Musk vs. Bezos: Who Wins the New Space Race? — Josh Tyrangiel

I'm at The Atlantic's office in SoHo. It's a lovely place.

It's perfect. It's like one of those fourteen days a year where you say it's really great.

Yeah, I mean, it really started with a fundamental question: what had Jeff Bezos actually been doing over the past ten to twelve years? I mean, you know, turning Bezos into a meme about his girlfriend, his yacht, and his muscles, and the Met Gala was fun for a lot of people, but what was the substance of his decisions? So I started looking into that. And almost immediately, I knew he had invested a lot of money and run Blue Origin. What I didn't know was that his heart was really in it. I mean, not his mind. His mind is completely clear with Amazon and a lot of the other rational decisions he makes. But a lot of his decisions over the past decade and more, including buying The Washington Post and changing partners, were driven by this desire to reach space in a very specific way. So as I began writing the report, it became very clear that this was not a one-man story, and that there was someone else racing him to the same thing, albeit in very different ways, and that was Elon. And the ways in which they clash and compete with each other, but they have already become as powerful as many nation-states and have begun to define what our future in space looks like, clearly became the real story. It's a complicated story to tell for all the reasons I'm sure you know. But that was the beginning. It was more like, what is all this about? And when we found the answer, it became about what the next ten, twenty, or fifty years could look like. And that's a story about, you know, dominance among the stars.

I thought that was the case. Then, when I dug into the history of Blue Origin and SpaceX and how they were actually founded, a few things surprised me, okay? So the first is that, you know, Bezos was famous for being a hedge fund analyst at D. E. Shaw when the internet idea occurred to him, that it was growing at roughly 2,300% per year. And very rationally, he said, okay, how do you benefit from that? Well, you benefit from it by storing things that a regular physical store can't store. There were about three million print book titles in 1993 or '94 when he was making these decisions. So, it's books. It's not a hard decision. He may love books, but in the end he was making a decision to build Amazon based on his thinking. When it came to Blue Origin, this was what he loved more than anything else in his entire life. And if you go back to the reporting when Amazon started becoming something significant, his high school girlfriend said, you know, the reason behind his attempt to become rich was to reach space. And that truth was obscured for about 20 years until it became very clear that it was real. But he founded Blue Origin in a way that was the complete opposite of how he founded Amazon. So he kept Blue Origin secret for three years. He had very few employees. He put all his money, you know, a lot of surplus money, into it. The first employee was a man named Neal Stephenson, who is a very famous science fiction writer. And Stephenson wrote a blog post in which he mentioned that most of his work revolved around trying to find out if there was a different way to reach space, other than the way Robert Goddard and Wernher von Braun laid out, and everything rocket science considered for a century. And so, they stayed for three years in Kent, Washington. Blue Origin was basically just an experiment. It was a science-fiction-flavored startup. Whereas two years later, when SpaceX launched, it knew no mercy. Elon came back from that somewhat strange venture where he was trying to get intercontinental ballistic rockets to launch what was essentially a science-fiction publicity stunt. He wanted to send rockets to Mars with robots to gather some soil and grow a plant. That was his whole idea. And what he learned when the Russians literally tried to charge him $8 million on the return trip, he said: "No, I can do it cheaper." There's a problem with space. It's kind of like a monopoly. I'm going to basically treat it the way a 1970s utilities regulator would treat the telecom sector. And from the start, his operating principle was: I don't need a different way to get to space, I need a cheaper way to get there. And operationally, he staffed up at an incredible pace. He had enough funding to keep his team really excited and motivated to work. And he did something great leaders do, which is he delegated. So, even though he knew he had a specific vision for how to reach space, he hired Gwynne Shotwell, who is among the best executives in American business history. They divided the work and took over the situation. In the meantime, Bezos, you know, took about 15 years to get serious about space. And there are all sorts of reasons for that. But I think you could say that Elon was the better operator for a very long time.

Yeah, look, the way I describe Musk in the article is that his life is like a browser with a lot of open tabs, right? He's constantly switching between things that distract him, things that interest him, personal crises, professional crises, and random statements that consume his time. And clearly the X platform is the always-open tab, right? I mean, he's always giving his opinion on this or that. What I want to say is that people are a mix of many things, you know. He's very strange and complex. And one of the reasons for SpaceX's success is not just delegation, but the foundational work he did between 2003 and 2012 to build a very skilled team, not big, but highly capable. They know exactly what they're doing every day. They have very clear business goals and work leadership. You can't credit him with all of their achievements, but you can credit him with the delegation, right? And the way he delegated tasks made SpaceX very successful. I'll talk in a bit about how that fits into a future action plan, but I think both things are true; he's personally a mess. And professionally, especially when it comes to SpaceX, that team is really very good. And what I'll say about Jeff is that I think you're right. He's very disciplined. So disciplined that when he started Blue Origin, he set a very strict ceiling for hiring and for the spending he was willing to make, which may have been a mistake. I mean, it didn't lead to this kind of great experimentation. One of the people I spoke to for this article is a former senior NASA official who visited both companies when they were in their adolescence. This person said: "Look, SpaceX is a pirate ship, right?" I mean, there are a lot of young people wandering around it. And they're all dirty. And they all look like lunatics. But Blue Origin was more like an academic lab. They had huge 3D printers. They had a beautiful, shiny place, but there was no one there. And there was no energy. As Blue Origin employees told me at the time, they noticed that SpaceX was having more fun. When they launched things and they exploded, they laughed about it. And they posted it on social media. Meanwhile, Blue Origin was largely secretive. Now, there's a good reason for that, which is that Jeff Bezos was the CEO of Amazon. And he had obligations to shareholders to ensure that this was the place where he primarily spent his time. So he'd dedicate Wednesdays and many Saturdays to working at Blue Origin's offices. This was while Amazon was turning into what it is today. It wasn't hidden, but it wasn't advertised either. And his focus was largely on Amazon. And he was unable, if you want to criticize him for something, he never found the number two who could read his mind, right? A lot of founders make themselves indispensable by being the only ones who can interpret what they want. I think I've dealt with them, and you've dealt with them too. And while this is somewhat admirable, this kind of thinking doesn't scale. So what you see with Blue Origin is that despite having a lot of resources and brains, it was always stumbling because of the mismatch between what Jeff thought he wanted and how they executed it. Meanwhile, Elon was becoming the Elon we know, and he was tormenting Blue Origin, and Jeff specifically, by tweeting about it and making all the childish jokes Elon can make. And so a role reversal happened. And someone at Blue Origin told me in the early days: "Look, the irony is that Elon took Amazon's strategy and built SpaceX." The strategy was to expand very quickly and worry about everything else later. And if you have to push yourself to the brink of bankruptcy to break through, do it. Whereas Jeff was very gradual, and logical to an excessive degree in some respects. And then he put his heart into it so much that he couldn't see that his involvement might be holding back progress.

Yeah, look, I think you're referring to the IPO prospectus, the SpaceX IPO prospectus, which really is the first true blueprint for what could be called a space East India Company, right? And what it lays out, and you and I wrote this in the story, is that you have to think of the work that both Blue Origin and SpaceX envision as a kind of ladder, right? The bottom rung is the ability to launch into space. And right now, SpaceX launches 80% of the mass into space. The closest competitor is the Chinese government, and the gap is not close. It's amazing how much they dominate launch operations. Blue Origin has launch capability, but it's not entirely reliable. So, the first rung is launch capability. And it's also not very profitable. It's an extremely expensive business. There are only a limited number of customers. And so, what Elon realized around 2015-2016 is that he needs to create some kind of guaranteed demand, which you point to as the second rung, which is boring old telecommunications, right? Telecommunications is a really good business. It generates a lot of profit. And so, Starlink, which works very well, makes up a large part of the payload that SpaceX launches. And Blue Origin also realized that to scale its launch capability, it needs satellite capability. So they have their own satellite constellation they're working on. It has a contract as a provider for a quarter of Amazon's future satellite launch needs. So satellite launches are crucial. What's new in the past two years, despite being theoretical and controversial, is both companies' push to move forward in space computing. Okay? That means data centers in space. The reason you'd want to do this as a business is that the NIMBYism phenomenon and the political cost of building data centers on Earth are very high and competitive. If you're one of the few companies that has vertical integration in rocket and satellite launches, and if you can provide computing in space, you'll have an unconstrained means to charge others and yourself for access to this enormous computing resource that you have satellites broadcasting data to Earth. There's a lot of science that has to make this possible, but the people I spoke to say: "Look, we're 5 to 7 years away from making this a reality." Quick digression: if we're 5 to 7 years away from achieving this, it raises questions about the massive data centers being built around the world. What will happen to them if you can access computing in space? Now, if this seems crazy, and I know it does, I want to emphasize that every other AI company is scrambling to enter this arena. For example, OpenAI and Sam Altman considered buying a company called Stoke Space a year or two ago because they didn't want to be excluded from the future. And many people really believe that data centers in space are the future. There's a fourth rung on this ladder, and I'd like to say this is where things become more concerning and more theoretical. It's that if you have your own AI model, which both Jeff Bezos and Elon Musk have, and you can connect it to unconstrained computing, you'll really reach a distinguished position. Then you'll actually be able to train and run models, and even if you're not first in the AI race today, you could catch up very quickly if you can experiment on your models, train and tune them without worrying about energy or computing capacity. And so, this becomes a vertical integration of enormous power in the future. And this is the ultimate goal that both are racing to reach. There are some differences in how they present these businesses, and undoubtedly Elon Musk is way ahead right now. But as everyone in the space industry told me, this is a challenge that will take decades. Just being in first place now doesn't mean he'll end up there. But when you think about what it means, it means future power over a lot and a lot of different things. And I think the US government is just beginning to realize the fact that it has citizens capable of becoming supranational entities.

That's a great question. I'm not scientifically qualified to answer yes or no. But what I can tell you from my reporting is that every one of these companies is racing full speed to publish. Google is launching its first orbital data center project, I believe, in early 2027. And Musk said he'll launch his project soon. I spoke with the head of the lunar sustainability program at Blue Origin, a man named John Culbertson, who previously worked at SpaceX. And he told me: "Look, I was skeptical, but what I'll tell you is that this makes theoretical sense." You need a steady power source. In space, you have continuous solar power. You need a steady cooling source. Point the solar panels the other way and you'll get a temperature of minus 250. They're talking about the moon. The telecommunications infrastructure is already there. There's no local NIMBYism, and the number of data center projects already halted in 2026 is on the mind of every AI company. So when I hear about data centers on the moon or in space, I first think it's a great name for a band, but this won't happen, right? But the more people I talk to, the more I realize they're pushing toward it with all their financial might. They believe this is the way forward in computing, and that demand for it will be enormous. Meanwhile, I understand your point of view. I read the SpaceX prospectus and thought about two things at once. First, it's an interesting blueprint for the future of the space business, even though the total market

The target seems like it was invented on the fly. But I also thought that this seems like the product of an investment banker's weekend on drugs. It doesn't seem real to me. But the more I dug into it and spoke with independents, they kept confirming that it's real. We'll be living with this reality within 5 to 7 years. I also learned that deadlines in the space sector are complete nonsense. Not a single deadline has ever been met in the history of space launch. But what I can tell you is that they're pouring money into this project, and both are pouring money into their AI models because they believe this cycle is the cycle of the future.