3 min read
|
Saved February 14, 2026
|
Copied!
Do you care about this?
The article critiques the push for data centers in space, arguing that the immense costs and logistical challenges outweigh the benefits. It highlights the growing risks of satellite congestion and the competitive edge of ground-based energy sources, suggesting that such ventures are driven more by hype than feasibility.
If you do, here's more
SpaceX's acquisition of xAI and plans to launch data centers into space have raised eyebrows, especially given the immense scale required. To support frontier AI, companies need hundreds of thousands, if not millions, of GPUs. xAI’s Colossus cluster has 200,000 GPUs, while OpenAI aims for millions. Achieving this would necessitate launching an unprecedented number of satellites—far more than the 15,000 currently in orbit. Such a massive deployment heightens the risk of Kessler syndrome, where space debris could cripple access to orbit.
Even if rocket and satellite costs drop significantly by 2035, they'd still need to compete with the price of running ground-based AI servers at that time. Ground solar power has consistently become more affordable, reducing the appeal of space-based data centers. Despite the apparent impracticality, major companies are investing heavily. SpaceX's push for an IPO creates a need for hype, while xAI faces financial pressure. Investors can rationalize pouring money into a project they know is unlikely to succeed, banking on short-term gains before the reality sets in.
Questions about this article
No questions yet.