The first public earnings call for a rocket company just turned into a robotics pitch, and Wall Street's not sure if they bought shares in SpaceX or the next Boston Dynamics.
The Summary
- SpaceX posted revenue above analyst expectations in its debut public earnings, but shares tanked on surging capital expenditures that signal a major strategic pivot.
- Musk unveiled plans for lunar robots and AI systems while defending his $1 trillion revenue target — a projection that would require SpaceX to grow faster than Amazon did in its first two decades.
- The real story isn't rockets anymore. It's autonomous systems for space, and the capex spike shows Musk is betting the company on it.
The Signal
SpaceX's first earnings as a public company revealed a classic Musk pattern: deliver on today's numbers, then immediately shift the conversation to tomorrow's moonshot. Revenue beat the street. Operating margins held. But capital expenditures jumped 47% quarter-over-quarter, and suddenly this wasn't a rocket company anymore. It was an AI infrastructure play that happens to own launch vehicles.
The earnings call spent less time on Starship than on what Musk called "autonomous lunar infrastructure." He described plans for robotic systems that would operate on the moon's surface, building structures and mining resources without human supervision. He used the phrase "totally nuts" twice — once to describe the ambition, once to dismiss analyst questions about near-term profitability.
"The $1 trillion revenue target isn't about rockets. It's about selling the operating system for off-world commerce."
Here's what the capex spike is funding:
- AI training infrastructure for autonomous space systems
- Robotics R&D for lunar and Mars surface operations
- Ground-based compute clusters to control fleets of remote robots with light-speed delay
Investors reacted by selling the stock, down 8% in after-hours trading. The market wanted a mature space logistics company with predictable cash flows. Musk delivered a pitch for becoming the AWS of extraterrestrial automation. Wall Street hates pivots. But this isn't really a pivot. It's the plan Musk's been telegraphing since he started talking about Mars colonies a decade ago. You can't send a million people to another planet. You can send robots first, then follow with people once the infrastructure exists.
The $1 trillion revenue target stops sounding insane when you frame it that way. SpaceX isn't competing with Boeing or Lockheed. It's building the backbone for a new category: remote autonomous systems operating beyond Earth's orbit. That's not a space company. That's a robotics and AI company that solved the launch problem first because you can't deploy robots on the moon if you can't get them there cheaply.
The Implication
If Musk delivers even half of what he described, SpaceX stops being a launch provider and becomes the largest customer for its own launches. Every robot, every sensor, every compute node on the lunar surface needs to get there on a Starship. The business model shifts from selling rides to selling the destination. Watch how fast NASA and commercial space companies start licensing SpaceX's autonomous systems. That's where the margin lives, not in the rocket engines.
For AI companies, this is the clearest signal yet that the next frontier for agents isn't better chatbots. It's systems that operate in environments where humans can't give real-time feedback. Train a model to build a habitat on Mars with 20-minute communication lag, and you've solved autonomy in a way that makes self-driving cars look trivial. The capex isn't crazy. It's early.