The real reasons Elon Musk merged xAI and SpaceX
Elon Musk just created the world’s most valuable private company. And he didn’t do it through rapid growth or a new product launch — at least not directly, anyway.
Instead, as reported this week, Musk merged his artificial intelligence startup xAI into his wildly successful rocket company, SpaceX.
Combined together, the two companies are now valued at an estimated $1.25 trillion. It’s the biggest merger in history. And because Musk controls both companies, he calls most of the shots when it comes to the deal.
A sci-fi twist
At first glance, the connection between rockets and AI seems tenuous at best. But dig deeper into Musk’s big picture goals, and the merger starts to make a lot more sense — even if there’s a decidedly sci-fi twist.
SpaceX has made a name for itself by building gigantic, reusable rockets that deliver satellites into orbit for cheap. The company also delivers people and cargo to the International Space Station on behalf of NASA.
That’s a lucrative business. SpaceX’s rockets are now America’s main method of getting things into orbit, and its cheap satellites have fueled the success of Starlink, Musk’s space-based Internet service. Fully 95% of the things America launches into space are now put there by SpaceX.
Simultaneously, Musk’s xAI has been hard at work building Large Language Models, like its core Grok model. Although xAI isn’t as well known or widely used as dominant players like OpenAI, its models still perform well in industry benchmarks, putting the company on the Large Language Model leaderboard.
Training models is expensive, though, not least because of the cost of electricity, and the challenges of finding room in data centers here on planet earth. That challenge likely hints at Musk’s deeper reason for merging his two companies.
Musk has previously pushed for the idea of launching data centers into space, a long-held, sci-fi-escque dream of his. This sounds outlandish, but it’s becoming a surprisingly mainstream concept.
Computers on satellites in orbit would benefit from plentiful, free solar energy. They could also potentially cool their chips by transferring heat into space, avoiding the insane power (and water) usage of terrestrial data centers.
The lack of cooling equipment and grid infrastructure means these orbital data centers could be smaller than those on earth. And they wouldn’t need to take up valuable real estate here on the ground.
By beaming their data back to earth, a constellation of data center satellites could greatly reduce the cost of training and operating Large Language Models. That could give a third-tier LLM company like Grok a huge advantage over its competitors.
Musk may also have an easier time recruiting talent for the well-respected SpaceX than for xAI. And he could use lucrative government contracts for orbital launches to fund AI development.
All of this will take time to develop, of course. But given Musk’s track record (for engineering at least, if perhaps not social network administration), the idea of flying data centers could come to fruition sooner than imagined.
When Musk said he would build reusable rockets that could land themselves upright, people mocked him. Today, that’s a key part of what makes SpaceX successful, and it’s being widely copied by companies and governments.
The same rapid development cycle could apply to orbital supercomputers, too.
In the short term, there are other advantages of merging the companies. Starlink customers will likely see more AI tools built into their Internet subscriptions.
Musk might also be planning to build more AI into his government contracts, including those in the defense space. Companies like Palantir make billions by selling AI services in the defense sector. Musk may be looking to use his existing SpaceX connections to get in on the opportunity.
Not a done deal
The deal isn’t officially done yet. Regulators could still balk at the idea of creating a mega company at Musk’s desired scale. And because the X social network sits under the xAI umbrella, concerns about Musk’s control of both information and access to space could crater the deal on national security grounds.
Still, assuming the merger goes ahead, Musk could have an unprecedented level of control over two of the 21st century’s most promising technologies . And, he would have an unprecedented ability to combine those technologies together.