Sam Altman and Elon Musk engaged in a heated exchange on social media this past weekend, highlighting the disconnect between the aspirations and the current realities of the space computing industry.
In response to Musk’s allegations of deceit, Altman remarked, “you’re the one selling public market investors on short-term space data centers.”
By addressing Musk directly, Altman is echoing sentiments shared by many industry experts who feel that public investors may be overlooking a critical point: establishing space data centers as a viable business won’t happen any time soon.
SpaceX’s ambition to deploy a constellation of orbital data centers aimed at handling AI inference tasks significantly contributes to the company’s staggering two-trillion-dollar valuation. Optimistic analysts argue that this capability could enhance SpaceXAI’s systems or function as a novel orbital cloud, making it a key player in the ongoing AI revolution.
However, conversations with industry professionals—ranging from founders of competing space data center ventures, experts at Google working on their own orbital project, to engineers exploring feasibility—consistently reveal the same conclusion: substantial progress in this field is contingent on advancements in affordable rocket technology and cost-effective mass production of powerful satellites.
Musk is likely to respond by touting the potential of SpaceX’s Starship rocket, which is set for its thirteenth test launch soon. If his team achieves frequent and reliable flights, it could potentially strengthen the business case for data centers in space.
Nevertheless, even if the upcoming test flight is successful in recovering both rocket stages, operational reusability may remain years away. Additionally, the priorities of SpaceX, including contracts with NASA and the expansion of the Starlink network, could further delay the rollout of space data centers.
In their presentation during the IPO preparations, SpaceX acknowledged that Starship might not be entirely reusable in the immediate future, necessitating the disposal of each second stage after launch, thus hindering the economics of space data centers.
This context makes Musk’s bold claim that “we start flying them next year” seem less convincing. While there’s no doubt SpaceX could launch a high-speed data processing satellite in the next year, the critical issue is the timeline for launching and manufacturing these at a large scale, which appears to push expectations into the 2030s.
