344
Productivity & Workflow355
Automation & Workflow225
Software Development251
Marketing & Growth193
AI Infrastructure & MLOps175
Writing & Content Creation204
Data & Analytics142
Photography & Imaging156
Design & Creative170
Customer Support133
Sales & Outreach125
Voice & Speech135
Education & Learning131
Operations & Admin87
Google launched a satellite to test one of its AI chips in space and says cheap space data centers would need many Starship launches first.
In short: Google launched a satellite to test whether one of its AI chips can reliably run in space, as part of a long-term plan for “data centers in orbit.”
Google launched a prototype “orbital compute satellite” on a SpaceX rocket from California. It is the first time Google has sent one of its advanced chips, called a TPU (Tensor Processing Unit, a computer chip built to run AI), into space.
The satellite was built by Planet Labs. Google says the goal is to prove the chip can work in orbit with enough power, cooling, and safety checks. Space is a harsh place for electronics, partly because of radiation (high-energy particles that can cause computer errors).
To reduce risk, the satellite will only turn on the TPU in 15-minute bursts. Google also says it is working on a future demo, planned for next year, using two satellites designed more specifically for this kind of computing. Those two satellites may connect using laser links (like a very fast, very precise beam of light used to send data).
Google also published updated research on the bigger idea: “space data centers” (think of a server farm, but floating above Earth). In that research, Google estimates launch costs would need to fall a lot. It suggests SpaceX’s Starship might need around 1,800 launches over about 10 years to deliver enough cargo to make costs drop toward about $200 per kilogram by 2035.
AI systems need lots of computing power, and that usually means large, expensive buildings full of machines on Earth. Google is exploring whether some of that computing could happen in space, but its own research shows the biggest barrier is still the cost and pace of rocket launches.
Source: TechCrunch AI