries. Then you have insurance, and then you have things like critical infrastructure. But to get below that, there's essentially no customer that has any money. So the real key is the barrier to entry.

I mean, we took Airbus's public pricing, which is $7,500 for a one-meter image, and we said we’d sell a 50 centimeter resolution image for $750. So a hundred percent improvement in quality at one-tenth the cost. And guess what. It’s still too fucking expensive. We need to be 90% lower than our current price in order to start making the commercial market bigger.

How do you know? How do you know you won’t get there and it’ll still be too expensive or there really isn’t any demand waiting? Do you have buyers saying, “at this price, I would purchase that data”?

Oh, sure. But we have buyers at any price, right? If you go on and just start selling a 50 centimeter image for $75, you increase demand. You have all these people not in these gigantic industries who are trying to solve a problem where their current best solution is to send a guy in a pickup truck to go look at a bridge to understand the length of a bridge. This costs $200 an hour. It turns out I can buy all the data I need for $75 and then just literally download it from space and run the code that gives me that answer. I just reduced some dude's cost by 90%. That price reduction will not happen overnight. The industry and data providers, including Umbra, need to grow, learn, and evolve.

In Umbra’s distribution, you had strong conviction around two fairly contrarian approaches in the industry. They were 1) Don’t build an analytics platform that competes with your customers, and 2) Provide abundant, free, archival data. Where did that conviction come from?

If you want to have a business that cares about your customers, you can’t compete with your customers and create channel conflicts. Especially when your customers are Google, Amazon, and Oracle who have the best AI/ML in the world. I mean, as a satellite company, you’ve got to solve a super difficult problem launching a satellite into space. And ideally, you want to be the best one doing that. And then to be competitive with Google for AI engineers to then go analyze the data is really expensive. If you do, you’re probably not going to be the best in all those products.

But maybe we’re the dumb ones. Almost every single other company who takes a picture from space, also analyzes it. Almost every single one.  If I’m “wrong” and make money and everyone else is “right” but loses money, I’m fine being wrong.

I would guess they’re doing that because they want to expand their market size, right?

But I'm not trying to grow my market or grow revenue. I'm trying to produce free cash flow. So until you can show me anyone who makes money doing the analytics on their data, I’m not convinced that’s the way to go. You know a liquor store produces more cash flow than every single publicly traded Earth observation company. A liquor store. What are we doing? So people have to change their mindset, and as my wife always says, “just make more money.”

If there's one thing you can emphasize to an individual founding a space startup, what would it be?

So I said this earlier, and I'll say it again because it's important. I would not wish starting a space company on my worst enemy. It is really difficult. But what I would say is, if you have to raise money, understand investors' incentives.

My co-founder, David Langan, is much smarter than me. When we started to talk to investors, he sent me an IRR calculator. I was like “What? Why do we need to calculate their returns over 10 years? We are going to make billions!?” And he was like “No one's going to give us money if they can't actually show a high enough return for their fund.” Which I didn't intrinsically understand until I got older.

And obviously, we've had some crazy years in the stock market. As an anecdote from being an LP in some funds, not a lot of VCs offer huge returns over long periods of time. Oftentimes, returns are lower relative to individual stocks or an index.

So if you talk to me about your business, and the response is that its capital intensity and time to market cannot produce sufficient returns to make investors a great IRR, and guess what, a 10X return over the full fund life is not a good IRR. If you can't produce a product, get it to market quickly, and 100x, you probably should not take venture capital. That said, if you don't have VC, great, build a great company and take as long as you want.

That’s a wrap on my discussion with Gabe! I hope you enjoyed it. Don’t forget to subscribe to SpaceDotBiz or follow me on Twitter for more insights and interviews like these!