How Elon Builds Trillion-Dollar Companies

0:00Look, I personally believe SpaceX is the greatest company of all time. probably the most vertically integrated American manufacturing company. When you're vertically integrated, it just lets you move with like a speed and flexibility that no one else has. You know, and that's where things like space data centers can just happen very quickly, shockingly quickly. SpaceX in 2019 was just a launch company. In 2019, they had 13 launches. The total global launch market, I would say in 2019, is probably like a five, six billion dollar market. Probably like a billion of profit if you capture 100 % of the market then.

0:31The company's valued at$36 billion in 2019, investing at 36 times the theoretical maximum earnings where you wouldn't be growing very much if you didn't find another market. I first invested in 2019, in cumulative invested$1.2 billion. That position's worth like$12 billion today. In the$800 billion valuation, hopefully in the IPO it's worth a lot more than that. I think this is the healthiest wealth creation event in history. I think that XAI is severely underrated. You think Elon's underrated?

1:09Welcome to Sorcery with Molly O'Shea, who's becoming very famous. It's great. And I'm Sean McGuire. I am a legendary ex-poster. Oh, shit poster. Post with a P-O-A-S-P-I-N-G. So we have, I think, four different routes we can take this conversation. One of them is just going to be exclusive to X because YouTube won't allow it. And we'll save that for the end. Which route do you want to take? One through three. You like math? You choose. Okay. I don't know what's behind each of the doors, so you choose. Okay, so let's start with SpaceX's upcoming IPO. I think a lot of people would love to have exposure to it, would love to know the inner workings of what's going on.

1:55The latest is that there's going to be data centers in space. So what is the update on SpaceX that people should know? Great question. I'm very privileged to be a major investor in the company. And I'm just going to start with the caveat that I take that as a huge responsibility. And so I try not to share anything that I don't think is already public. but I'm pretty good at packaging what's public because I follow it very very closely

2:30look I personally believe SpaceX is the greatest company of all time I think the company is just getting started I and I have many other great portfolio companies but I'm willing to say SpaceX is on a different level of every company it's also just so fun to work on like launching rockets into space. One thing for me that's kind of been a lesson of watching how Elon builds companies is he, I think he focuses on the bottlenecks and the most important thing. And then once that starts to clear up, then he starts to focus on kind of what you can do with the thing. Examples are, you know, the company didn't really start working on Starlink until, like really it became a serious issue in November 2018, which was two years after they'd been doing reusable Falcon 9 flights.

3:22So my impression was he intuitively understood and the company understood that Falcon 9, they passed through this giant kind of challenging phase of getting Falcon 9 to be reusable, and they knew they were going to have a lot more launch capacity the next couple years. And so it was finally time to start thinking about what do we do with all this extra capacity, and that's when they got really serious about Starlink. I think that a few things happened around data centers in space in the last six months or so. I think it's very similar. The same way that you need a Falcon 9 reusable to become reliable, I think Elon and the company have gained a lot of confidence that Starship is going to be flying reliably and reusable this year in 2026.

4:11That's my strong sense. And very quickly, within two years or so, I think the company is basically going to have an excess capacity. They could launch all the just Starlink data satellites. They wanted all the direct-to-sell satellites they wanted, and they'd probably still have excess launch capacity in 2028 or 2029. And so on the one hand, you start thinking, what else should we be doing to be filling this capacity that's going to come in three years or so? And on the other hand, we all know what's happening in AI. We're limited by power, limited by regulation, which is probably even more restrictive.

4:54And kind of the combination of excess supply, launch supply coming online with Starship, married with this extreme urgent pain point on Earth, got the company thinking very seriously around, like, can we actually do data centers in space? Does the math pencil out? And, I mean, Elon has spoken about this quite a bit publicly. Like, the answer is yes. The public doesn't yet, like, the public hasn't yet gotten to see the full argument, like, hasn't gotten to see the math. But I'll just say, like, I think the math is really compelling. And I think data centers in space are, and I was on TBPN maybe six months ago saying, you know, I don't, I didn't say I don't believe in data centers in space.

5:38I said, like, it's the right macro. I'm not sure if it's the right micro. Like, I haven't done all the math. I can say I have now done the math, and I think it's going to be absolutely giant. Is that what that big algorithm was when we walked in? I don't even know which algorithm. You have a huge whiteboard over there. Crazy. Unfortunately, I don't, you know, I didn't use the whiteboard. That's not mine. That's someone else's. Sorcery is brought to you by Brex, the financial stack trusted by more than 30 ,000 companies, including one in three venture-backed startups in the U.S. Nearly 40 % of startups bail because they run out of cash.

6:15Rex is literally built to help founders avoid that. Unlike traditional banks that let your money sit idle, chipping away at it with fees, Rex's designs help you spend smarter and move faster. Their all-in-one solution combines checking, treasury, and FDIC protection into one powerful account. You can send and receive money globally at lightning speeds, get 20 times the standard FDIC coverage through their partner banks, and even high yield from day one. With same day and even same hour liquidity, access your funds anytime. Companies like Scale AI, DoorDash, Service Titan, HIMSS, Anthropic, Flexport, Robinhood, and Plaid trust and use Brex.

6:56Start today at brex.com slash sorcery. That's B-R-E-X dot com slash sorcery. So of those things, I also want to talk about Starlink. Could you break down that business for people who don't know? Sure. Starlink is an incredible business. You know, it's right now the kind of the first wave of Starlink was providing consumer internet from space, primarily in rural areas. So the way it works in simple terms is you launch a rocket. The rocket, like the initial V1 had about 60 Starlink satellites per launch. So you have these satellites in the sky that we've now kind of gone to two or three generations past that.

7:42We now have about 9 ,000 satellites in space. These satellites can talk to each other using laser links. So they're passing huge amounts of data, you know, terribus per second, back and forth in the constellation. And then you have this giant space infrastructure. and then on the ground people have user terminals, which is just a little dish that's 10 inches wide or 12 inches wide or so, that is what's called a phasory antenna and it's basically a bunch of radio waves that you can steer to basically lock onto a satellite and move from one satellite to the next as every 20 seconds or so as a satellite is moving out of view.

8:25All right. And so basically all it takes is these ground terminals at your house or on your car or whatever, or on an airplane now. You have these satellites in the sky. Data is moving from satellites to the ground and in space. And we kind of built the equivalent of all the long-range fiber optic cables that we had on Earth. So if you go back to 1999 or so, Global Crossing is a big deal. They laid fiber optic cable across the Atlantic Ocean, made a lot of money moving data. Now these fiber optic cable equivalents are in space. You don't need the cables, it's just lasers talking to each other in free space.

9:07And so it's really the best way in history to move data long distances. Starlink is now moving like a large fraction of the world's long distance data. And so anyways, the consumer business was Act One, then started to get into the enterprise business. Aviation was one of the first markets where that started to do well. I think aviation is going to, and so now we're seeing it. Qatar Airways was one of the first airlines. United signed up. JetSuite X was actually the first airline if I remember correctly and we're kind of getting this flywheel going now where when people get on an airplane and they use Starlink and they have this magical experience I think a lot of times they're realizing especially if you live in a rural area like Starlink internet on an airplane is way better than the internet than like the cable internet you have or dial up or whatever you know depending exactly where you live and so I think more and more people are getting awareness that way and will buy Starlink and And then now there's a new product, which is, well, and then the next product after that was the government product, Star Shield.

10:16And now kind of the newest product in the Starlink family is Direct2Cell. So this is cell service from space, pretty insane. It's, you know, and so Direct2Cell initially is for like the kind of most, technically it works everywhere, but SpaceX is limited with how much data can move into really dense places like dense cities. Starlink has a weird property that compared to terrestrial networks, like with the terrestrial network, when you build a cell tower, you have coverage just within a couple miles or whatever of the tower. With a Starlink satellite, it has coverage over everywhere it passes on its orbit.

11:02And so a lot of time it's passing over deserts or oceans or places that are not very populated. And so we have an excess bandwidth capacity in remote areas and a shortage in dense cities. And so Direct-to-Sell is initially working with many of the global carriers. T-Mobile in America was the first one. and when T-Mobile has bad cell service or low coverage, then they can kind of move on to the Starlink network and data will move directly from space to your phone. So basically these satellites are mimicking, they're spoofing cell towers and it's going to be an absolute, you know, SpaceX just announced that at the end of last year they had 9.2 million subscribers on the core Starlink consumer side and 6 million Direct to Sell subscribers.

11:57Direct to Sell is growing very quickly. And when we go like five years from now, all of my underwriting is, I think Direct to Sell will be probably even bigger than the original consumer internet business. And what are the economics behind that? Maybe even compare it to legacy businesses. Yeah, the economics are unbelievably good. There have been some very sophisticated analysts that have kind of put out their estimates. I may know too much here, but I will say it's incredibly good margins, incredibly strong unit economics.

12:35And it's very positively disruptive. The same way how Africa leapfrogged with cellular communications and instead of putting landlines everywhere they kind of went straight to cell phones. Starlink is really a leapfrogging technology. You don't need to have ground infrastructure. And so we're seeing schools in Ghana using Starlink so that kids can learn. But anyways, the unit economics are, I think it'll all come out in the S1, and the unit economics are incredibly good. In our interview with David Sutter, we propped up one of the hats with a red buzzer button and he pressed it whenever he didn't want to comment on something.

13:24I'll just have to get a no comment. But that was a good answer. Yeah, I'm trying to give you something. Maybe this could be a fun question. It's like the unintended consequences of what it would be that the impact of like unleashing internet and sell to the world. I mean, I at least notice it when I'm taking a flight. United, I think it's fake. But well, it's not very it's not very heavily penetrated right now. It's definitely not it takes, you know, I would say within I haven't seen the numbers in a while I think it's probably within nine months or so most United fleet will be upgraded But they install Starlink when a plane goes down for its like annual or biannual maintenance And so you kind of have to wait for the plane to hit its scheduled maintenance window I guess that's a good thing.

14:12Yeah, it's a good thing Turing is training the next generation of AI with tasks that require real expertise and real world judgment. That's why companies like NVIDIA, Anthropic, Salesforce and Gemini partner with Turing. Turing builds realistic reinforcement learning environments and data systems based on real operational traces. The kind of infrastructure frontier labs need to train superintelligence. Visit Turing.com slash S-O-U-R-C-E-R-Y. Well, what do you think the unintended consequences of like all this access would be? Yeah, I call this internet everywhere. Look, I think it's going to be insane.

14:52And I think it's going to feed into a lot of the other Elon companies. When you have a self-driving Tesla, now you pair that with Starlink and whether it's direct to sell or whether it's they put a dish on the roof of the car, which someday I don't know why they wouldn't embed you know like a dish directly in the car because then you just hit you have a bigger you know an optimized radio um i like imagine you're in your car and you have starting like and your car is driving yourself and you have perfect internet everywhere and your car starts to become you know your living room and you're doing your zoom calls or uh you know watching sporting match or talking to your grandkids or whatever it's just like we're going to to go to where we have internet everywhere.

15:40Optimus is going to use this. Military drones are going to use this. Kids in Africa are going to use this. Something that I think SpaceX has been very good about is whenever there's been a major natural disaster anywhere in the world, they're basically offering Starlink for free because all the terrestrial, usually both power and cellular communications get knocked out. And so anyways, like it's just, this is an incredibly resilient additional layer of data movement that candidly is better than all prior terrestrial approaches in many settings, not in all, like especially not in cities, but in like in most situations.

16:27It's pretty cool. within SpaceX itself what is what is a component that excites you a lot is it visiting Starbase like what is the what has it been like to be invested in the company well I'll just say like the vibes of the company right now are unbelievable like the team knows the team knows that it's on the precipice of history and like everything to now has just been a warm-up like it's really it's toy stuff compared to what's coming in the next five years. Everyone in the company understands this. The first thing is Starship. 2026 should be the year of Starship.

17:12I don't think people will fully appreciate in 2026 what it means. When you go five years out, it's really going to revolutionize a lot of industries. And then I think in roughly 2028, it's going to be the year of direct to cell. And SpaceX talked about this publicly, but it's going to take time. They just acquired a bunch of spectrum. It takes time to get antennas and phones optimized for the spectrum. And they need a bunch of V3 satellites in space. But 2028, it's going to be major cell service from space. I don't know exactly when the space data centers will start, but I think it's one of the biggest market opportunities ever.

18:02And then at some point, you know, at some point in the not-too-so-so future, like, they're going to land infrastructure on the moon and Mars. And, you know, then eventually land optimists on Mars and then humans. and like all this is coming faster than people think. So, and yeah, look, so I was, yesterday I was at SpaceX in Hawthorne and I'm going to Starbase in two days. It's just this kind of crazy thing. Like when you go to Hawthorne now, like the first time I went to Hawthorne, I felt like, oh my God, this is the most amazing place I've ever seen. It's a literal rocket factory. And now after going to Starbase, like Hawthorne feels like the tiny small little like old it's the you know like the Super Nintendo compared to like a modern like PC and it's just I wish other people could get the experience of seeing what SpaceX building the scale of infrastructure and then the last thing I'll say is just the amount of vertical integration that's happening in the company like I think it's It's probably the most vertically integrated American manufacturing company.

19:18And when you're vertically integrated, it just lets you move with a speed and flexibility that no one else has outside of China. And that's where things like space data centers can just happen very quickly, like shockingly quickly. Is SpaceX the first Elon company you've invested in? Yes. I first invested in 2019, you know, and in cumulative invested probably$1.2 billion. And across all the different funds, that position's worth like about$12 billion today. And, you know, in the$800 billion valuation. And so hopefully, hopefully in the IPO, it's worth a lot more than that. And, you know, but I really feel like it's just the very, very beginning of the company.

20:13And then I've been lucky after that to invest in most of the other companies. And I think they're all like I'm maximum bullish on all of Elon's companies. I think he is still extremely underappreciated. You think Elon's underrated? 100 percent. Not like he's I would say he's only gets like 10 percent of the appreciation he deserves. Some of you may not have heard this yet, but our sponsor, Public, just launched something called Generated Assets. and it brings AI into investing in a way I've honestly never seen before. Here's how it works. You type in an idea like AI-powered supply chain companies with positive free cash flow or defense tech companies growing revenue over 25 % year over year.

20:52Publix AI then dispatches a swarm of agents that scan every single US stock, evaluates them, and instantly builds a custom index around your thesis. What really stands out is how clearly it explains why each stock is included. And before you invest, you can even backtest your idea against the S &P 500. So you're making decisions with real context, not just guessing. And beyond generated assets, Public lets you invest in stocks, bonds, options, crypto, all in one place. They'll even give you an uncapped 1 % match when you transfer your investments over from another platform. If you want to build a portfolio that actually reflects your thesis, visit public.com slash sorcery.

21:28Paid for by Public Investing. Full disclosures in the description. What was the process to first get into SpaceX? Because I can't imagine that environment was at all as enthusiastic as this environment. No, look, I know this sounds crazy, but SpaceX in 2019 was a highly contrarian investment. And it was very hard. It was a very hard investment to convince people to do. And I didn't even really believe in it, you know, the year prior. and so the like the rough the read and I know it sounds crazy today but a few things one SpaceX in 2019 was just a launch company in 2019 they did 13 launches the total global launch market it depends how you count it depends how you count China some of the government launches but I would say in 2019 is probably like a five six billion dollar market like addressable market And at the time, the way people thought about the theoretical maximum EBITDA margin you'd have on a launch is like 40%.

22:37So it's about$2 billion of EBITDA. If you capture 100 % of the market, it's like$2 billion of EBITDA as a potential thing. And so then it requires a lot of capex investment and depreciating assets. And so probably have like a billion of profit if you capture 100 % of the market then. And the company's valued at$36 billion in 2019. And so like investing at 36 times the theoretical maximum earnings where you wouldn't be growing very much if you didn't find another market was like people just thought it was insane or stupid. And when I say people, there were some visionary investors that had invested very, very early you know I believe in giving credit where credit is due and you have people like you know Founders Fund and you know you have Valor and you have VY and you have a bunch of funds you have Steve Jurvetson from DFJ like you have people that really really understood the long-term potential way before I did but in 2019 it was pretty contrarian or not pretty like extremely contrarian and you know another thing was elon at the time like tesla when we first invested was about a 70 billion market cap company and so and it was you know a couple years like shortly after production hell and it just like the perception wasn't you know what it is today and for Starlink basically every prior internet constellation had failed and I think if I if I deserve credit for something it's that like with my background in physics and some in communications I was able to like understand that Starlink was actually going to work you know a year before it was working and to understand like what questions to ask to to really like underwrite that um and but you know so that was it there was just this weird moment in time where people were assigning basically zero credit to starlink because prior internet constellations had been failures and were assuming that the launch market was not going to grow and starship was like basically people were assigning zero value to starship um and so i don't know there was a weird i would say two-year period where it was underappreciated and then after Starlink got to about a billion of revenue which was very fast it was about 12 months after they got their first dollar of revenue they were at about a billion of run rate like I'd say at that point it became an extremely consensus investment so there was this weird two-year period where you could buy a lot of shares where people didn't get it and now it's like good luck finding shares no one wants to sell and you know it's going to be I'm very excited for the IPO it's going to be pretty incredible how do you think that like initial caution or I guess hesitancy on the business and having those early supporters has informed Elon in bringing in new investors because I feel like it's such a tight circle.

25:54Yeah, look, I, and that's part of why I am very careful with what I say, and I don't want to say anything that's not already public. I mean, I think to operate at the scale that Elon operates, he needs to, as much as possible, be able to trust the people around him and, you know, and, like, trust that people are going to, when you are as aggressive as he is, there's going to be times where things don't look as good and people need to kind of be patient during the hard times and very trustworthy with information for what's on the roadmap. And so, yes, it is a pretty small group of people and I'm lucky to have been able to invest in most of his companies.

26:44How did you get into the other companies? What was the process like? I mean, once I started working with SpaceX, I got to know other people. That's basically it. And really, like, the teams in the other companies are just as good. They're just kind of earlier in their development cycles. Which of the other companies do you think is most underrated? Honestly, I think they're all severely underrated right now. Like, I think that XAI is, like, severely underrated. And it makes sense. Like, its revenue scale compared to OpenAI, Anthropic, Google is obviously, like, much smaller. But it just kind of blows my mind that people don't understand how fast the company is moving.

27:37You just have to look at the derivatives of progress. And then also just look at the bottlenecks. I believe the bottleneck, at least for inference, is going to be power. And even if you have the best model in the world, if you can only deliver it to 1 % of the market or 2 % of the market demand, And you're heavily limited in your growth rate and expansion opportunities. I think XAI is going to, like Elon's the best in the world at atoms. And I think atoms are going to be a decisive factor in the AI race. And it just, the way Elon builds companies is different than other people. The way I describe it is he builds up potential energy and then he converts up potential energy into kinetic energy.

28:32Whereas a lot of other companies, whenever they have potential energy, they're just immediately converting it into kinetic energy. Elon builds gigafactories that are not producing any cars or revenue for like five years and then immediately start making a million cars. And the company spent a decade trying to get from almost a decade from Falcon 9 to Falcon 9 reusable. And then very quickly they get Starlink going. You know, now they've been spending almost a decade on Starship and it's like zero dollars of revenue from Starship today. But it's about to be an unbelievable amount. You know, like with XAI, they've been building, you know, the biggest coherent training clusters in the world.

29:16And they're just not really focused on monetizing them right now. And I just it's kind of crazy to me that people don't understand the pattern yet. And it's not to say that I'm bearish on the other foundation model companies. I just think XAI is going to shock people with the scale that it gets to quickly. But I'm bullish on all of them. I'm bullish on the Boring Company. I'm bullish on Neuralink. Tesla's a public company, so I have no comment on Tesla. But I'm literally insanely bullish on all of them. What are your thoughts on just the massive spend in the category? I think it's rational. Okay.

29:54A year ago, I probably would have been more bearish on the spend. I think that there's a lot of things that have happened that make me more bullish and think it's rational. One of which, ironically, I think the SpaceX IPO will actually help all private companies this year, including the AI companies, just because it creates a lot of liquidity. and it draws a lot of capital into tech. And it will just create a lot of liquidity that can go into the next wave of build-outs of AI companies. I also, Anthropic is rumored to go public in 2026. I think that that will also catalyze a lot more dollars into the category.

30:46and then you know very importantly I just think we've hit a point where the model like I think we have hit AGI you know and there's different definitions but I think we've hit the point where like I think we've definitely hit the point where in most fields of math or physics or computer science like the best models are better than the average PhD student at like a top five research program. And so it's just like, I think we've hit this point where the economic value is going to be so extreme. And we're seeing it right now in coding is probably the first real vertical. And then also just like the consumer kind of next generation Google equivalent, next generation search is ask whatever question, ask your, you know, like my son has eczema and like being able to ask questions about that and get a lot of ideas from the different models.

31:41Like it's just unbelievably useful and has replaced a lot of my behavior in many people on the planet. But I think the whole field is becoming more and more useful at such an extreme rate that it justifies the CapEx investments. That's my personal opinion. And hopefully lots of liquidity events to come. I think the liquidity events in 26 are going to be major and I think are very important for keeping the whole the party going yeah I just interviewed snowflakes ER CEO Sridhar and I asked him great guy great guy and he was so fun and we yeah it was great it was great um but they had one of the biggest software IPOs of all time so I asked him about this plug my partner Pat Grady and Carla Schenbach they led an amazing investment in snowflake and you know it was an incredible investment.

32:35They invested$100 million at a billion dollar valuation, you know, like a few years before IPO. But a question I asked him was, do you, like, what do you think on this immense private lockup? And do you think these companies should go public? And if you want to know the answer, you have to listen to the episode, obviously. But I'm curious from your standpoint. So as we move into, I also asked Vlad this as well, because I mean, he's a big retail So he's a big business guy. I actually took graduate math classes with Vlad at Stanford when we were kids. Really? Yeah, really. Like real analysis, Math 205, ABC with Brian White, this professor that wore koala shirts every day.

33:22Anyways, I've seen Vlad come a long way. And he's a freaking legend. So do you think most of these were past the stay private forever phase and now we're actually to the let's go public, this is good for you? Vlad was like, it's actually much better than you think. Yeah. Look, I just think it's very company and timing specific. And like for SpaceX, this data center opportunity came up. and it just creates an incredible catalyst to go public. And I think it's 100 % the right thing for the company, like given our needs. And then there's other companies. I mean, like there's other companies like Stripe where I think it makes sense for them to stay private right now.