344
Productivity & Workflow355
Automation & Workflow224
Software Development251
Marketing & Growth192
AI Infrastructure & MLOps174
Writing & Content Creation203
Data & Analytics141
Design & Creative170
Photography & Imaging156
Customer Support131
Sales & Outreach125
Voice & Speech135
Education & Learning131
Operations & Admin87
A public back-and-forth between Sam Altman and Elon Musk is drawing attention to doubts about whether data centers in space will be practical soon.
In short: A social media argument between Sam Altman and Elon Musk is putting a spotlight on expert doubts about “data centers in space.”
Sam Altman and Elon Musk traded sharp posts online over the weekend. Musk accused Altman of being a scammer, and Altman responded by saying Musk is the one selling investors on “short-term space datacenters.”
The argument points to a bigger debate about space data centers. A data center is basically a warehouse full of computers that runs online services. A “space data center” would put some of those computers in satellites in orbit, like moving a server room into space.
According to TechCrunch, many experts think this will not be a serious business anytime soon. The main reason is cost. You would need much cheaper rocket launches and a way to build powerful satellites in large numbers at low cost, not just a few one-off machines.
SpaceX has talked about using orbital data centers to do “AI inference” (when an AI system uses what it already learned to answer questions or process requests, like a calculator doing a new problem). But even if SpaceX can launch a capable satellite next year, experts question when it could do it at large scale.
SpaceX’s Starship rocket is expected to attempt another test flight soon, and Musk argues that frequent, reusable flights could make the idea affordable. Still, the report notes that fully reusable operation could be years away, and SpaceX also has other priorities like NASA missions and its Starlink internet network. For now, the key question is whether launches and satellite manufacturing can become cheap and repeatable enough, and that may not be answered until the 2030s.
Source: TechCrunch AI