- Good afternoon. My name is Eric Berger, and I'm the Senior Space Editor at Ars Technica. I write about and love all things face. This is the first in a series of discussions on science and technology that we'll be doing monthly on Ars Technica, and I really appreciate y'all joining the first one we're doing here. And I'm really thrilled because today we're joined by Caleb Henry. He's the Director of Research at Quilty Space. And he's really an expert on satellites, which is perfect because we're gonna be talking about mega constellations and satellite internet, and most notably Starlink and Quilty's recent report on the financials of that company, which of course is private. So welcome, Caleb. Thanks for coming on. - Yeah, thanks for having me. I'm glad to be here. - So I mentioned that Quilty had done a study of Starlink revenues. And maybe let's just start out by like, the question I have is how do you go about modeling that since so much of their numbers are proprietary? - Yeah, it's a good question and one that we do get often. So as a research firm, we've created dozens of models for publicly traded companies where every three months they disclose, their revenue, their EBITDA, how much money they're making, how much money they're losing, their indebtedness and so on and so forth. With Starlink and with SpaceX, we didn't have those details. And yet the biggest question that we were getting over and over from our subscribers, which are largely companies within the space industry and financial firms that invest in the space industry, was what about Starlink? Is Starlink successful? Will Starlink have longevity as a system, or is this a flash in the pan? Is this an attempt at something that we don't view as sustainable, and ultimately, therefore, shouldn't have to view as a long term concern? And so when we went about trying to make this model, we said, "All right, let us collect as much data as possible about every single thing that we can on Starlink. So looking at how much of star, excuse me, how much of SpaceX's workforce is dedicated to Starlink? How much money are they making? You can kind of split it into two buckets. The first is all their costs, right? So their headcount, their launch costs, their manufacturing costs, their ground and user terminal costs, and anything else that goes into this. And then you look at what's bringing in money, their consumer service, their government business. their increasing strides into what the satellite industry refers to as mobility. So connecting planes, trains, automobiles, mainly cruise ships, and large airliners, things like that. And so we took all of these figures and put them together as best we could. And then we also did a ton of homework on trying to understand things that they hadn't stated. So for example, how much does a Starlink satellite cost to build? SpaceX hasn't said that number. We did our homework by researching how much it costs to build satellites of various sizes and what we've seen SpaceX be capable of relative to the rest of the industry. And between those and other sources pulled together figures that we felt were defensible or their core costs. And in some instances, SpaceX has given figures or targets like how much it cost them to build a launch, a Falcon 9. So we pull all that together, put it into a model, and then use that to get what we believe is a good proxy for how Starlink is doing as a business. - So just to step back for a second, and Starlink is a constellation of about 6,000 satellites in low Earth orbit. I think they're about the size of a mini fridge, maybe a little bit larger. And they're launching one or two batch of 30 to 60 of these every week. So they're building a new Falcon 9 upper stage. They're refurbishing the first stages. They're building the satellites. They're mounting them, and then they're connecting them and tracking and delivering all these services. They have to build this enormous network of ground stations to receive signals and send them, dump them onto the internet. And they've done this all in about five years from their first operational launch. And so that's an enormous expenditure, enormous effort. And so I'm just curious, like what did you find sort of five years into this project where they are financially? - Yeah, I mean, just piggybacking off of the 6,000, 6,000 satellites is out of like eight or 9,000 for all of humanity. So you look at the scale of this thing, it's absolutely enormous. And so based off of our model, and the outputs that it gave us, I think I'll say I was shocked. I think we as a team at at Quilty Space, we're shocked. but our conclusion was that this year, 2024 calendar year, coming to 2024 ending in December, of course, will be the first year that Starlink generates a profit. And so if you calculate profit or looking at this as a cashflow positive business. So taking their EBITDA, earnings before interest taxes, depreciation, amortization, subtracting capital expenditures, that leaves them, we project with about $600 million of free cash flow on revenues of 6.6 billion for the year. I do wanna emphasize that this is the first year we see that. And I think if we had done this study a year before, we would've said, "Well, Starlink is a money losing business." And same for the year before that and the year before that. This is a massive cost to get to this point, but we believe that now Starlink has reached an inflection point where based off of the breadth of their subscriber base, the efficiency of the machine, the machine that builds a machine, and they're launching and building and laying out of their ground infrastructure network, that they have reached a point where they are then the cost of the business itself and that they will continue to do so as the years go forward. - Do you have a sense of how much they've invested to date to get this system up and running? Like are we looking at like $10 billion development costs? I'm sure how you calculate all that, but what's your given us with the? - I'm sure it's in the model. I don't recall it off the top of my head. If we were gonna look at, just their CapEx for 2023, we have that estimated at about $2.5 billion and going up this year to somewhere around $3.1 billion in spending and you can see them adding new features and so on going forward. A big question is where Starlink, if Starlink ever have sort of a steady state rhythm? I think we've seen one element stabilize that was uncertain in the early days and that was how long the satellite's gonna last, right? Because they were launching these spacecraft and then observers were seeing some of them come down well before their five year design life. But if you look at the average age across the fleet today, while they do the orbit some, the overwhelming majority, like over 90% are staying longer than four years. These are spacecraft are, they're getting their money's worth out of them essentially. So I think there is some variability that could be introduced if and when Starlink introduces additional upgrades. But our model has them continuing to stay in that sort of, cashflow positive territory, assuming that they don't do these major lurches for new investments. - So they're up to about 6,000 now. Do we have a sense of how big they want to grow the constellation before they potentially get to this steady state you talked about? - I don't. SpaceX has never as far as I'm aware, given an upper statement on how many they anticipate launching or needing. I think at the very beginning of this program they had done a call, and said that they needed about a thousand satellites, a thousand for Starlink to be economically viable. I think that was a low number. From what our model suggest, it looks like it's become economically viable now five to 6,000 satellites. And they have regulatory filings for 12,000, and maybe upwards of 40,000 satellites. Like others, we've used their filings with the ITU at the international level and the FCC at the domestic level to get a sense of like what's the upper limit of what they're gonna do. It doesn't mean that they will build out 12,000 or that they will build out 40,000, but we know that they could go as high as that. And so that's an element I would say that we are are watching for. SpaceX continues to launch at a faster and faster clip with their rockets, but they're also starting to put fewer satellites rocket, as they introduce features like their direct to device service. And so I think a lot of how fast or how aggressively they continue to scale their network will be influenced by the future design of the satellite, their version three as we call it, and the timeline for operationalizing Starship. - Yeah, I'm gonna ask you about that in just a second. But first, are they truly a global network now? Like do they have satellite, how would do the entire planet? Or are they still targeting like strategic areas worth the highest population? - Yeah, so I think from a pure like can Starlink beam down over the Earth's surface? Yes, should be global now with the one caveat being, I don't know how they disperse the satellites that have crosslinks. And I think there's still some of the early spacecraft that don't have that feature. And that's gonna mean that they could have patches here or there if they don't have a ground station in view on your surface. This is what makes places like the Southern Pacific hard to cover, for example. But as they build out the network of ground stations and as the satellites increasingly have crosslinks, their coverage goes up. And so I haven't heard anyone at this point complain about not having coverage. - I mean it was pretty impressive that they were in a remote area of the Indian Ocean and we were seeing a video Starship coming back from space. Obviously have some coverage for pretty good globally. - Yeah, I think that's the biggest testament to the crosslinks working 'cause that was a difficult part and I think SpaceX wanted to have that from the very beginning. But laser crosslinks are a really challenging piece of technology. You gotta get two beams to point at each other at moving objects in space and hold that link. So it took longer than the rest of the spacecraft to get going, but now they have it and they can use it for things like showing the rockets returning. - Yeah, how important is Starship with obviously it's much larger pay, not only payload capacity, but sort of the size of its fairing to really chunk a lot of stuff in there. How important is Starship to the long-term viability of Starlink satellite network both just from a financial standpoint, also from the direct to sell capability that they're worked on with T-Mobile and potentially others? - Yeah, so I think we all remember a year or two ago when Musk get sent out this kind of urgent email saying, "Everybody we need to get this vehicle working." Otherwise the business model falls apart for the luxury constellation. I think that the V2 mini has served as a really effective gap filler for Starlink. And that's a the V2 mini is what? Somewhere around 730, 750 kilograms of spacecraft. It's about 2 1/2 to three times the size of the earlier iterations. And that has helped them with increasing the amount of capacity that they can offer like four to 5x what it was before. And I think the satellites are somewhere between 80 and 100 gigabits per spacecraft so it basically gave them breathing room on the way to their next generation. And so now what we're waiting to see with Starship is how much more capacity can SpaceX put up into the network and at what kind of pace? And frankly, I don't think that SpaceX has even figured this out. If you look at their FCC filings, they've submitted two different architectures to the agency based on kind of like how they might fill out orbital plans. And this looks very much influenced by whether or not they have Starship flying at a regular cadence and whether or not Starship is actually lifting the target up-mass that Musk has described for the program. Now he said that they wanted to launch a hundred metric tons and going up to 200 metric tons, but some of these early launches have been more in the 40 to 50 metric ton range. So there are a lot of moving pieces here and that's why what happens after 2024. We ended our model projections at six months from now because there's a great deal of uncertainty that comes into when Starship starts launching, how much up mass it can carry, and what the design will be for the V3, which I think people who have followed the program might say, why are you calling the V3 and not the V2? That's based off of a comment that Musk had made at IAC about. I think it was last IAC, he anticipated or projected launching the V3 version of this satellite in a year's time. So that will put those launches in this fall, this Q4. We'll see if that schedule holds, but the cruelty team that red as V2 mini being the actual V2 permanent now and then moving into the next generation as soon as possible with their next generation rocket. - Yeah, one of the things that anyone who has followed SpaceX for any time period understands that their plans change almost on a daily basis. And some of that is, you know, could change based on Elon's mood, but a lot of it is like they are just plotting the straightest line to the future that they want to bring about and their ideas. They get more data, they get more experience, they have better ideas, and they just sort of say, okay, well, we're not gonna go down that culdesac anymore. We're gonna go this new way because this is the better way. And then so I think sort of going out six months into the future and leaving it there was probably a wise decision. I'm gonna ask Caleb some more questions, but I wanna remind people who are watching this that on YouTube you could submit questions. We're taking those. We're gonna have plenty of time, for shoot your Q&A with Caleb as we get a little deeper into this chat, which is gonna run about an hour today. I'm gonna ask you Caleb, you mentioned consumer customers, obviously the US government is a customer, in terms of revenue, do you have a breakdown kind of of how much of Starlink revenue's coming from government versus how much is coming from consumers? - Yeah, so at this point, we believe that the majority of their revenue is coming from the consumer sector. Like well over 3/4, probably, well over 80% I would say, it's coming from the consumer market. And then of that consumer market, it's something like half of it is coming from the United States at least. And then the rest of it, coming from other parts of the world. Actually, I think we had it at like 60% US, and then 40% rest of world with that likely equaling out in 50/50 maybe by the end of this year or sometime in 2025. And I think it's worth remembering that Starlink started as a consumer business. This was actually something that other operators would use to say why they weren't competing with Starlink initially, but so Starlink is direct to household. They were connecting the war fighter. They were connecting DOD. They're connecting airplanes and ships and things like that. And so it's only more recently that SpaceX for Starlink has begun gaining market share in those verticals that are more of the territory of kind of the historic thick satellite services provider, yeah. - So the other kind of quote mega constellation, we've seen come online in the last five years is OneWeb. And they obviously had some significant disadvantages in that they didn't have their own rocket. Russia invade Ukraine, and stranded some of their satellites and they had to find alternative launches. Some of them even launched ultimately on the Falcon 9, because that was the only really Western rocket that was available to get them into space. And I haven't heard a whole lot about them, but I know you're really well-connected on OneWeb. And I'm just wondering like how are they doing financially? I don't think they're doing much consumer services, but are they finding government customers and like is that a sustainable business? What's your take on kind of OneWeb now that they're getting deeper into operations? - Yeah, so I'll use this as a plug to say that I'm writing a book on OneWeb. And there will be lots of details on why OneWeb looks very different from Starlink and vice versa. But what I think is interesting about where OneWeb is today is yet they've completed their Gen1 network. All eyes are now on what they're gonna do for their Gen2. And based off of what their new leadership at Intelsat has said, I believe that a decision on their Gen2 constellation will be imminent. I think this summer, OneWeb has had an interesting journey in that, I think what we are seeing is a lot of competing operators using that as their way to compete against Starlink. So you look at what Intelsat has done. They placed a $250 million contract with OneWeb. And I believe that has the option to expand to 500 million Eutelsat before they had acquired it, had put like 200 million or so dollars worth of backlog onto OneWeb because even though there are different networks, there's still this big advantage that comes from having low Earth orbit from low latency. And I remember talking to one customer of theirs, I attended a OneWeb launch last year. Yeah, it was the last launch they had out of India. One of their customers was telling me that they literally couldn't connect to geo satellites because they were in two mountainous of the country. And they were excited about one way because they were able to have better look angles. You also see their owner, Eutelsat has a strong presence in Africa, in Africa and Europe and other parts of the world, especially as a global operator. So I think between what you see these other satellite operators doing and what some emerging players have done, whether it's Joomla or in Saudis, they've really banded together behind OneWeb as a another network. And I think this is interesting because it reminds me a bit of the launch scene where one of the things that actually helped SpaceX become a significant player in the launch sector was the faith that SES put in them back in 2013, 2014. Companies like having multiple choices. They don't want just one rocket. They didn't want just the Russian rocket or European rocket. They want Russian, European, and American, ideally at least three. And I think we're gonna see the same thing with the low Earth orbit. There's not gonna be satisfaction with just having a single network. And this creates an opportunity for OneWeb and it creates opportunities for others, and again, at least one more. - It's funny you bring up SES. It's a funny story I can tell about their first launch on a Falcon 9. It was the first one point version, 1.1 Rocket think it was the seventh launch overall Falcon 9. The first 1.1 to fly from Cape Canaveral. The first one had gone from Vandenberg. And so they were like, they were madly sort of trying to strip down and rebuild the launch site to accommodate the bigger 1.1 rocket in Florida. And SES shipped over the satellite because SpaceX told everybody in January or something. And they came over and they saw this like launch site and disarray when the executives came over and like they had this very heated conversation with John Muratore, and some other SpaceX officials in Florida. Basically saying, how could you tell us to ship the satellite? You don't have this big air conditioned facility, it's basically a tent. Your launch site is in ruins, and they were like, we're taking the satellite back 'cause we don't wanna leave it in Florida for like six months. And Muratore told them that, "Look, go back to Luxembourg for six weeks and then come back, and you'll have a grand story to tell your, or you'll have a story to tell your grandchildren about how these maniacs at SpaceX completely rebuilt the launch site in six weeks. And lo and behold, they came back and they were blown away. SpaceX ended up launched that mission I think in January. They were like a little bit later for other reasons. But anyways, yeah, SES and some other companies like Iridium showed a lot of faith early on in SpaceX to help get them going. What about Project Kuiper or Kuiper, however you choose to pronounce it. They obviously have some benefits. They don't have a a rocket either, but they have a much wealthier backer in Amazon to really kind of fund and support that program. Do you think because they're gonna be five to 10 years behind Starlink, they'll be able to compete? It does seem like they've got some pretty strong use cases with the Amazon's existing customers to really use that network. But sort of what's your sense of the viability of Kuiper going forward? - Yeah, I think, like you said, wealthy backer and everyone knows Amazon. And interesting, I'll kind of try to steer back to the conversation without going too far off, but one of the things that the industry had always looked for was where were the wealthy entrants gonna come from? Because satellite operators, like your biggest satellite operator is today, or you know, now through mergers and whatnot. You look at the Viasat which merged with Inmarsat making about $4 billion a year. Intelsat and SES when they merge gonna be making about $4 billion a year. We have our projection for Starlink at $6.6 billion revenue this year. But Verizon and other MNOs spent like $80 billion on spectrum alone. - Wow. - Just to bolster their terrestrial networks. So there was always this expectation that either these kinds of telcos or Silicon Valley would get into the satellite business and that money would be no issue for them because what's an insurmountable cost or nearly insurmountable cost for, I just gave you the headlines for the biggest satellite operators in the world, everybody else who's smaller than that, it's a huge challenge to try and deploy one of these networks. So sure enough, it was Amazon, and I think Amazon was a bit of a surprise. A lot of people expected it would be Facebook. OneWeb was a secret project at Google once upon a time. So there's the expectation of command of Silicon Valley, but no, Amazon is getting into this. They've got real expertise and building, and shipping producing consumer products. And so I think one of the things that was really interesting was the fact that they wanted to talk about their user terminal before they talked about their rocket deals or before they talked about building the satellites. Because I think the industry has long understood that having a low cost antenna, low cost user terminal was the key to making this business close. And so Amazon started with that, and said, "All right, now that we have solved historically the most difficult part of this business, let's go forward and do everything else." I am curious to see how, I do think that there's a challenge in the launch part of it. They have now procured vehicles. They have four different launch vehicles. They've got ULA's Vulcan. They've got the Ariane-6 from Europe. They had to purchase a few SpaceX, Falcon 9s, because other vehicles were falling behind schedule. And they've got the New Glenn from Blue Origin. They say that they're still gonna be able to, they're guiding towards the FCC target of having half of their network up by July 2026. Hoping to see a few of those launches, and I think this year is what they're talking about. I guess to answer your question, like I do think that they've got this embedded customer in themselves and in cloud networks. They have a very easy base to launch off of. One could see them bundling hyper internet with Amazon Prime and having a direct conduit to customers. And again, going back to the kinda previous comment about consumers liking choice. I think one of the things that really helped Starlink get as far as they they've gotten is the fact that prior to them you had two incumbent players. You had Hughesnet and you had Viasat. They were both in Geo, and even though they had really pushed the bar as far as geostationary satellites are concerned, they are and we're operating some of the most powerful geo satellites in the world. These things were taking like seven years to build. - Yeah. - And what happen was they would launch a satellite, all these customers would come onto the network. They would fill it up in three years, and then there's no more room and bandwidth demands grow. Just like us doing this video call a year from now is probably gonna require 10% data. So you ended up with this kind of war against attrition that these operators couldn't really accommodate more customers. And then Starlink came onto the field with a flood of new capacity and they were able to soak up all of this latent demand that the incumbents couldn't meet. So I think that the market will welcome a second player just like it welcomes Starlink. The challenges do the competitive advantages that Amazon has. Are they as effective as the competitive advantages that SpaceX has? I mean, SpaceX is on the launch side, Amazon's I would say is on the ground side. - So we haven't talked about China's mega constellation plans. They're talking about thousands, potentially tens of thousands. Europe has a constellation that may or may not get off called the ground called IRIS squared. And I think other countries, excuse me, I'll look into this as well. I guess just the question that, the last question I have before we get to reader questions is, is how sustainable is all this? Have we learned anything from the thousands of satellites now that SpaceX has up there? Like is there enough room in LEO for all of these constellations? Do we know yet? - Yeah, so I don't know. [laughs] I don't know how much room there is for all of the constellations that are out there. What I will say is that the operators that are in service today have certainly, or either in service or are soon to be in service, have selected some of the ideal orbits for these kinds of things. And so just to give an example, Starlink is at 550 kilometers today. They're above the space station, but below the threshold for what used to be the guiding post, the 25 year rule, which stated that operators should de orbit any defunct spacecraft within 25 years of end of service, end of life, at that orbit, if a satellite fails, it will come back down naturally. And so you don't have to worry about it being up there for centuries or millennia, eons. You go slightly above that at around, I wanna say it's 590 to 630 kilometers and in three bands is where Kuiper is expected to be. And I would defer you to an astrodynamicist to tell you exactly where 25 starts and stops, but iscenic go as high as 650 kilometers. So you could say that they are also in that safe shell of kind of space sustainable orbits. You know, below that you get your very Leo, feed LEO orbits. Below the ISS around 300, 400 kilometers, where you start to deal with a lot of atmospheric drag. And so you kind of have to design your spacecraft a bit differently. Starlink has talked about putting satellites there. We'll see if they have a different design for that. I don't know if they will, or have talked anything about that. And then you kind of go above 785 kilometers is where, you've got Iridium. And then around 900, you start to get defunct rocket stages, which are bad company or an orbit. So you don't wanna put anything there. A thousand is Telesat, the Canadians 1,200 is OneWeb. I think a thousand is also where the space development agency is at. And as you go higher up, you can still have low latency with fewer satellites. But then you have this risk of if you have a failure and you are not hiring somebody for active degree removal, like an asterisk scale or a clear space that could be up there for civilization type timelines, timeframes. [laughs] - Yeah. - So yeah, I think the best orbits people are selecting now, it doesn't mean that there's not more room. Low Earth orbit is defined as going up to the 2, 000 kilometers. I wouldn't be surprised to see more maybe go a little higher, maybe more go a little lower and try to tease out some of the optimal orbits in between here and there. But it certainly gets harder. And I think this has been a reason why some companies have chosen to invest in LEO networks or invest in partners that are active in LEO because they saw this is not something where 90 countries can claim to the low Earth orbit constellation. There's not room for that many, I can tell you that much. - Great, so we're gonna get to questions, again, you can submit those, and we'll take as many as we can in the next 30, 28 to 30 minutes. Andrew asks, how likely are various Starlink competitors to be able to be profitable without the vertical integration of launch services? So that obviously, it's a question about SpaceX has the Falcon 9, which is this huge competitive advantage. How can the other companies compete with that? - Yeah, so I would remind people that there are satellite operators that are profitable today and they don't own a launch vehicle. So this isn't a new thing for them to do. It's harder with lowered or with low Earth orbit because your costs are going up. But I think it's still, and everybody, it depends on how they structure their business. What is the lifespan of your satellite? And I think one of the things that we're seeing here, satellites that are going up in design life, even Starlink is talking about adding additional redundancy to future satellites in the V2 minis, right? They've got two solar arrays, either the V2 mini or the next generation of them are supposed to have two propulsion units and more redundant power systems. Power is something that consistently causes, it's a leading cause of spacecraft failures and anomalies. So adding redundancy means that those spacecraft will last longer. If you can stretch the design life and still get the same revenue from it or comparable revenue from it, then your business case has a greater chance of closing. You look at what Telesat is doing, I think their spacecraft are supposed to last 10 to 12 years. So you get the network up and if you spend what? Three or four, $4 billion getting that there, there's a point. And then we did a report on Telesat's Lightspeed as well. I can't remember the number off the top of my head, but there is an inflection point where you break even and then you make money. For geo satellites, it was seven years. A geostationary satellite was typically designed for 15. So you spend three, four years building it, getting it launched and calibrated. Seven years, recouping your costs, and then the next seven years is the gravy train. That was the way back when everybody was watching TV from satellites. The pivot to low Earth orbit is because of a recognized need for scalable broadband and the advantages that a low Earth orbit network has for internet services versus TV. But at the risk of going too far off topic, there's certainly a pathway for this. And for each operator, it's a function of what are their strengths, how long do your satellites last, what have you used as your value proposition? And even going back to the Amazon paper versus Starlink example. Hyper has, excuse me, Amazon has data centers all over the world that they can use to place their gateways at. And one of the things that I've heard more than once over the past year is that for even geo spacecraft, the number of gateway sites that are being needed to pull down all of this bandwidth, 'cause they're offering more capacity too, is starting to go so high that they have to entertain project financing. They have to borrow money to pay for the ground network, which used to be something they only did for the rocket or the actual manufacturing of the satellite. So I wouldn't underestimate the fact that there could be serious advantages that other players have. It's something I wouldn't sleep on with Hyper. And I would watch to see what other players are doing to show where they can be strong in the market. - Joey says, "I've noticed that Starlink has an agreement with Roger's communications for supplying Starlink for satellite to phone communications. It's currently SOS texting. Will Starlink allow voice soon?" - So I'll just parrot what I've heard SpaceX say and that's they plan to start with low data rate services. So texting is something that you can do very easily and send thousands of text messages with a little bit of bandwidth. Voice starts to take more data, video takes even more data. So I'm sure there's definitely a timeline in the roadmap for voice service. I don't know what it is off the top of my head. I think they recently said, they've got like 8% of the satellites they need to start their direct to device service. And I would anticipate that scaling as the years go by. The only other data point or detail that I would add is that I think it's really interesting, and cool that you see so many cellular network operators or mobile network operators partnering with satellite companies on direct to device. Historically, the terrestrial companies and satellite operators used to fight tooth and nail over the best way to use spectrum. And oftentimes, it was your cellular networks that were saying, we can make better use of that satellite spectrum so we have more customers. And then there's satellite operators saying, "Our services are just as important as yours and don't play that game." So now we're kind of in this new era where the two are working together to provide, the FCC calls it supplemental coverage from space. I'm excited to see where that's gonna go. I've never seen this much uniformity in terms of goals and directions between the two industries. - We could do a whole nother hour on direct to device technology. So we'll pull back from that abyss, I think right now. Gerald asked, "How much does it cost SpaceX to launch a Falcon 9?" I think he's probably asking like what is their internal cost to refurbish a first stage, throw a new second stage on there and get it to the launch site. - Yeah, so at one point, I think SpaceX had publicly said that it was $28 million was their target for it. We believe today that they are below $20 million per launch and actually lower than that. So I'd say I would put it in the mid teens for how much it cost them internally. And that's going down as they increase the reuse of the vehicle. Recently they've launched their 20th, maybe 21st use of a first stage rocket. And as they can amortize the cost of the booster over a greater number of missions, that only helps them with their business case. Another interesting data point that came up between now and when we actually published the report, SpaceX does these really cheap ride share missions, for small sets. And there was that New York Times article where they were confused of, excuse me, accused of anti-competitive practices. And then Musk came on the, yeah, his platform X, and said that he doesn't believe they've ever lost money on a launch even for these ones where the rocket's not all the way full. And so one of the things, at Quilty, we certainly anticipate revisiting our Starlink model and maybe putting out another update six months or 12 months. Don't hold me to a date, but when we reevaluate this, We'll look at all of the new data that has come in. And one of the questions that we'll surely ask is, okay, if we take this at face value and really believe that SpaceX has never taken a loss on one of their rideshare missions, then that kind of gives you a figure and say, "All right, at the very least this is an internal cost." - Right. - So we'll continue to study that, but it it's certainly come down as the years have gone on. - Certainly a powerful argument for reuse that you could launch 15 to 20 metric tons to low Earth orbit for basically a million dollars a ton. So Falcon ask, what do you is the ceiling for Starlink user base? I suspect this question is like how many consumers are there out there really who would benefit from satellite internet that cost wise about a hundred bucks a month of their service? - Yeah, this is a good question and so we didn't do like a total addressable market, a TAM assessment- - Right. - In this report. One way that you could start to draw boundaries around this and then you kind of reign them in to get to a figure is you can look at what the ITU says for how many people are offline. This was the inspiration for the name of an older company called O3B, used to stand for the other three billion, 'cause back then humanities population I think was six billion and half were offline. I don't recall the number at present. I think it's still around three, maybe I think it's 2.6. I think the IQ figure was 2.6 billion people did not have internet access. And so you can start with that as your high figure and then say, all right, of those 2.6 billion, how many can afford a connection? And your number's gonna shrink. I'm making things up off the fly. So you say you cut it in half and you go, okay, how many of those are in countries that will authorize this service because Russia and China and others have not and may not, probably won't ever. - [Eric] And so then you probably won't though. - You keep shrinking this and so you go, all right, 2.6, and then maybe you end up with a number that's much smaller. But even if that's pulling a number outta thin air, even if it's only 200 million people, that's still prior to Starlink. The two million is the number that was collectively connected by Hughes and Viasat. So you could say, all right, these two players had only scraped 1% of their total addressable market, that gives you ample room to grow. Again, purely hypothetical, but if the question is there concern that they would run out of room? I wouldn't expect that anytime soon. - Two countries, I think you said you were watching were Indonesia and I think India as well. Can you kind of give us an update on where they are SpaceX is in getting into those markets? - Yeah, so when Quilty space, we did our own kind of webinar to give a highlight of, our conclusions from the model. Indonesia and India were the two countries that we described as future needle movers because Starlink had not yet been authorized there. Since then Starlink has received market approval to provide service in Indonesia and that's a big deal, 'cause Indonesia is the fourth most populous country in the world behind the US, China, and India, and is a vibrant market for satellite communications. There's something like 270 million people I think it is, and spread across 17,000 islands. So where satellite internet excels is where it's really hard for you to get terrestrial connectivity. It's when it's too expensive to build a fiber, too expensive to build a cellular tower or fiber to a cellular tower. And so those are your traditional, your Verizon's of the world, or your Oranges or [faintly speaking] don't wanna put that infrastructure there. So now that Starlink has access to that market, we're really waiting to see what kind of adoption they get because it's a huge place. And if I recall correctly, I think something like, at one point, there were like 40 foreign geostationary satellites that we're serving that market, and the Indonesian government has gone so far as to bankroll satellites. I think it's the local operator, PSN, Pasifik Satelit Nusantara, that is helping the government with providing connectivity through a geostationary satellite. The government has also built fiber through the Palapa Ring. So bottom line, Indonesia has invested very heavily in building out its broadband infrastructure. There's a large addressable market there, and it's good news for Starlink that they have access to it now. India historically is a very difficult market for operators to get into. Hughes has wanted to get into that market forever. I think OneWeb is still trying, notwithstanding that they've got- - [Eric] It's crazy. - Got [faintly speaking] [laughs] - That's crazy given their ownership and they launched on Indian rockets, wow. - It just tells you something about how challenging from a regulatory perspective India is to get into. But people seem to believe that the barriers are gonna come down a little bit and that there will be more opportunity there. And that's what 1.4 billion people total, of which there's surely a percentage that does not have connectivity today. And again, the ITU have those numbers. So that's a large total addressable market that you're waiting to see. After that, going from memory, there's a handful of countries that haven't authorized, but I think they're closer to like a hundred million people. And they'll surely still be significant to the business case. But I think less so compared to India and Indonesia. - I saw Elon tweeted a congratulatory message to Narendra Modi. So I know he's working at political angle to get into India. We'll see if that's successful or not. Alexander asked, "Do we know why they targeted a five year design length for the Starlink satellites and not longer?" - Yeah, that's a very good question. Frankly, I don't know, but I can speculate. [laughs] So why a five year design life? One of the ideas or ideologies that has sort of underpinned the small set movement is this notion that you can have spacecraft live shorter lives and then you retire them and replace them with newer tech because bandwidth needs are always increasing. Demand signals are changing. And if you have an asset that lasts 15 years, you could run into obsolescence. I mentioned the geo guys that they launched these 15 year satellites, and they are out of capacity in three. That makes for a challenging business case. So if you lower the design life, then you can have a more responsive constellation to your demand. Another one that I would say for SpaceX is because of the orbit that they have chosen, they are close to the Earth and it's a more radiation friendly environment than satellites and medium Earth orbit geostationary orbit or even the higher end of low Earth orbit. And when you don't need as much radiation hardening that allows you to use lower cost components. And so they can have a satellite that lives for a shorter period of time, but it's also gonna be considerably cheaper. And we had estimated for the early satellites that they like the V0.9s, V.1s were costing like $200,000 at staggeringly low price. And then as they increase in mass and power and redundancy, that figure climbs closer to a million dollars a satellite. But I think in order to keep costs low and to stay responsive. It was in space like this favor to go with a shorter design life. - Great, I know we talked a little bit about direct to device coverage earlier and we don't wanna go too deep into it. Technomadia asks, thoughts on how quickly direct to device will have coverage. How will the race play out with Apple slash Globalstar play out? Maybe just, I would ask you like, there's several entities that involved in this. There's AST, there's Apple Global Star, there's SpaceX, there's Link, I don't know the industry, so I'm not that well-versed. I'm just wondering like do you have a sense of who's ahead in that sort of direct to device race? - Yeah, so I mean you've got like five or so major players in the West right now. I think you did run through, you've got AST, Link, Starlink and then, oh, so actually, let me bracket it a little bit differently. You've got your players that are trying to reuse cellular network spectrum, that's those three AST, Link, and Starlink. And then you've got a class of players that are trying to use Spectrum that satellite operators already have claim to. This is typically S or L Band. So narrowband frequencies, they don't have as much bandwidth, but the signal tends to be really robust and this lends them well to things like emergency services, emergency SOS and that's what you see with the Globalstar and Apple. They would fall into that camp. Look, I think that it's a matter of when not if, or direct to device services in terms of how they roll out. The big question is just the how? We were seeing a couple of different methodologies going forward. You've got this camp where operators are working together with your cell phone companies, and I mentioned that's really exciting because they have often been at odds with each other in the past, and maybe worked together just on the fringes doing some like cellular backhaul work. But I think that that's exciting and I think that we will see that go forward. The FCC has been very supportive of this and has wanted to create a clear regulatory framework for that to go forward. And then you've got your operators where they have existing assets. They have been waiting for something like this to happen for a long time now, and are also trying to figure it out. And who's to say that there won't be both. I think we can look at the business model of Iridium, for example, where Mat Desch has just had to fend off [laughs] round after round of questioning where people come and say, "Well, isn't Starlink gonna eat your lunch?" And the Iridium revenue is kind of up until slowly creeping up until the right, yeah, every quarter. And it's because they offer a different type of service. L Band is a different flavor of connectivity and if you need emergency connectivity, that's who you're gonna turn to. So there might be, here's a service where, all right, you wanna stream YouTube by the Grand Canyon? Sure, beam it over Starlink or AST or Link. If you are stuck in a Shenandoah and you just need to press a button and get somebody to help you, that might be your Apple Globalstar, or some future service from Viasat and Inmarsat or Iridium. Yeah, I'm excited by the progress that all these players are making. And I'm just waiting to see how their approaches play out. - It's very much sort of a wild, wild west in the sense of we don't really know where things are headed. Lots of activity interest certainly in that area. WinQuik asks, how much is Starlink a product of dysfunctional US ISP regulation? [Caleb laughs] - That's a good question. It's probably a bit beyond me. My expertise is solely within the space sector. So what other ISVs are doing, I couldn't speak to. What I will say the one area where I noticed this overlap was with the FCCs, RDOF program, the Rural Development Opportunity Fund, Rural Digital Opportunity Fund where Starlink had been awarded a significant chunk of money and then you had kind of an administration change and the new FCC leadership rescinded that, citing unproven technology. I think that's gonna be proven to have kind of bitten the FCC. They've already penalized other operators that they gave funding to under RDOF for failing to meet some of those targets. And I do think the satellite industry association has certainly argued on the hill that terrestrial connectivity companies have had the year of regulators and congress much more effectively than satellite operators over this in years past. So I think that a fair playing field for whatever type of connectivity infrastructure a company chooses to use, whether that's a satellite, whether that's cell towers, whether that's fiber, whether that's hacks or something brand new. I think a fair playing field is the best approach and giving consumers choices is something that they'll appreciate. - Great, so we've got just a couple more minutes. Lucas asks, do you think the V2 or V3 satellites will have the community to communicate with things above it? So obviously, we've been talking about all Earth facing communication, but SpaceX, NASA are interested in going to the moon, they're interested in going to Mars. So I'm wondering like, we've got the deep space network right now, which is set up with these large dishes on the ground to get signals back from Mars and other spacecraft in deep space. But that's that the decent network is certainly at capacity and there's very little, I think, expansion room for, you know, commercial activities in the moon and elsewhere. So I'm just wondering how much of a role does Starlink have to play, if any in deep space communications? - Yeah, I mean here, again, it's another area where I'm gonna have to profess ignorance. [Caleb laughs] I know that NASA has contracted with a number of companies to study ways to supplement the Deep Space Network. There's also conversations around how do you provide connectivity on the moon. I think SpaceX has floated the idea of Starlink for Mars. So everything that SpaceX does is with this long-term vision of getting to Mars. And I think to the extent that having a relay between multiple orbits or multiple celestial bodies aids them in that quest, I would expect them to do it. - Yeah, there's a Mars clips program that NASA's put out that SpaceX is competing in, I think to do communication relay services, which would be some form of Starlink which again, I don't fully understand either. Do you think Starlink will be seen as a monopoly by US and EU regulators in the near future and thus will be forced to open up to wholesale customers? TTG Movie DB, ask that question. - Okay, look, I think the idea of a monopoly here is silly. There are 50 to 60 geostationary, we're talking just about satellite communications. You've got somewhere between 50 to 60 in geo alone, and then you can probably add another handful in low Earth orbit. This is not a market where SpaceX is alone. It's not one where they are the first. It is one where they're extremely competitive. So I would just, I mean, I don't know. I think that what SpaceX has done has made people more aware of what satellites can do and the potential that satellite connectivity frankly has always offered. It's just never been fully realized. And now that they've done it, I'm excited to see more players come on board and try to do similar things. If certainly, there was the Wright brothers and you had your first airplane and you had more airplanes, and Henry Ford, you get your first car and you get more cars. I'm assuming Ford made the first car. I'm not 100% sure, but I would hope that you'll see a similar trend and you've already got, yes, there's Starlink, you also have OneWeb. And OneWeb does have a nontrivial customer base, remind people of. We've got Kuiper coming up. There's certainly the expectation of at least one, if not more constellations coming outta China, Telesat that in Canada, and maybe the IRIS Squared network out of Europe. So I see more of these coming forward. And just the other day, Yasat out of the United Arab Emirates ordered two low Earth orbit satellites. They haven't been fairly detailed on what they're for, but that is the pattern that was exemplified by Starlink, by Kuiper and by Telesat Lightspeed before they pulled the trigger on full fledged LEO internet constellation. So I think it's a space that's really dynamic and I'm excited to see more players come in without regulators sort of jumping the gun. - Okay, great. So I've just got one more quick question for you. Is SpaceX now a launch company or a satellite communications company? [Caleb laughs] - My quick answer is, I think today they're a satellite communications company. I think it's interesting that Stefan Israel from Arianespace used to in the early days, like 2015, 2016 when Starlink was just announced, wasn't even called Starlink yet. And he would try to court customers and say, "Do you want to fund your competitor?" And no one really took him seriously. [Eric laughs] Now people are taking him very seriously. [laughs] a huge impact. They're the largest satellite operator in the world. They have literally more than doubled the number of consumer subscribers for satellite internet in the world. To me, it's their biggest, and I'm biased, because this is the part of the industry I care about. I don't really watch Bruce launches that much, but this is a humongous, nearly unrivaled impact that they've had on the industry. - Well, thank you very much, Caleb. That was a fascinating discussion. I thought you just brought so much knowledge to the table, so thanks very much for joining us today. I appreciate it. - Thank you. - Also, I just wanted to say, thank everyone for tuning in today, and say we'll have a transcript and this conversation up on the homepage in a few days. Thanks very much.