SpaceX's Real Crown Jewel Isn't Starlink or AI (ft. Julie Zhu)

Welcome back to The Rundown, interview edition. Today, we are talking to Julie Ju, research analyst at Moffitt-Nathanson and co-author of a massive 93-page research report on SpaceX. SpaceX is a super interesting company, as you guys know, and also one of the hardest companies to analyze. They have rockets and Starlink and AI and data centers in space, and of course, there's the Elon factor. So in today's conversation, we talk about it all with Julie, why she believes the launch business is SpaceX's crown jewel, why expectations for Starlink may be too high, and what it's like analyzing a company that has ambitions of going to Mars. This was a great conversation. I think you guys are going to really enjoy it. So let's get into it. All right, guys. Today, we are joined by Julie Ju, a research analyst at Moffitt-Nathanson and the co-author of a 93-page deep dive on SpaceX. Julie, welcome to The Rundown. Thanks, Dave. Happy to be here. I'm super excited for today's conversation. I'm a big fan of Moffitt-Nathanson and the research you guys put out. Before we get into everything, I'm just curious from your perspective, you've been covering SpaceX probably, you know, leading up to the IPO. It's probably been a big part of your life for the last few months. What has that been like, especially with all the hype around it? Were you able to have any conversation outside of SpaceX this summer, or was it all just SpaceX all the time? SpaceX has definitely been top of mind. Like you mentioned, I co-cover it with my colleagues, Craig Moffitt and Nick D'Aldeo. I'd say it's kind of been a firehose of news both before the IPO and post-IPO. Obviously, it's exciting to be able to talk about all the research and work right about now, but we were lucky to be involved in the process from the get-go, so have insight into the company leading up to this date. I'm really curious, a company like SpaceX, there's so many mini-companies inside SpaceX, and there's so many unknowns. You've got the rocket stuff, you've got the Starlink stuff, the AI stuff, there's data centers in space, there's the Elon factor. How did you guys approach it as analysts when it comes to such a wide-ranging company with so many unknowns? Right. Love that you use that word, because that's probably how we'd bucket the different segments too. Maybe for those who aren't as aware, SpaceX reports both in three different segments. The first is the space segment, so that's the rocket stuff. You've got the AI segment, which has both Grok and the space data centers, terrestrial data centers, et cetera, and then connectivity, which generates a lot of the money today, which is the broadband or internet business and your discussions about a mobile business that may be on the come. And then to your point about all these unknowns, the way we thought about it was really the Rumsfeld matrix, so the known knowns, known unknowns, et cetera. You've got the stuff that equity research analysts like ourselves are super excited about, the stuff that you can do price times quantity, math on, project it out, do your DCF, et cetera, and that would be things like your Starlink business, the number of people who buy the internet service per month and then multiplied by the price per month. You've got the known unknowns, which is something like an orbital data center. Do we know if the technology is going to come to the fore? Probably. Do we know if the economics are going to be as advantageous continuously into the future? That depends on whether the bottleneck of terrestrial compute continues. And then you've got the stuff that's the unknown unknowns, asteroid mining, going to Mars, colonies kind of outside of Earth. That stuff is probably something that you can't put your finger on today, but because SpaceX owns the bottleneck to launch, you can imagine that once these opportunities potentially come up in the future in the space ecosystem, SpaceX is going to be able to kind of hold the moat on those areas too. I think I saw in the report you guys were estimating, what, $750 billion for the known unknowns. How did you come up with that? Was that just like, all right, we're just going to take half and then add another half on top? What was the process there? Yeah, and that's for the unknown unknowns. Unknown unknowns, thank you, yes. Yeah, the way to think about it, you're right, is that we have a target price of $131. $170 of that is the traditional math and valuation metrics you might use. And then the other part is related to these unknown unknowns. And candidly, part of it is just thinking about how big the opportunity out there could be and being really cognizant of the fact that it'd be really high hubris of us to say we know what the number is going to be. Not a satisfying answer, but a way that we try to put a framework around it. Yeah, I like that you're factoring it in. And speaking of the details, so like you said, you have a neutral rating on the stock with a price target of $131, which is right around what the stock IPO'd at. I think what stood out to me was, while most of the hype in conversation and the revenue for that matter is coming from Starlink, right? You said the crown jewel of the company was the launch business, which again, that's kind of how the company started. It's a rocket company. But the launch business only accounts for 22% of the revenue. So why do you consider the launch business to still be the crown jewel? Yeah, it's a great question. If you look at the launch business's reported financials, you're right, it's kind of tepid, right? It's like 7% growth last year. The margins are all over the map, but really not that exciting. But you have to remember that the way SpaceX reports that business is based on external revenue only. So anything that comes from launching Starlink broadband satellites up into space, and there are about 10,000 of them in the constellation today, that's not factored in to the space financials. Beyond just looking at the numbers, you have to think, again, going back to the concept of the bottleneck, without having this launch capability and to just ground everybody, the US sent about 200 launches into space in 2025. 170 of them were from SpaceX. And then 165 of those were from Falcon 9. And so all of that is kind of the dominant factor of SpaceX. Without having the ability to send things up reliably into space, and having the kind of technical ability for the reuse of the Falcon 9 rocket, you're not going to be able to have the constellation for Starlink broadband. You wouldn't be able to have the conversation for orbital data centers to begin with. All of that stems from the rocket business. And clearly, I mean, that's where SpaceX has a huge advantage compared to the competitors. I mean, you can make the case that Blue Origin is, you know, their closest competitor about a decade away. For me, though, I guess what I wonder is, is space overrated, right? Everyone talks about this TAM, and space is going to do this, and unlock all these different things, and all these different industries. Maybe space is just going to be used for defense and taking pictures of the Earth, and maybe some satellite connectivity. And that's it. Is that really like a massive TAM there? It's easy to get excited about because it's space, and it's cool. But the economic TAM, is there actually a massive one there? So also a great question. And the point that you brought up on defense is a good one. I think you saw this in their latest quarterly earnings, too, about a couple of weeks ago, where a lot of the growth in connectivity business came from the government enterprise unit. A bit of a black box, honestly. But you can imagine that given the kind of galvanization geopolitically that's in place today and going to continue in the future, you're probably, unfortunately, going to see more of a defense use case in the future from kind of space-related products or solutions. I think part of that calculation, too, comes back to our idea of this potential value coming from the space bottleneck. It's just tough to put definitively a number around all these things that could potentially take place in space. There's a lot of good reasons, scientifically, why certain manufacturing might be better done in a zero vacuum environment or nearly zero vacuum environment, things like that. It's just tough to say today that X percent of the trillion dollar TAM comes from this. But yes, all that to say, large opportunity ahead. Well, I mean, SpaceX in their S1 says their biggest TAM right now is the AI segment, right? Whether it's data centers in space or just their AI business. I mean, the data centers in space, I think you guys said it's pretty far away. How do you kind of see the AI portion, especially their efforts to be like a Neo cloud right now, right? They're renting out all these data centers that they've built up over the last couple of years. They're making a lot of money on that, especially renting them to Anthropic and Google. But is that just like a one-time blimp because they're taking advantage of the shortage right now? Or do you see that as like a sustainable business in the long term? So a lot of different threads to pull. I'll do them one by one. The first is Brock and obviously the cursor acquisition closed a couple of weeks ago. That piece is a bit of a question mark in terms of how much that piece of the business is going to monetize and frankly, whether or not they're going to take the lead in the landscape or frontier labs. Beyond that, there is the whole tension between the US and Chinese and how each country kind of brings their frontier lab to the fore. That's difficult to say with conviction that Grok is definitely going to win. We've seen early data from different benchmarks that from a coding perspective, cursor has helped Grok kind of go ahead and beyond where they were historically. But again, all of these things are product cycle related and these product cycles move quickly kind of on a two-month basis. The second piece that you asked about was the terrestrial data center component and the orbital data center. In our view, those both sort of hinge on whether or not the supply and demand in compute continues to be very tight. If you think about it, the reason why you would have an orbital data center is because of nimbyism, permitting. It's very difficult to build out data centers on the ground. You use it as a release valve because when you look at the economic comparison between a terrestrial and an orbital data center, you'd see that it makes less sense to go up into space even if you can solve the problems of radiation, cooling, etc. in a vacuum. The reason is because 80% of the value of the data center comes from an NVIDIA chip. Those things should last for at least five years but we've spoken to many operators on the ground and for those servers, you have to kind of take out the tray and fix it or have some sort of human intervention anywhere from 3% to 10% of the time. You can't do that easily in space. So then what happens when you shorten the useful life of your most expensive total cost of ownership component in orbit? And then you come to terrestrial data centers. If you continue to have kind of tight supply and demand, then it's possible to imagine that you have the, what we call infrastructure as a service pricing of $30 to $50 billion per gigawatt. That's for context, like two to three times what the New York Times earn today. I saw you guys had CoreWeb on a few weeks ago. I know that those companies are kind of moving towards having shorter dated contracts but really it's only possible for SpaceX to have these contracts because of their balance sheet and who they are as a business. It's tough to see that if the market turns and it's not as tight anymore that they're going to continue to have this persistent pricing. Yeah, that's a great point and that's why I'm really curious to see how all that plays out. I mean, you make a good point about Grok. They were kind of left for dead a while ago. This cursor acquisition seems to be paying off now. I've heard good things about Grok 4.6. So, I mean, I think they're probably still in it but, you know, we'll have to see. I do like how you did have a hot take in the report about moving on to the connectivity part of what you said. How, you know, everyone's talking about Starlink. I think Starlink is the sexy part of SpaceX's business and it's because like it's a consumer product essentially, right? People know about Starlink. People know people that have Starlink. People, you know, have used it before. So, especially if you've flown to United and you use Starlink on the flight, it's incredible, right? And so everyone's like, oh my God, this is going to be a massive business. But you in the report think that it might be more of a niche business. You know, there's always this kind of talks about how Starlink could potentially replace cell phones and internet services and all these different things and the TAM gets massive. But you guys are a little bit more contrarian saying that it might be more of a niche business. Can you talk more about that? Yeah. Also, I used the United service a few weeks ago. Incredible. Mind-blowing. I'm like literally streaming multiple YouTube videos and I'm like, how is this possible on a flight? It was amazing. Yeah, exactly. And that's kind of the power of the LEO satellite constellation versus the geostationary one that most people probably experienced before. But for the connectivity business would break it apart into two. Again, to your point earlier, they have so many different sub-businesses and TAMs they're going after. The first is broadband and that's the internet business. And they have already seen a lot of traction and success, not just with users who put it on the roofs of their houses, but also obviously the United, other airline and aviation partners as well. When we did the math and the way that these satellites work, not to get too technical, is that they shoot out these spot beams. And within each spot beam, you have enough capacity to serve a certain number of households. And these spot beams are, think about them as like maybe a flashlight. As you do things like lower the constellation, which SpaceX has done in the past year or so, you get a more concentrated light. So you get more capacity for those households. You can serve more households. If you have a stronger beam, you can obviously also serve a ton more households. But when you do the math on the constellation that they have today versus the constellation that they'll have, we estimate by the end of 2030, 2031, what you realize is that in the US, the densities of households that you're able to serve maybe edge up from like the least dense, most rural areas to the third to fourth last decile. And what that means is just in New York City, where I reside today, you're probably not going to see people put Starlink up on your roofs and that naturally limits your TAM. And then when you move beyond the US for broadband, you've also got a bunch of different geopolitical considerations. You can imagine Russia and China are going to be completely off limits. China is building their own constellation. I think it's called thousand sales. You can imagine they're going to want their trading partners to use that constellation versus something that is US-based with a high profile person like Elon Musk. So broadband, we're bullish on, but probably not as bullish as the rest of the street. And then you also get to your mobile business, which is the second part of connectivity. I think you saw a certain amount of skepticism from our team that SpaceX is going to become the fourth carrier in the United States after Verizon, T-Mobile and AT&T. And part of the reason is physics related. When you think about high school physics, like Maxwell's law, when you move away in distance, your signal that propagates weakens by a square of that distance. Did you learn that in high school? I don't know if I learned Maxwell law in high school physics. I might not have been in the smart classes. I don't know. It's been a minute, but it is a really important concept here. You're not going to be able to serve with the same capacity when you have a constellation, which by the way, the mobile constellation is a completely different constellation from the internet constellation. You have a satellite that's 220 miles in the air, if not further, versus these macro towers that are maybe, call it five miles away maximum. And so it's just a difficult proposition. And then at that point, what you're going to ask for as SpaceX is probably what's called an MBNO agreement for the carriers who help wholesale in areas that you can't get to. The more I read about the physics challenge that you have to overcome in order to make a phone carrier, the less convinced I become that SpaceX is an actual player here. And I think you're right. I mean, it makes sense as a great alternative to have or a service to have on top. If you're driving out into the rural areas and you have satellite service, that's fantastic. In fact, I think T-Mobile has that because I think I've seen it pop up on my phone sometimes. But as a standalone service, I don't really know what the SpaceX play is. So I appreciate you kind of explaining that and I'll have to look up Maxwell. Was it called Maxwell's theory? What was it called again? Yeah, Maxwell's law. Maxwell's law. I'm going to have to do some more research on that Maxwell's law to really dig into the science. But so more big peek, kind of moving on from that. SpaceX did release their first quarter results recently, and that was after you guys published your report. Growth was strong. And I think what was kind of surprising was the amount of CapEx spending that SpaceX is doing. Maybe it's an indication that they're kind of doubling, tripling down on the AI side of the business. Was there anything from the first earnings report that changed your outlook on the company? Yeah, so there were a couple of interesting tidbits that came out of the second quarter earnings. The first of which is they spent a lot more than expected on CapEx, as you mentioned. And most of that kind of over, overspend is the wrong word, but most of that spend above and beyond what consensus thought was in the AI segment. The team had mentioned that on the terrestrial compute side, they were seeing payback periods of under a year. Normally you'd see it kind of two years or so. And so as a result, they're trying to build clusters, both as a way to couch kind of grok competitiveness in the industry, but also as a way to potentially monetize in the interim as people like Anthropic, like Google, are looking for clusters. Again, that brings up the question, is this pricing going to persist? Because if not, this is a lot of spending. The other thing that came up on their quarterly earnings is they had mentioned this small cell or they call it a femtocell mesh network in order to build out the mobile ambitions of the company. As we mentioned earlier, like it's difficult to see that this is going to be a solution to build a standalone carrier business. Most of all, because these small cells are going to run into the same permitting issues, maybe not as aggressively, but if you think about putting kind of a radio tower on someone else's broadband dish and asking people to put it on their roofs, it's a really big ask. And you're going to run into kind of the same, like I mentioned earlier, permitting. I also do understand like this is not a high margin business either. If you look at Verizon, AT&T, all of them, it's like these aren't 80% margin businesses. This is like a relatively low margin business. So like the obsession over this business unit is kind of interesting. Exactly. And if you think about the spectrum needs that they're going to have when they do a terrestrial build out or if they do a terrestrial build out, how do you balance out whether you're going to spend the money on getting compute, which you're seeing under your payback or building out something to your point where the question is really around, are you going to see a ROIC in that business or a return on investment in that business? Yeah, absolutely. So there's a lot of questions there. I want to wrap with this final question. I think this is probably the biggest question. The SpaceX Tesla potential merger, right? Everyone is talking about it. Elon's even hinted at it a little bit. Obviously, he doesn't want to, he hasn't said exactly. Where do you see this happening? Or do you see this happening at all? I mean, just kind of walk me through your thought process because I keep flip-flopping. Some days I'm like, oh, this is definitely going to happen. It makes too much sense. It'll always be one Elon empire. And then I'm like, well, I don't know. Maybe he doesn't want there to be all under one company. So I'm really curious to get your thoughts on it. Look, your guess is as good as mine. And I can understand when folks talk about the potential synergies, but in my mind, the types that they have, people talk about the TerraFab chip, JB. These are all things that you can do at an arm's length and you can do kind of separately and have transfer pricing in there. So it's just, it's not super clear to me where the rationale comes in. All I can say is when the time comes and we have to build the merger model, not really looking forward to it. I was going to say, that's going to be, you might be locked in a room drinking Celsius and ordering DoorDash for a couple of weeks there. That's going to be a tough one. I mean, I think that's the question there. Cause the other thing is like, does Tesla's valuation take a hit or like does the enthusiasm for Tesla stock take a hit? Now that there's SpaceX because of all the Elon fanboys, this is just me speculating, all the Elon fanboys now have two options to buy, to invest in Elon Musk now. So like you just put that all under one roof and maybe there's some like alpha there. So I don't know, that'll be very interesting. Julie, I really appreciate your time. Again, your research report was fantastic. I love all the work that Moffat Nathanson puts out there. We'd love to have you back on in a few months to see where things stand and see what the revised forecast is from Moffat Nathanson. Awesome. Thanks, Dave. Thanks so much, Julie. Well, all right, guys, I hope you enjoyed that conversation with Julie. I thought this was a really interesting one because SpaceX is such a difficult company to wrap your head around because so much of the valuation today depends on what the company might be able to build in the future. And some of the things they're trying to do is very, very ambitious. I also really liked how Julie broke down on why she thinks the Starlink business is overhyped and some of the technical challenges that it faces. Let me know what you guys thought about today's conversation. Does it change the way you think about SpaceX? Drop your thoughts on Spotify and YouTube. And while you're there, if you have five extra seconds, consider giving us a five star rating or thumbs up as well. You know, all that engagement really does help us out and it helps other people find the show. Thank you guys again for listening, watching and commenting. Shout out to Mike for all the work behind the scenes. And we'll see you guys back here on Monday.