PARIS — Capella Space of the United States and Iceye of Finland are venture-capital-funded startups building constellations of radar satellites for government and commercial customers worldwide — assuming there are enough of them to close a business plan.
It’s risky, and while both companies have raised sufficient capital to start launching their constellations, both have also been hit square in the face by launch delays.
So far, it has not proved fatal. Iceye launched a demonstration satellite in January that failed early but yielded enough data to permit the company to perform a thorough redesign in time for the next launch, which has been delayed from November 2017 and is now set for this November.
Capella, which was scheduled to fly on that same delayed flight — a dedicated SpaceX Falcon 9 purchased by Spaceflight Industries — has used the delays to scrap its early design and move to what had been viewed as a second iteration.
The two companies’ chief executives — Payam Banazadeh of Capella and Rafal Modrzewski of Iceye — made no secret of the difficulties of securing customer commitments before operational hardware is in orbit. But that’s what their investors are asking for.
Addressing the World Space Business Week conference organized by Euroconsult, the two CEOs compared and contrasted their approaches to a market that in many respects they still need to create.
Here are excerpts from their remarks:
At what point do investors say they need to have customer commitments?
Banazadeh: They are already asking. It’s been difficult to get commitments without having anything [in orbit]. These are contingent commitments for a service that will come in six months or 12 months, without any money down.
In return, I am willing to give [these early customers] something they wouldn’t get 12 months from now. But [investors] are looking for customers now.
Back in the Skybox and Planet days [two optical Earth observation satellite constellations since merged into Planet], when they were starting the companies, it was easier. You didn’t need to do as much on the commitment front. LoIs [letters of interest] and phone calls might have been sufficient. That’s not the case anymore.
Modrzewski: Investors always ask about [customer] commitments. They ask before you launch, they ask before you’ve raised the funds, and they ask after. It’s a question of how many you have and how convincing you’ve got to be in order for them to actually believe in it. I agree: It’s really hard to convince any customer to give you money before you actually have something to show. They want to see more, which is fair.
There is a long way to go before we will be able to provide the ultimate product we are advertising, which is a three-hour revisit time, reliably, over any point around the globe. It’s not going to be until 2020. Right now we are telling customers: Why don’t we work out an agreement where the moment we launch you get the basic data? You tell us what works, what doesn’t.
We launch something new, you tell us what works. We make those iterations and as long as customers understand the data-relevant approach, and the fact that we are building up a constellation, we are fine and they are fine.
You’ve both confronted launch delays and have used them to redesign your hardware. How do you do that when you can’t know if a delay is a month, or three months, or more?
Banazadeh: It’s called aerospace power and we played it really well. Honestly, we got lucky. We were supposed to launch in November last year and we had the satellite ready to go. It got delayed to February, then April. Then it was June, then August, and now it’s in November.
At some point we made a decision: This is getting delayed, what if we move to the next iteration of our satellite?
The first bird was going to have a 3-meter-square antenna. We went from that to an 8m2 antenna and changed a bunch of things. We added propulsion. It was almost a complete redesign. If the launch hadn’t been delayed from April to June and then August, it would have been tricky. Now we’re launching the nexts iteration of our satellite and the first iteration is going to be in our lobby.
Modrzewski: It would have been really difficult for us not to have that first launch. Our approach is to assume every launch is going to be delayed and we don’t know how much. We buy multiple [launch-service] providers and then wait for the one that’s actually on time. We make the satellite fit into all of them.
Our first satellite, Iceye-X1, was key in terms of our learning curve. There has been a massive amount of lessons learned. We identified multiple problems in this first satellite. Many customers gave us good pointers and we turned around the entire redesign in less than 10 months.
Let’s assume the SpaceX/Spaceflight launch goes off as now planned, in November. What are your deployment schedules after that?
Banazadeh: Our first satellite, Denali, is going up in November. It’s purely a prototype. We are expecting lots of problems and we’ll be debugging the hell out of it. Our second satellite, Sequoia, is going up in June 2019. That is still a test satellite, but we’ll start delivering imagery from it and do demos. Sequoia is 90% identical to the constellation.
Our third satellite, called Whitney, is going up in August .
Our goal is that by Q1 2020 we have six satellites — three satellites each in two orbital planes. That gets us a maximum revisit of six hours.
By the end of 2020, we’re going to add another six. That will drop [revisit time intervals] to three hours globally.
We have set up our manufacturing process to scale up really quickly. Our satellites are 37 kilograms. The first generation is 1-meter resolution, the second is half a meter and we have about 10 minutes of imaging per satellite per orbit.
Modrzewski: With the financing we have received we have booked four additional launches — one in November, another in December, then February and April. Our goals is no fewer than at least six within 2019. We aim for eight, depending on how we perform and how quickly we can turn around consecutive iterations, so we’ll see if it’s six or eight.
These are the ones we’ve financed, so we can confidently promise this. The goal is to get down to hourly revisit rates. To get down to three hours, we need 18 satellites. Depending on what the market says about revisit rate, we will either grow the constellation further or keep it at that.
Banazadeh: We have realized that even though you have Rocket Lab and Virgin Orbit coming on line in 2019, realistically it’s going to be difficult to build a constellation to provide hourly monitoring even if your satellites are in the 100-kilogram range. If you’re 100 kilograms, you put on on a Rocket Lab and it requires 36 rockets to build a constellation of 36 satellites in the different orbital planes to get realistically to a maximum revisit of one hour. It’s a challenge that most companies, across all different sensor types, don’t really talk about, and that most investors don’t pick up on.
So we needed to fit three of our satellites on a single [Rocket Lab] Electron. That’s where we got our 37-kilogram number, because the goal is a constellation of 36 satellites in 12 orbital planes around the globe. Now we can do that with 12 launches. It’s a phased approach that allows us to scale up pretty quickly.
Government: THE customer or a customer among many?
Banazadeh: We are a commercial company with commercial aspirations, but the reality is the market today for SAR [synthetic-aperture radar imagery] is government. It’s defense and intelligence, so we’re tackling that heavily.
Being the first and really only American commercial SAR positions us really well to tap into the U.S. government market, and we’re after that. We’re after international governments and the commercial market, but we realize it’s going to take much longer because the market doesn’t really exist and creating new markets will take longer.
On the commercial side, we’re seeing traction for applications where frequency and reliability matters — a lot of monitoring applications. Infrastructure monitoring, pipeline monitoring, a lot of applications where customers are using helicopters or manually send out crews to the middle of nowhere to take a look.
Modrzewski: We started with the same customers already using SAR and we are trying to be complementary [to larger radar satellites]. We are looking at governments and big corporations. Ultimately Iceye’s idea is to spread global-change detection capability and allow people to buy it is a subscription.
It’s interesting that it’s not mentioned very often that the commercial market isn’t as mature as we all think it is. The ultimate destiny of a company like ours is to focus on the existing government market. It’s very challenging for anyone to finance a company that relies on the government as the main market. It probably isn’t the best thing from our perspective, but it actually is true: The low-hanging fruit, the existing market, is the government market. We see how Capella fits better into the U.S. market and really they have an edge there.
Banaadeh: I don’t see much competition between us [and Iceye]. We’re aiming at different parts of the market with different approaches.
The real disruption I am worried about are in-situ sensors that will be deployed in the millions, if not more, in the next decade. At the end of the day, the products are not satellites. We’re competing against anyone who can provide any information to all customers globally. If someone can do that cheaper on the ground with these little sensors everywhere such that you have the status of everything moving and you can see changes — that could disrupt my business.