Closing Bell - Manifest Space: Infrastructure as a Service with Loft Orbital COO Alex Greenberg 1/11/24

Episode Date: January 11, 2024

What if you could get access to space to run applications or collect data without building your own hardware or launching a spacecraft to orbit? Loft Orbital is setting out to provide space infrastructure as a service, running some of its satellites currently in orbit and aiming to conduct “virtual missions.” Loft Orbital Co-Founder & COO Alex Greenberg joins Morgan Brennan to discuss the company’s business model, computing services from space, and its plug-and-play approach to space.

What if you could get access to space to run applications or collect data without having to build your own hardware or launch a spacecraft to orbit? What if a company could do it for you and offer their space infrastructure as a service? Well, that's what startup Loft Orbital is setting out to do with its virtual missions. It's already running some on its satellites currently in orbit. But the first spacecraft completely dedicated to this business model, EM6, will launch via SpaceX's Transporter 10 mission later this quarter. EM6 has a couple of cameras on it. It has some RF sensors. It has an inter-satellite link, so it has real-time connectivity to the ground.

And it has a bunch of GPUs and other compute systems that allow us to run customer software apps on board. So our business model for EM6 is almost to treat it like a bunch of GPUs and other compute systems that allow us to run customer software apps on board. So our business model for EM6 is almost to treat it like a data center. We're essentially renting out capacity of both the sensors and the compute on board to any customer who has any application they want to run. Think computing capability for AI workloads. Alex Greenberg, the startup's co-founder and COO, says Microsoft is both a customer and a partner on EM6, and others have signed on as well. Loft is essentially selling

computing sessions from space, meaning the satellite gets utilized almost 24-7 and can generate revenue around the clock. How is Loft making this possible? Well, it buys standard satellite buses from vendors like Airbus and outfits them with hardware and software, quote-unquote, abstraction layers that enables a plug-and-play approach to any sensor capability that's needed. Basically, Loft Orbital is and has been commoditizing the satellite payload process.

And now it's taking the next step to offer access virtually. I'm Morgan Brennan, and this is Manifest Space. Let's just start at the beginning. Tell me a little bit about Loft. Absolutely. So I think the best way to think of Loft is we are a space infrastructure company. That means that we operate satellites and we fly our customers' missions as a service. So a customer mission can be a physical sensor or payload. So for example, a hyperspectral camera that we fly for NASA, but it can also be software. So AI software, machine learning software that we fly for a customer who wants to run some kind of edge compute workload on board.

So we do both of those things. Company is about 200 people. We're headquartered in San Francisco, but we have a big presence in Golden, Colorado, so right outside of Denver, and then Toulouse, France as well. That's actually our largest office. And in terms of traction, company is doing a smidge under $100 million in revenue. We've been around for seven years, still growing quickly, and launching a lot of satellites next year. So that's what we're focused on right now. a smidge under $100 million in revenue. We've been around for seven years, still growing quickly and

launching a lot of satellites next year. So that's what we're focused on right now. And I want to get into all of that. But first, just one more question on this, because it is fascinating, this idea of basically a mission as a service. How did this idea, you're the co-founder, how did this idea come into being in the first place? Yeah. So I've been in the space startup game for a very long time. I was at Spire, which was a venture-backed space company. Now it's a public company. I was there in 2014. And that's where I met my co-founders, Pierre and Antoine. So we were working on the product and business development teams there. And we saw both from

our experience there, as well as just being in the industry, how challenging it was for companies or government agencies to get satellites in space to collect some kind of data or provide some kind of space-based service. So we saw the supply side of the industry being pretty broken, but we also saw the demand side really starting to pick up. We saw so much more activity happening at LEO, so many more customers relying on satellite data or services from space for critical applications. So we saw like massive demand, but really broken supply. And we thought, hey, can we fix this? And really the way we're going about that is by just making the customers life easier so they can focus on their specific payload, their specific mission without having to think about how do I build a satellite? How do I integrate a satellite, get it launched and operate it for a number of years? That's all really challenging,

really capital and resource intensive. And so we are taking all that away from the customer and offering it as a very simplified service that they can just simply buy. It sounds great. How are you able to do it, given the fact that it does tend to take time and it does tend to be capital intensive? Yeah. So, you know, we're really we had that that original question from day one. And so our approach was always to be take a very different tack than what we would call a legacy aerospace company. So look at any traditional aerospace company. They're all phenomenal engineering organizations, but they are not really product guys. Usually their strategy is to take customer requirements and then build the satellite as a one-off.

So most satellites that are in space today are a unique one-off design that are purpose-built for a specific mission or a specific application. Loft kind of looked at that model and said, that's really difficult to scale. It's hard to get economies of scale when you're building something always as a copy of one or as a single design. So what we're doing is essentially standardizing and productizing the satellite. So we actually don't build what's called the satellite bus or the chassis of the satellite ourselves. We work with partners who are already mass manufacturing them for other applications. So in our case, we've worked with four different bus providers, but primarily we're working with Leo Stella, who's building the Black Sky Constellation and Airbus OneWeb Satellites, who's obviously building the OneWeb Constellation. So our approach is take those buses, buy them off

the manufacturing line, don't really make too many modifications to them and just treat them as commodities. So for us, that's really half the battle. What we build and what our product is, are essentially hardware and software abstraction layers that go on top of that bus that allow us to plug and play any mission, any payload, any sensor capability on board without having to modify the bus or modify any of the onboard hardware. So what we're essentially doing is productizing this entire platform and turning it into something that is very modular and very turnkey. Now, once the satellite's in space, we need a way to operate it, right? You need usually a large team and a lot of people

and a lot of custom software to command and control and operate a satellite. And we've handled that problem as well. So the second side of our whole product set is something called Cockpit, which is a system that allows customers to request things that they want their mission to do on the satellite, and Cockpit is the automation platform that takes those requests

and actually turns them into taskings that are sent up to the satellite and executed on board. So we're making it possible for even a non-technical person to very easily operate a satellite. And so our customers don't have their own antenna networks. They don't have their own ground stations.

They don't have their own satellite operations team. They just have the software that makes it super simple for them. It's very cool. You have several satellites in space. You're going to be launching more as well as we look at 2024. You talked a little bit about it, but the virtual aspect of what you're doing and what that's going to look like, especially with the launch of your next satellite. Yeah, absolutely. This is something we're really excited about. So

going back in time a bit, when we started the company, our mission was really just make it simple for any customer to launch a payload in space. And payloads are cameras, sensors, radios, you know, anything physical that you would put on a satellite in order to collect some kind of data. What we saw, though, over time, and it's really kind of come up, especially in the last 18 months, is lots of customers not just wanting to fly their own physical payloads, but customers who want to apply software applications on board. And the reason they want to do that is because they want to process and analyze data as soon as it's collected to derive some kind of insight from that data in real time. So they want to understand where a wildfire is right now, or how many troops are in a certain area, or what the effect of an extreme weather event is right now, or how many roads are available in a certain area. These are questions that need to be answered within minutes, seconds even. And if you have to wait hours or days to downlink raw data, get it into a data center, get it in front of an analyst, and then have a satellite operations engineer retask the satellite with new instructions,

you're going to basically have lost whatever insight you needed. So what we're really trying to provide is the ability to allow for time-sensitive understanding of what's happening on the ground. And the way we do that is by allowing folks to upload algorithms to the satellite, process and analyze data on board using our own edge compute capabilities, and just make that really, really easy for them. That's wild. So that's already happening with some of the satellites that are on orbit right now. Exactly. So we have a couple examples of that. We just announced one of our customers, a company called Agenium, who is performing a mission where they uploaded a neural net

to one of our existing satellites, YAM-3. It's actually been in space for two years. So we're still finding new ways to use it and new ways to monetize it. And that customer was able to basically detect dark chips onboard the satellite using their own AI application. So that was really cool. But what we're really excited about is YAM-6, which is our next satellite launching in March of next year on

a SpaceX launch. And that is the first loft satellite that is fully dedicated to what we call virtual missions. So this idea of flying software payloads in space. And so YAM6 has a couple of cameras on it. It has some RF sensors. It has an inter-satellite link. So it has real-time connectivity to the ground. And it has a bunch of GPUs and other compute systems that allow us to run customer software apps on board. So our business model for EM6 is almost to treat it like a data center. We're essentially renting out capacity of both the sensors and the compute on board to any customer who has any application they want to run. We're partnered with Microsoft here. So Microsoft is both a customer and a partner on this mission. We've got a few other customers that have signed and we actually just kind of

opened it up for commercial use a couple weeks ago and have already signed a few new customers. So we're really excited about the capability that Yam6 offers to the industry. I think it's the first time where anyone is offering their satellite as a deployment platform for customers to run their own software applications on board. That's not something you usually see in the space industry. No, it's really fascinating and I have a lot of questions for you. But first, how many customers do you have? How many customers can one satellite support? Yeah, so the way we kind of allocate resources on our satellite is essentially by more of a, you know, kind of demand based way. So sometimes we'll take a satellite and fly one customer on board if their payload is taking up all of the available mass, volume, power, budget, etc. We fill that that mission with or that satellite with one customer and the customer pays accordingly.

But we also have satellites that are more what we call ride shares, where we're taking a number of different payloads from different customers and then fitting them on board the satellite. So, for example, EM3 and 5, I think between them have something like 11 active customers. We've got another satellite, EM8, that's launching later next year that has four customers on board. But we also have some constellations where there is a single dedicated customer. So we actually have two, 10 satellite constellations. One is flying next year. The other is starting to fly in 2025. One's for a commercial company. One is for the Space Development Agency. And those are dedicated to those single customers. So we're really just selling the space on board or selling the capacity on board and one customer can buy all of it, or we can divvy that up among

many, many customers. Okay. So with virtual missions, then I would imagine you can have many, many more customers than you could, you know, if you were having to support a customer's hardware. Exactly. With virtual missions, kind of the atomic unit we're selling are sessions or kind of time of compute resources that we're selling on board. So there's a lot of customers you can fit on there. And what's exciting for us from a business model perspective is the monetization potential is much higher because we can utilize the satellite almost 24-7 and generate revenue from it at all times. Interesting. The fact that Microsoft is a partner and a customer on this, I mean, we're having this conversation in a year where generative

AI, for example, has just exploded into the public sphere of discussion and into the business community for new applications. How much is that enabling your ability to do this with these virtual missions, if at all? Well, so first off, I think Microsoft and Loft share a vision that AI is going to operate satellites one day, whether that's actually tasking the satellite, but also making intelligent decisions on board the satellite based on information that is observed by the satellite or data that's observed by the satellite. So that's the world we're trying to bring about. And there's really two things you need to make that happen. And Loft as an infrastructure company is trying to provide those two things. The first is actual compute hardware on board. And then the second is making it really easy for

any entity, any customer to deploy their own software to run on board that satellite. That is not a trivial problem today. It's actually quite difficult. So that is actually what we're partnered with Microsoft on, is a software development kit and a framework that allows any developer in their own cloud environment to write a software app and click a button and then deploy it to EM6 and eventually other satellites and run that onboard. So that's really kind of the nature of our partnership with Microsoft. They've also purchased some of the onboard capacity of EM6 for their own applications as well. So they're also a customer.

You talked about the fact that one of your customers has been able to use your technology and one of your satellites for dark ships, for example. I mean, are you seeing, especially at a time where we are focused, for example, on the shipping lanes in the Red Sea and geopolitical risks and this need to be able to monitor the world in a much greater capacity, are you seeing demand for that type of product right now? Is that something that you have multiple customers looking to acquire that capacity from you? It's funny you ask because we just announced

another customer, a Japanese company called Space Compass, which is doing something similar with ship detection. So again, Loft is really just an infrastructure company. We are a platform in space that makes it easy for any entity to observe what is going on and with virtual missions actually derive insights and understanding of what that data is actually saying. So we're a platform for observing a lot of geopolitical events happening in the world. It's unfortunate kind of how some of those things are evolving, but that's certainly driven more demand for not only data from space, but actual real packaged information from space. And that is what we are trying to provide. How much of the demand you're seeing or how much of the customer onboarding are you seeing is

coming from the commercial space versus the government space, whether it is NASA or whether it is on the defense side? It's definitely divided in a good way. So in our business, we're about 60% commercial today, 40% government. I see that saying roughly in that range, maybe 50-50 at some point. Obviously, a lot of investment is coming from government, but I think there's also a lot of commercial companies who are coming up to serve both government needs and commercial needs on the analytics side. So from our perspective, there's a lot of demand from government, but a lot of demand from commercial entities as well. Okay. Just in terms of this whole idea of mission as a service, how much time are you cutting down on this idea of commoditizing this whole process? How quickly then from start to finish? I realize

every situation is going to have its own special details to it, I'd imagine, or customizations, but how quickly from start to finish can you actually get a satellite, you know, from development to on orbit and operating? Yeah, that's a great question. Something we're quite excited about. So going back to what I said earlier, most legacy aerospace companies think of a satellite and design satellites as one-offs. So first, what always happens when a satellite mission kicks off is the satellite is designed. It's usually a co-engineering process between the customer and the satellite manufacturer. And only once that design is fully validated can hardware actually be procured. And then because you're buying something as a

one-off where everything is custom, it's probably not going to work the first time you plug everything in and integrate everything together. So there's a long period of integration and test and troubleshooting and debugging. And so for that reason, it can take five years or more for an ambitious satellite to get designed and get launched. In Loft's world, we do things very differently. Again, because we're effectively commoditizing and productizing the hardware, the core hardware, so all the core Lego blocks or the core pieces of our platform are the same across every single satellite that we fly.

The buses are standardized and common. Our hub, which is kind of our universal payload adapter, that is common as well. So we're actually able to go and procure all of those pieces well in advance of when we actually need them, such that they're available for the customer off the shelf. So when a customer signs with us, all of the hardware that's going to fly on their mission is either truly sitting on a shelf or has been ordered over a year ago and is well on its way. And the reason we do that is because no matter whether

you're fully productized or whether you're fully custom, the supply chain and space is always a challenge. So we're always trying to figure out ways to essentially work with the supply chain where things are always going to take 12 to 24 months to be delivered. So from our perspective, you know, for a physical mission, we can get a customer a space in six months after they show up with their payload, if not less. It really depends on how long it takes for their payload to get delivered to us. And then for a virtual mission, a software payload, it can be hours. Eventually, we just kind of expose the API to the customer and they deploy their application. Right now, it's taking a little bit longer because there's a little bit more handholding, but our vision is that that becomes

fully self-serve at some point. So how quickly then is Loft continuing to grow? And just as importantly, what is the vision for the company? Because I know you have investors and you've done funding rounds. Is that something you'll continue to do? Is there a plan to go public at some point or not even thinking about it yet? Yeah, so on the growth side, Loft is still growing considerably. And I think we definitely benefit from being able to have that growth happen in the private market.

So just to throw some numbers I mentioned earlier, revenue is approaching $100 million. We grew almost 150% last year. We're expecting growth on that scale next year. So really continuing to grow fast, which is really exciting. We have great investors who are really supportive, who are long-term minded. We do plan on, of course, delivering liquidity to them at some point. Right now, we are really focused on scaling the company from a cadence of doing

one to two satellites a year to 10 plus. And so we're investing in production. We're investing in the ability to execute missions on a recurring basis. We're investing in our team. And so right now, in that kind of heavy investment period, we do want to stay private. We will probably raise one more round of capital. But then after that, we'll be looking to exist either as a standalone public company or some other outcome. But that's really a decision for the future. Right now, we're just focused on building. Okay. So what does that mean for 2024? Milestones to be watching and beyond? Yeah. I think first off, I mentioned it, but we want to launch EM6 and really offer virtual

missions truly for the first time to the space industry. The second thing is we have a busy manifest of a bunch of other satellites. So mentioned it, but launching over 10 satellites next year. We're celebrating the one year anniversary of EM5, which launched last January. And that was the only satellite we launched this year. So we're going from a cadence of about one a year to 10 plus a year. And there's a lot of work behind the scenes to make all of that happen in growing and scaling.

And so the fruits of that labor are all going to be on exhibit in 2024. Final question for you, who are your competitors and what is your outlook for this market looking out 2024 and beyond? Yeah, I think we've seen a lot of companies float around the term satellite as a service, mission as a service. It's almost become a bit of a platitude at this point. We don't see any other companies directly kind of implementing that business model in the way that we are. That being said, though, there are quite a few satellite manufacturers that I admire a lot, which are incredible companies. So I think, you know, the folks who are building small satellites,

some of the names that are on some of the SDA programs, whether they're big primes or they're small specialists, those are the people that we would view as our competition. Got it. Appreciate the time. Thank you so much for joining me. Yeah, thank you. Have a good one. That does it for this episode of Manifest Space. Make sure you never miss a launch by following us wherever you get your podcasts

and by watching our coverage on Closing Bell Overtime. I'm Morgan Brennan.