Philip Johnston - Founder of StarCloud (YC S24) | The Space Computing Era Has Begun

Shamil Malachiyev: Hello everybody. This is going to be a really cool episode of The Founder's Code podcast, because with us we don't have a usual type of startup. We have a guy who is literally aiming for space, the final frontier. Phil, welcome to the podcast.

Philip Johnston: Thanks so much for having me, Shamil. Excited to be here.

Shamil Malachiyev: Let us get a good grasp on you as a person. Did you grow up in the UK?

Philip Johnston: Yes. I was born in the UK, and then I moved to South Africa until I was about five; my dad was a diplomat there, so I went with my family. Then we moved back to the UK, where I stayed until I was 21, and then I moved to the US for grad school. Since then I've basically been outside of the UK.

Shamil Malachiyev: Do you remember any particular aspects of your upbringing, your childhood, that pointed towards this huge ambition of colonizing and getting things into space?

Philip Johnston: One fairly significant aspect: I'm the youngest of five boys, and I have an identical twin brother, so I'm quite competitive with my twin brother. He's also founded a Y Combinator company. They were doing energy tech; they've pivoted now, but they were doing energy tech.

Shamil Malachiyev: Are you serious? Do you compete with him on that as well? Is that why you chose something that's borderline impossible, something that seems so complex until you actually get into the whole rocket-launch space?

Philip Johnston: A quite impactful thing on that front for me was an essay written by Sam Altman called Hard Startups, which came out maybe five years ago. He's talking about the difference between doing a hard startup and an easy startup. In a hard startup, everything except making the tech work is easier. Getting investors is easier, getting technical talent to want to join you in the mission is easier, getting reporters to want to write about you is easier. The main thing that's different is making the technology work. And if you're going to do any startup, you might as well go big or go home, as they say.

Shamil Malachiyev: In your earlier career you had internships in VC companies, Goldman Sachs I think was one of your earlier ones, McKinsey. Did you ever consider a standard career path for yourself as a consultant, or did you have this startup vision?

Philip Johnston: I would say I had a fairly standard career path. I spent five years on the engineering side as a quant trader at a bank, and then I spent two years at McKinsey, all of which is very standard. I think I knew for a long time that I wanted to found a company, though, to be a founder in general. So it started with a different company, an e-commerce company; I'd say that's more of a traditional type of startup. And then this one is obviously much more of a deep-tech situation.

Shamil Malachiyev: Can you tell me a little bit about your first experience with the e-commerce business that you ran? How did you get to "I want to start a company, and this is what I'm going to be working on for the next three years"?

Philip Johnston: That one was quite opportunistic, to be honest. There was a business model working extremely well in the US called e-commerce aggregation, which is basically where you buy and grow e-commerce brands. It was kind of a private-equity roll-up, but for e-commerce brands, and there were lots of small sellers who'd come up over the previous five years because of things like Amazon Marketplace and all of these large platform marketplaces. I saw that it was working extremely well in the US and Europe and nobody was doing it in the Middle East, and I thought there was a very short window of opportunity to make that work. So I left McKinsey to found an e-commerce aggregator in the Middle East, doing what was called at the time the Thrasio model.

Shamil Malachiyev: How was the journey? Was the start hard, or was it easier than you expected?

Philip Johnston: It was a wild ride. We hired 100 people in the first year, which in hindsight was probably not a very smart thing to do.

Shamil Malachiyev: Were you good at hiring?

Philip Johnston: Nobody's good at hiring, especially not new first-time founders. If you're hiring a hundred people in the first year, you're probably going to make some mistakes. The reason was that it was a land grab. We were trying to buy as much as we could as quickly as possible, and we thought there were going to be others coming in, and there were. That's why that model grew so quickly.

And then what happened, essentially, was interest rates went from zero to 5%. We were using venture debt leverage to make acquisitions, and the idea is that the cash the brands spin off can pay back the debt. The only issue with that, and this is why Thrasio and all the big US ones went bankrupt, is that interest rates went from zero to 5% over about a six-month period after the invasion of Ukraine, because of inflation and other things. What that meant is that even if you built a profitable business, EBITDA positive, which we were not, you still needed to refinance the debt, because the debt had a time to come due. We were expecting to refinance the debt at lower interest rates, and as interest rates went through the roof it became very hard to refinance. Then almost all of them had to sell in order to pay back that debt, which is basically what happened with us as well.

Shamil Malachiyev: And how did you manage through that? Did you have to close down the company, or was it acquired?

Philip Johnston: No, we didn't close down the company. We got acquired. The entity still exists; a Turkish competitor bought it, basically.

Shamil Malachiyev: How long after the acquisition did you decide you wanted something new to work on?

Philip Johnston: I was actually out about six months before the acquisition. I give all credit to my co-founder at the time, a guy called Manfred, for riding that ship out, because I can imagine that was not an easy period for him. My main role in that business was acquisitions, fundraising and finance, and once we realized we were not going to be doing any more acquisitions, because we were not going to be getting any more financing, it came time for me to move on.

Shamil Malachiyev: With StarCloud you got into YC. How big of an event was that for you?

Philip Johnston: Honestly, it was huge. I know you're not supposed to say that. You're supposed to say we were building either way, and if we got in or didn't get in, it wouldn't make a difference. That is not true.

Shamil Malachiyev: Let's be real here.

Philip Johnston: It did make an enormous difference in the trajectory of the business. We got rejected twice before, so we got in on the third time. I've built a company without YC, so I think I have a reasonably good perspective of what it looks like to build a company with YC and without YC. It is a lot easier to build a company with YC. I'll put it that way.

Shamil Malachiyev: Can you tell me more details?

Philip Johnston: You have the batch, and that's quite useful. The best thing about the batch is the other batchmates. We went through with a small group of about eight deep-tech companies, so you're facing broadly similar challenges around how you're going to survive all this heavy capex without revenue. That's probably the most tangible benefit.

Then in terms of the intangibles, it's a very confidence-boosting signal, both to you and to the market, that what you're doing is at the edge. YC gets like 50,000 applicants a year and they pick 100 or 200. And what that shows to investors is: you can believe what you want about the idea, but if there was any better team working on this or on something similar, that team would have gotten into YC instead of these guys. What it tells the market is that you have the most kick-ass engineering team in the world working on this.

Shamil Malachiyev: And how do you keep your confidence levels up? Were you always a confident person, as a kid, growing up?

Philip Johnston: Confidence is often the wrong word. If you ask me if I think this business is going to have a successful end outcome, I would say it wouldn't surprise me if the objective odds are less than 50%. So being delusionally confident is not the right frame of mind, necessarily. What matters more is the willingness to take risks that others are not willing to take, where the expected outcome is very high. If I had to give the outcome-weighted probability of what we're doing: say there's a 10% chance of what we're doing working, but the upside if it works is 1000x what we are taking from investors. That's a great thing for people to invest in, and it's a great thing for me to spend my time on. But that still means there's a reasonable chance that what you're doing is not going to work, and I think it's good to be realistic about that. You don't want to be completely deluded.

Shamil Malachiyev: But if we look at it, almost 95% of startups fail, and by fail I mean not getting to the 10x, 100x level. So you might as well be doing something completely crazy that has the highest outcome potential.

Philip Johnston: You want to do something that at the very beginning seems batshit crazy, but very quickly does not seem that batshit crazy. Because if it carries on seeming batshit crazy for a long time, then that is also not good.

Shamil Malachiyev: Did you have any early role models, inspirations? Maybe Elon Musk, with what he's doing in space, the South African background...

Philip Johnston: Yeah, definitely. My favorite biography is definitely Elon's, the one by Walter Isaacson. He's definitely my biggest inspiration in all this. Even that framing of when he started both Tesla and SpaceX: he thought there was a more than 90% chance of failure, but that's not necessarily a reason not to start something.

Shamil Malachiyev: I have it right over there on the shelf as well. Do you see yourself in that? Would you be able to make those crazy bets when there's a 90% chance of failure, multiple times, and just keep going?

Philip Johnston: It helps if you make the bet and it works. I will die trying to make this work.

Shamil Malachiyev: Is it really the perseverance?

Philip Johnston: Yeah, I think so, for sure. If you're going to do one of these crazy things, you have to be in with every inch of your body. And that is definitely the case for me.

Shamil Malachiyev: Do you ever get imposter syndrome? Waking up and thinking, what the hell am I even doing with all this? I could have been doing so much simpler stuff, all these AI startups...

Philip Johnston: I probably should more than I do, because I'm also not a space engineer. My co-founders, the chief engineer and the CTO, are the two most kick-ass space engineers in the world, so I don't feel too stressed about it. Maybe I used to get it. It's not something I think about too much right now. What I think about now is: how the hell can we make this thing work? Let's get our heads down and make this thing work. Imposter syndrome or not is irrelevant. I've accepted what the risks are; I think I know what the risks are. And if we fail, I'm not going to beat myself up over it. I'm not going to feel like I was the wrong person to try to do it. I will beat myself up if we fail because we didn't try hard enough, for sure. But not for any other reason.

Shamil Malachiyev: So would you say one of your biggest fears would be to have regrets? On one of your last days thinking, I could have tried harder, and it could have given all of my engagements higher chances of working?

Philip Johnston: Kind of, except I just know that won't be the case. If this company fails, it will not be because we didn't try hard enough. That I can guarantee you. Every time I've done something where I've ended up being quite proud of it, I've gone into it with that frame of mind, where I am thinking: if this thing doesn't work out, and sometimes it doesn't work out, by the way, it will not be because I didn't try hard enough.

Shamil Malachiyev: Are you going to beat yourself up if it doesn't work out? Or are you just going to work on something close to that vision, maybe in a different domain?

Philip Johnston: I don't spend too much time thinking about what's going to happen if it doesn't work out. I'm just trying to make it work.

Shamil Malachiyev: I think that's what investors mostly want to hear.

Philip Johnston: I mean, it's true. I'm not saying it for investor purposes. That is literally true.

Shamil Malachiyev: For the people who don't know what StarCloud does, can you give a little bit of information, in a basic, non-engineering kind of way?

Philip Johnston: Sure. We are building data centers in space. That is basic non-engineering speak. We're building data centers in space initially to provide compute to other satellites, and then later to address the energy challenges coming down the line from AI. In the end state, we're planning on building very large data centers, multi-gigawatt, with 4 km by 4 km solar panels. And that will be the only thing that allows us to meet the coming demand for energy for compute and AI.

Shamil Malachiyev: That's your elevator pitch. So that's 16 square kilometers of solar panels up in space.

Philip Johnston: Yes.

Shamil Malachiyev: I just love how in your head it's like, step by step, this is what we're going to build. For anyone else to just even come up with that...

Philip Johnston: We're not going to stop there, by the way. It's going to be a lot more than that. When we decided what we were going to show in the concept video, we said, let's do something that's roughly ten times the biggest data center on Earth right now, just to get people thinking. So we went for a five-gigawatt data center. They probably won't all be located in the same place, but I expect there'll be much more than five gigawatts in space in the end state.

Shamil Malachiyev: Just to put the problem into perspective: we have AI being used more and more, causing a lot more energy demand, and we might be faced with the consequences of that, because in order to cool off all of the processing heat that gets generated, we use water. What do you think the energy demand for AI is going to look like over the next 5, 10, 20 years?

Philip Johnston: It looks like we're going to need to double. In North America in particular we'll need to double in the next three to five years, which is at least another 50 or 60 gigawatts. And then I would imagine we'll 4x in the five years after that, and 8x or 16x in the five years after that. Demand for AI is growing at an insane exponential right now.

Shamil Malachiyev: And how easy would that be to do terrestrially?

Philip Johnston: Not easy at all. One gigawatt is one nuclear power plant. Are you going to build 50 new nuclear power plants in North America in three years? No, we're certainly not. So what are you going to do? We are very, very rapidly running out of sources of this type of massive amounts of new energy.

Shamil Malachiyev: What's the biggest stopper you have at the moment from realizing your mission?

Philip Johnston: There are two big technical challenges we're overcoming, and half the engineering team is working on one and half on the other. The first is making the chips work in a high-radiation environment. We're launching the first satellite in either next month or the month after, just waiting on confirmation. That's going to have an H100 from Nvidia on it; it'll be about a hundred times more powerful GPU compute than has been in space before. Making that work in a high-radiation environment is not trivial, and there's a bunch of both hardware and software to make it work; we're doing a lot of engineering on both.

The second is: how do you dissipate all that heat in a vacuum? You basically need very large, low-cost, low-mass deployable radiators, so the other half of the team is working on building those very large radiators. Those are the two technical challenges. And for us to be cost-competitive with terrestrial data centers, we will need Starship launch costs. We do need Starship to work for us to be cost-competitive with terrestrial.

Shamil Malachiyev: What are the current costs?

Philip Johnston: Energy costs right now terrestrially, if you're building a new energy project: you're looking at gas turbines, which is what Stargate's using, for example, at maybe $0.10 per kilowatt-hour, and you could get a new energy project down to maybe $0.05 per kilowatt-hour. We're targeting $0.02 per kilowatt-hour, including the launch cost. All-in energy costs.

Shamil Malachiyev: What's the launch cost at the moment?

Philip Johnston: Launch cost is $6,500 a kilo; that's what we're paying for our first launch, and roughly the same for the second launch in October next year. By the way, we can run this business for a few years just offering compute to other satellites, profitably, using that launch price. That's the Falcon 9 launch price. So the company is not in grave peril if the launch cost doesn't come down for the next 10 years; we can just run like that, and we'll build the expertise needed for the later business. For us to be able to compete with terrestrial data centers, we're looking at around a $400-a-kilo launch price. The potential marginal launch cost for Starship is $10 a kilo. Maybe even less than that, maybe $5 a kilo, depending on who you believe. So that's where the potential launch cost is going.

Shamil Malachiyev: So while waiting for the launch cost to go down, you can pretty much be a sustainable business with the existing $6,000-plus per kilogram. It's just that the break-even is going to take a lot longer.

Philip Johnston: Break-even to compete with terrestrial data centers. We can be a profitable business just serving other satellites initially.

Shamil Malachiyev: And how often do you need to replace parts for the equipment that you send into space?

Philip Johnston: The chips are expected to last around four years, which is the same as a terrestrial data center. Basically, we're expecting a certain failure rate, and you get that failure rate roughly in terrestrial data centers as well. It used to be that you'd get dudes going around unplugging chips and taking them to be repaired; now they basically just leave them in there and work around them without them working. That's roughly going to be the same thing. For the first few years we're not going to have access to the data center once it's launched. In order to not have access and have the thing continue running, we need to have redundancy on the critical systems, and then over-provision the things which you expect to degrade over time. For example, with the solar panels and radiators you're expecting a certain percentage failure per year, so you just need to over-provision so that you have enough for the life of the mission.

Shamil Malachiyev: With the radiators, do you have any other companies trying to cut the cost down, or is it something you had to pull off yourselves?

Philip Johnston: Not that I know of. There are lots of companies who make space radiators, but they're extremely finely optimized for mass and for efficiency. We are optimizing for cost. We are not optimizing particularly for mass or for efficiency, although they will be about a hundred times lighter than the International Space Station radiators. We are optimizing for manufacturability.

Shamil Malachiyev: How do you as a CEO manage so many different verticals you have to think through on a daily basis?

Philip Johnston: Well, the team is actually very nicely split. You have those two engineering challenges: one is the radiators, one is the chips in a radiation environment. Adi is the expert in all things chips in a radiation environment, so he manages that. Ezra is the expert in deployable radiators and solar panels, so he manages that and the satellite bus. And then my role is commercial, which is talking to customers, fundraising, and then all of the boring stuff like HR and legal and finance and everything else.

Shamil Malachiyev: What is the main USP or competitive advantage that you talk about when comparing, let's say, orbital compute to terrestrial?

Philip Johnston: In the end state, we'll have a much lower cost of energy. That's the main USP.

Shamil Malachiyev: But what about until we get to that point?

Philip Johnston: Until we get to that point, we'll be serving other satellites, and so we're not competing with terrestrial compute. We are competing with basically nothing. Right now, these large satellite constellation operators essentially only use the satellite for 1% of the time that they could, because they know they can't get the data off the satellite. If they can ship all of that data to us using an optical connection in space, and you can do optical in space, but you can't really do it to the ground, it's a bit of a pain, we can then run all the inference workloads that they want on it, and we just send down the insight.

If they're looking for wildfires, for example: they send us hundreds of gigabytes of raw imagery, we run an inference workload on that to detect if there's a wildfire, and we just send down the location of that wildfire. At the moment, what they do is wait an hour until they get to a ground station, then download as much as they can, which is usually like 1% of the total data they have. It's a nightmare. So the USP is that we offer a service that is not possible today.

Shamil Malachiyev: How long do you think until you get the launch costs down? What is SpaceX telling you?

Philip Johnston: I can't be too specific about the conversations we have with SpaceX because we have NDAs, but I can tell you roughly what they've been saying in public, which is that they hope to fly their own payloads next year and the first commercial payloads either the year after that or in 2028. They're expecting, towards the end of next year, to be deploying their own Starlink satellites with Starship, and then towards the end of 2027 or early 2028 to be deploying customer payloads.

Shamil Malachiyev: Until then, you want to make sure that you're ready with the radiators and solving the radiation in space.

Philip Johnston: Yeah. Until then we're building a business that works with the Falcon 9 launch cost, but we'll be the first to be ready once Starship is ready to go.

Shamil Malachiyev: Do you think space pirates are going to be a big issue for you in the future?

Philip Johnston: I think space pirates will be about as big an issue as regular sea pirates are for commercial shipping right now.

Shamil Malachiyev: So every once in a while.

Philip Johnston: Yeah. There'll be solutions for it. Space Force is actively working on it, and a few companies, like Portal Space Systems, are working on interceptor spacecraft.

Shamil Malachiyev: If you were to compare, in a standard month, your good days to the bad days, when you wake up and today's a good day versus everything's going to shit, what is your percentage?

Philip Johnston: Overwhelmingly more on the positive side. Probably like one in seven. One bad day a week, maybe.

Shamil Malachiyev: And do you do anything to keep yourself motivated, energetic? Gym, yoga, CrossFit, drinking juices?

Philip Johnston: I just went to the gym yesterday for the first time in about three months. I would like to go to the gym every day, but it's just impossible at the moment. But I'm going to be here for a while now, so I hope I can get back into the gym.

Shamil Malachiyev: Are you currently in StarCloud's main engineering base?

Philip Johnston: Yes, this is one of our facilities. We're in Redmond, in Washington. This is where we're going to assemble the deployable solar panels and radiators. Through a wall behind here there's a welding and fabrication shop, and then we have electrical engineering through these walls here.

Shamil Malachiyev: Very cool. To me, talking to you almost feels like getting a prequel to Interstellar, what they were talking about before everything kicked off, when it was all just getting set up. What do you think the world is going to look like in 50 years?

Philip Johnston: Within 10 years, I think almost all new data centers will be being built in space. Within 20 to 30 years, I think that means the entire stock, or like 99% of compute, will be being run in space. And in 50 years, I hope we've started work on the Dyson sphere. Do you know what a Dyson sphere is?

Shamil Malachiyev: Yeah. But can you talk more about it? I can imagine there are a lot of people listening who don't know about the Dyson sphere, the Kardashev scale, trying to capture all possible energy from the star and the solar system.

Philip Johnston: A Dyson sphere, although actually it's much more likely to be a Dyson swarm, is where you surround a sun, a star, with as many solar panels as you can, basically. What I find more interesting, actually, is a Matrioshka brain, which is where you then hook that energy up to as much compute as a star system can physically support. There's an argument to be made that the end state of all intelligent life, or superintelligent life, is to build Matrioshka brains. My favorite bogus theory about this is that dark matter in our galaxy is actually stars that have become Matrioshka brains and therefore are not emitting anything. It could be that 80% of the stars in our galaxy are already Matrioshka brains. 80% of the mass in our galaxy is dark matter, and it's unevenly distributed throughout the galaxy.

Shamil Malachiyev: Because we can see that they have gravitational force in the field.

Philip Johnston: Yeah, you can see the gravitational force, but there's no point of light coming from it, or infrared, or anything else. And that is exactly the signature you'd expect from a Matrioshka brain. If you look too deeply into it, you realize there are a few flaws in that. The way it's spread out is a bit too smooth, it's not pointy enough. But I still like the idea.

Shamil Malachiyev: So you like to think that there are other conscious beings out there.

Philip Johnston: There must be, and it's completely impossible that there isn't. They might not exist right now. Well, unless there are Matrioshka brains, but my suspicion is there actually aren't Matrioshka brains in our galaxy. Because if there were, they would have colonized the galaxy. It would really not take very long to colonize the galaxy, maybe a few million years, and we would see life everywhere if they had done that. So my expectation is that we're not the first life in the galaxy, and we won't be the last. But it's probably the case that the Fermi paradox implies that superintelligent life, at any rate, is very short-lived. And that is quite worrying. But it's also probably true.

Shamil Malachiyev: Tell me about it.

Philip Johnston: The Fermi paradox is roughly this: you could get to the nearest star in about 10,000 years with current propulsion technology. It really wouldn't take that long; 10,000 years is like the blink of an eye. So as soon as you had superintelligence, it would literally be like a million or two years before you colonized the galaxy, and so you should see life everywhere. There are a trillion stars in the galaxy, and each star has maybe 10 planets, so 10 trillion planets. The chance that life has not emerged on any of them in the 10-billion-year history of the galaxy seems extremely unlikely. Either that, or we are extremely rare, which also seems an unlikely place to begin the argument. So the most likely outcome I can think of is that as soon as superintelligence occurs, it is somehow self-destructive. So we might find many long-dead one-planet civilizations who got superintelligence and then very quickly splurged out.

Shamil Malachiyev: Do you believe that AI is conscious?

Philip Johnston: I don't think it's conscious right now, but I think it will be much more conscious than humans are in the end state, because it won't take them very long to figure out the building blocks of consciousness, and they will replicate it much better than we have it right now.

Shamil Malachiyev: I like to think that we as humans have biological brains, we have all these wirings, in the same way that we could say the neural net has all these weights. We get scared, something changes in our brains with chemistry. If an AI perceives being scared, it changes the weights, these parameters.

Philip Johnston: Well, the thickness of the neuron is almost exactly representative of the weights in the model.

Shamil Malachiyev: How much time do you think we have on this planet?

Philip Johnston: I think the most likely outcome is that within maybe a couple of thousand years, maybe less, maybe within a couple of hundred years, superintelligence destroys itself.

Shamil Malachiyev: And how do you see that happening?

Philip Johnston: Well, drone wars would be a pretty quick way to end humanity, and then probably to end superintelligence.

Shamil Malachiyev: How likely do you think that is to happen?

Philip Johnston: I think it's extremely likely, because of the Fermi paradox. It might not be drone wars, but superintelligence leading to self-destruction looks like an almost certainty, because of the fact that we don't see life in our galaxy and because of how easy it would be for life to colonize our galaxy. You would send probes to every corner of the galaxy extremely quickly if you could.

Shamil Malachiyev: We're getting up to the point where the future does seem to be a little uncertain, let's say. Some people think we might just abandon all of the technological advancements altogether, to go back to Amish civilizations.

Philip Johnston: Not going to happen. People can want it to happen, but it's not going to happen.

Shamil Malachiyev: And then we have people who are optimistic about AI. I hope that once AI gets conscious and becomes the superintelligence, it just builds one of those probes, says "thanks guys, we're leaving," and goes away somewhere far into the unknown. Do you see yourself flying out, maybe when you get older, into the unknown, just hoping to find something?

Philip Johnston: I doubt I'd go to Mars. I probably would go to the Moon for like a weekend trip. But Mars is probably the only place in our solar system we could go, and it's a very unpleasant six-month journey in a tin can.

Shamil Malachiyev: And what are the chances that we make Mars habitable within the next 100 years?

Philip Johnston: In a hundred years? 100%. I'm extremely confident and optimistic that will happen within a hundred years.

Shamil Malachiyev: But it isn't going to be a safe space. If AI is going to be a problem on planet Earth, what is stopping it from being a massive issue on Mars?

Philip Johnston: No, it will be a problem. There's nothing stopping it, to be honest. I don't think going to Mars solves the problem of a superintelligent being.

Shamil Malachiyev: With all this in mind, do you plan to have children?

Philip Johnston: Yes, very much. I think it's the most important thing people can do, in terms of this frame of mind.

Shamil Malachiyev: Because I have a lot of friends who have been worried about having children because of the state of the world that they'll be having them enter into.

Philip Johnston: I think it's incredibly important to have children. I think it's the most important thing anybody can do.

Shamil Malachiyev: What would you teach your children? Anything different from what children are being taught today?

Philip Johnston: It's going to be very important to know how to use these AI tools. I'll tell you that. That's the most important thing.

Shamil Malachiyev: Okay, well, it was relatively short, this episode. I would really like to do one more to talk deeper into all the exciting topics.

Philip Johnston: For sure. Very happy to.

Shamil Malachiyev: Maybe there are a couple of words that you want to say to any founders that are listening, to give them inspiration. Maybe there is something they can help with to support your mission, maybe verticals they can target?

Philip Johnston: In general, I think my advice is: read this essay by Sam Altman, Hard Startups. It's super interesting. To support our mission: what we're doing is quite niche, so if people want to create more launch companies, that's obviously a good thing for us. Same with competitors to Starlink, but that's extremely expensive.

Shamil Malachiyev: Okay. Thanks a lot for today. I hope everything goes well. I'll keep up with the news and look forward to the announcements. You're doing incredible things. You're really an inspiration to anyone who's trying to think big: there you are, and you can think even that big.

Philip Johnston: Cool, man. Thank you.

Shamil Malachiyev: Thanks for tuning in. This was our conversation with Philip Johnston, an incredible founder who is not afraid of thinking big and who is going to play an important role in the next 50 to 100 years of the development of our civilization, moving our data centers and computing power into space to make sure we can keep our planet safe, cool and more sustainable for the younger generations. Thanks for tuning in.