Sam Altman is calling out Elon Musk's space data center narrative as investor theater disconnected from engineering reality. The OpenAI CEO spent the weekend telling public market investors that Musk is selling them a short-term story on orbital compute that the physics simply does not support, and space-compute engineers working on rival projects largely agree with him.

Why Is This Dispute About Timing So Important?

SpaceX's roughly $2 trillion valuation leans heavily on the promise that a fleet of orbital data centers will power artificial intelligence inference at scale. That thesis is the load-bearing wall under SpaceX's current stock price. If you strip the space-compute narrative out of the investment thesis, the valuation multiple compresses fast. Altman's weekend jab matters because it directly challenges one of the core financial assumptions driving SpaceX's public market valuation.

The exchange also reveals a genuine tension in the AI infrastructure market. OpenAI's compute demand is enormous, and every dollar the market allocates to orbital compute is a dollar not chasing terrestrial capacity that OpenAI can actually rent this quarter. Framing SpaceX's data center plans as short-term investor theater serves both the truth as most engineers see it and OpenAI's competitive interest, though those two motivations can coexist.

What Do Space-Compute Engineers Actually Say About the Timeline?

Talk to the people actually building orbital compute infrastructure, and the answer is consistent across the industry. Founders at rival space data center startups, engineers on Google's orbital compute project, and independent analysts running the launch-cost math all reach the same conclusion: scale space compute requires two things that do not yet exist at price.

• Reusable Rockets: Rockets that fly cheaply and repeatedly, with unit economics that make frequent launches economical. SpaceX's Starship is expected to fly for the 13th time as soon as July 16, but operational reusability is still years of iteration away.
• Mass-Manufactured Satellites: High-powered satellites that can be manufactured en masse and equipped with multi-megawatt power systems and thermal management in vacuum.
• Ground Infrastructure: Ground-station bandwidth sufficient to move inference traffic from orbit to users on Earth, with total cost of ownership competitive against terrestrial data centers.

Subject-matter experts peg the timeline for scale manufacturing and launch of orbital data center satellites to the 2030s, not the 2020s.

What Did SpaceX Tell IPO Investors About Reusability?

SpaceX itself conceded the reusability issue during its initial public offering road show, telling investors that Starship may not be fully reusable in the near-term and will need to expend each of its second stages on every flight. This is the unglamorous line buried in the prospectus that Altman's weekend post is effectively pointing at. Expending the second stage on every launch does not close the business case for economical orbital compute; it eliminates it entirely.

Musk's rejoinder, that the satellites start flying next year, is technically defensible but strategically thin. SpaceX can absolutely put a satellite equipped for high-speed data processing into orbit in 2026 or 2027 as a demonstration. The harder question is when the company can build and launch those satellites at the volume required for meaningful inference capacity. By consensus of the people doing the work, that is a 2030s question.

How to Evaluate Space Compute Claims as an Investor or Industry Observer

• Distinguish Demo From Scale: A single orbital data center satellite proves the concept works. Dozens or hundreds of satellites operating at full capacity to handle real inference workloads is a different engineering and manufacturing problem entirely.
• Follow the Launch Manifest: Even if Starship achieves full reusability tomorrow, NASA commitments and Starlink buildout will consume the launch manifest for years before orbital data centers get meaningful flight slots.
• Check the Unit Economics: Ask whether the cost per kilogram to orbit, combined with satellite manufacturing costs and ground infrastructure, actually beats the total cost of ownership for a terrestrial hyperscaler data center.
• Watch for Prospectus Caveats: When a company tells IPO investors that a key technology may not work on the timeline they are marketing to retail investors, that is the actual story, not the marketing pitch.

For the broader AI market, the read is that the ground game still wins for at least the next five years. Terrestrial hyperscaler buildout, including Meta's 5GW Hyperion data center project, ongoing high-bandwidth memory shortages from suppliers, and distributed compute experiments, is where inference capacity gets added between now and the end of the decade. Space compute is real, and eventually it will matter, but investors treating it as a near-term revenue line are pricing a 2030s option as if it ships next quarter.

The gap between what SpaceX is telling public market investors and what its engineers are telling each other is the actual story. Orbital compute is a credible long-term business with a roadmap to the 2030s, and SpaceX is the company most likely to execute it. The problem is the timeline mismatch between a public market pricing 2030s outcomes into 2026 revenue and a launch program that still has to demonstrate reusable second stages.