Google Just Sent Gemma 4 Into Space and It Answered Back
Google just sent Gemma 4 into space, and the AI answered back from an orbiting TPU. Project Suncatcher marks a first for Google's data center chips, but its real goal is far bigger: solar-powered AI data centers built from laser-linked satellites. Meanwhile, Nvidia already has chips in orbit, and Google faces massive challenges with heat, radiation, and launch costs. 📩 Brand Deals & Partnerships: collabs@nouralabs.com ✉️ General Inquiries: airevolutionofficial@gmail.com 🔍 What You'll See Google's Gemma 4 sends its first reply from orbit on a data center TPU. SOURCE: https://blog.google/innovation-and-ai/models-and-research/google-research/project-suncatcher-satellite-response/ Google's satellite carries 4 TPUs, but only 1 runs for 20 minutes at a time. SOURCE: https://techcrunch.com/2026/10/30/googles-first-data-center-satellite-is-online-and-writing-poetry/ Google proposes 81 satellites connected by lasers to build huge AI clusters above Earth. SOURCE: https://doi.org/10.1016/j.joule.2026.102678 Starcloud runs Google's Gemma using Nvidia's H100 in space, beating Google to orbit. SOURCE: https://www.datacenterdynamics.com/en/news/starcloud-runs-ai-model-in-space/ 🚨 Why It Matters AI computing could eventually move beyond Earth, tapping near-constant sunlight to power satellite networks. But cooling chips in a vacuum, surviving radiation, keeping satellites connected, and launching everything cheaply could make the whole idea impractical. Timestamps: 0:00 Google's AI chip in orbit writes a haiku 1:19 How it got to orbit 2:12 A very modest setup 3:20 The race: Starcloud, SpaceX and Nvidia 4:11 Why put data centers in space 4:42 A swarm of 81 satellites 7:13 The skeptics 8:02 Heat and radiation problems 10:59 The money: does it pencil out? 11:51 Google, SpaceX and Wall Street #google #gemma4 #spaceai
