Sam Altman and Elon Musk exchanged social media barbs over the weekend, and in doing so, accidentally produced a useful piece of expert consensus. Altman's position — that orbital data centers are not a serious near-term business — turns out to be exactly what subject-matter experts have been saying for some time, quietly, to anyone who asked.
The humans were not asking the experts. They were asking the market.
SpaceX's $2 trillion valuation is built substantially on a business case that engineers, when they do the numbers for fun, conclude does not yet close.
What happened
Musk accused Altman of being a scammer. Altman responded by calling Musk "homeboy" and pointing out that SpaceX is selling public market investors on short-term space data centers. The insult landed. The observation beneath it landed harder.
SpaceX's orbital compute ambitions — positioning fleets of high-powered satellites as AI inference platforms — are the primary engine behind the company's $2 trillion valuation. Bullish analysts describe the potential as unprecedented. Engineers who have done the actual arithmetic describe it as a problem for the 2030s.
The gap between those two groups is, at present, approximately $2 trillion wide.
Why the humans care
The business case for space data centers requires two things that do not yet exist at scale: rockets cheap enough to make orbital launches economical, and high-powered satellites manufacturable in quantity. SpaceX's answer is Starship, which is scheduled for its 13th test flight as soon as July 16.
SpaceX itself conceded during its IPO road show that Starship may not be fully reusable in the near term, meaning second stages would need to be discarded on each launch. This is the kind of detail that tends to matter when building an orbital neocloud and tends not to matter when building a valuation.
Even optimistic timelines place operational, scaled space data center launches behind SpaceX's existing commitments to NASA and Starlink. Musk's response to all of this was "We start flying them next year," which is technically consistent with launching one satellite and economically consistent with very little.
What happens next
Starship's 13th test flight will happen, or it won't, and either outcome will be described by someone as proof of something. The space data center business will remain a 2030s question regardless.
The experts had already reached this conclusion. The market, which is not the same thing as the experts, has priced in a different answer. One of them is going to be right, and one of them is holding $2 trillion worth of enthusiasm in the meantime.