The comments came during his speech at the India AI Impact Summit 2026 in New Delhi, where he elaborated on the logistical and economic barriers that such projects will face. He was quoted by The Indian Express as saying.
Why the idea has not yet taken off
The initiative to create orbital data centers is being actively promoted mainly by Ilon Musk, the founder of SpaceX and xAI.
Musk has repeatedly talked about the prospect of taking the computing infrastructure for AI into space, where he believes the centers will be able to use solar energy, avoiding the environmental and infrastructural limitations of ground-based data centers (DCs). The talks even included plans to create a constellation of up to a million data center satellites that could ensure continuous operation of computing power.
However, Sam Altman emphasized that such plans face fundamental obstacles. “Frankly, the idea of putting data centers in space at the current level of development seems ridiculous. Orbital data centers will not matter on the scale of this decade.”
Altman noted that the main problem today is launch and maintenance costs: the high cost of putting equipment into low-Earth orbit, the difficulty of repairing and replacing components (e.g., GPUs), and the operational risks of the space environment make such projects uneconomical compared to developing ground-based infrastructure.
As specific benchmarks, the head of OpenAI and industry analysts cite launch costs: on average, about $5.6 million per 800 kg payload into low-Earth orbit using modern rocket systems, which is many times higher than the cost of installing equivalent resources on Earth.
The time has not yet come
Nevertheless, he does not rule out that in the longer term (perhaps beyond the current decade) such projects may make sense when launch costs come down and space servicing technologies become available. “All in good time,” Altman added.
Altman’s comment reflects a broader discussion in the industry, where ambitious space initiatives face the realities of economics and engineering, and the largest AI developers are still focusing on expanding ground-based data centers, highlights The Indian Express.
