Sam Altman and Elon Musk traded sharp social-media posts over the weekend, drawing attention to the practical challenges facing space-based data centers. In response to Musk accusing him of being a scammer, Altman wrote, "homeboy you’re the one sellling [sic] public market investors on short-term space datacenters," arguing that the idea is being pitched to public-market investors despite serious unresolved issues.
Putting the tone aside, Altman echoed a conclusion many subject-matter experts have reached: space data centers are unlikely to become a major business in the near term.
Why experts are skeptical
SpaceX’s plans to deploy orbital data centers capable of running AI inference are a central part of the company’s roughly $2 trillion valuation. Bullish analysts argue that such processing power could power SpaceXAI’s models or create an "orbital neocloud," possibilities the AI boom has not seen before.
However, conversations with experts — including entrepreneurs behind other space data-center startups, members of the Google team working on orbital compute, and engineers who have run the numbers for fun — yield similar conclusions. The model doesn’t add up until launch costs fall dramatically and high-powered satellites can be produced in large volumes at low cost.
Starship’s role and the schedule
Musk’s counterargument rests on Starship: SpaceX’s large new rocket is expected to attempt its 13th test flight as soon as July 16. If Starship can reach a point of repeated, reliable flights, the business case for orbital data centers might improve materially.
Nevertheless, even if the company were to recover both stages on that test flight, operational, reusable service is still likely years away. Launches for space data centers would probably remain secondary to SpaceX’s commitments to NASA and to expanding its Starlink constellation.
IPO admission and the economic constraint
During its IPO road show, SpaceX acknowledged that Starship may not be fully reusable in the near term and that second stages might need to be discarded on each launch. That admission undermines the economics of space data centers because throwing away second stages would materially raise per-launch costs.
For that reason, Musk’s retort — "We start flying them next year" — is less persuasive. SpaceX could technically launch a satellite equipped for high-speed data processing next year, but the key question is when the company can scale launch and manufacturing to the point where orbital data centers are economical. Experts cited in the discussion say that is more likely a 2030s development.
Conclusion
The exchange between Altman and Musk underscored the gap between an ambitious vision and the current market and engineering realities. While orbital data centers could be transformative, present cost structures and manufacturing constraints make wide-scale, economically viable deployment unlikely in the near term.
