SpaceX CEO Elon Musk said the company’s planned orbital AI data centers will rely mostly on technology already developed for Starlink V3 satellites, as investors assess SpaceX’s AI infrastructure ambitions ahead of its blockbuster public listing.
TL;DR
- Musk said orbital AI satellites do not need “magic” technology and will use much of SpaceX’s existing Starlink V3 work.
- The first AI satellite is expected to deliver 150 kilowatts of peak power and 120 kilowatts of sustained compute power.
- SpaceX expects its Bastrop, Texas, AI satellite factory to reach meaningful production volumes by the end of 2027.
- The company is pitching orbital computing as a way to address power limits facing Earth-based AI data centers.
SpaceX is leaning harder into artificial intelligence infrastructure as it prepares for one of the most closely watched IPOs in technology history.
CEO Elon Musk said the company’s planned AI satellites, designed to function as computing nodes in orbit, will use mostly technology that SpaceX has already built for its Starlink satellite network. The comments came during a company video discussion released ahead of SpaceX’s expected public listing, which Reuters reported is targeting a valuation of around $1.75 trillion.
“Part of what we want to convey here is that there is not some magic that is necessary, that doesn't exist,” Musk said, according to Reuters. “A lot of this is technology we've already made for the Starlink V3 satellites. We don't think this is a super hard problem compared to the things we already do.”
The company’s orbital AI data center plan is built around satellites that would use solar energy for power and radiate heat into space for cooling. SpaceX argues this could help address one of the biggest problems facing terrestrial AI infrastructure, mainly power availability.
According to the presentation cited by Reuters, the first proposed AI satellite would generate around 150 kilowatts of peak power and 120 kilowatts of sustained compute power. Musk said that would be roughly comparable to a single Nvidia GB300 AI server rack, which typically consumes around 140 kilowatts at peak power.
SpaceX engineer Ian Dahl said the satellites could be simpler than broadband-focused Starlink spacecraft because they would not need the large phased-array antennas required for communications services. Instead, the AI satellites would rely on technologies already used or planned for Starlink V3, including solar arrays and thermal-management systems.
The company also pointed to Starship as a key enabler of the project. SpaceX said the rocket’s fully reusable design could eventually allow it to launch the large volumes of solar panels, radiators and chips needed to scale orbital computing infrastructure.
Topics For More Insights
Musk said SpaceX expects its AI satellite factory in Bastrop, Texas, to reach meaningful production volumes by the end of next year, which would place that milestone around the end of 2027. Reuters separately reported that SpaceX is looking to begin orbital AI computing demonstration missions by late 2027, ahead of an earlier “as early as 2028” timeline.
The timing is important. Reuters reported that SpaceX’s IPO demand is approaching four times oversubscribed, with investor orders surpassing $250 billion against a planned $75 billion raise. The company has been marketing the offering by highlighting its launch business, Starlink’s global network, and a longer-term pitch around space-based AI infrastructure.
However, the same ambition is also creating investor scrutiny. Reuters noted that SpaceX’s proposed valuation would put a significant “Elon premium” to the test, especially as the company’s future outlook depends partly on markets that are still largely unproven, including orbital AI data centers.
For now, SpaceX is trying to frame orbital AI computing as an extension of its existing capabilities, not a distant science-fiction bet.
The larger question for investors is whether existing Starlink and Starship technology can turn space-based AI infrastructure into a scalable business, or whether it remains one of the boldest assumptions behind SpaceX’s public-market debut.
