Trending
Prev Story
Next Story

Google’s Suncatcher satellite runs Gemma 4 in orbit and writes a haiku

The Next Web (TNW) •
November 01, 2026 • 02:35 PM
• By Ana-Maria Stanciuc
1
1 reads 0 reactions

“Blue marble below”: Google’s Suncatcher satellite sent back its first AI response from orbit, a haiku by its Gemma 4 model.

1 / 6
Swipe Photos

Google’s prototype data centre satellite has run an AI model in orbit and sent its answer back to Earth. The onboard Gemma 4 model wrote a haiku about being in space, Travis Beals, a senior director at Google, wrote in a blog post on Friday.

The satellite launched on 1 October for Project Suncatcher, Google’s research effort to run AI computing in space. Google called the haiku the first AI response from a data centre class TPU, its own AI chip, in orbit.

“Blue marble below / Data streams in silent grace / Starlight feeds my core.” Gemma 4, on the Suncatcher satellite, as published by Google

Google sent the prompt during a pass over a ground station. The poem came back as the satellite flew over Svalbard, Norway. The team then tried space-themed limericks and prompts asking the satellite to report on its own systems. The prompt asked for a haiku written as a TPU orbiting Earth, Thomas Claburn reported for The Register .

The satellite holds four TPUs and generates up to 1 kilowatt of power, Tim Fernholz reported for TechCrunch . Because of that power limit, it can run only one TPU at a time, for about 20 minutes. Testing its systems took about three weeks, and a Google spokesperson told TechCrunch the satellite was working normally.

Omar Sanseviero, who works on Gemma at Google DeepMind, wrote on X that the model ran on stored solar energy.

Google will now test how the TPUs cope with heat and radiation in orbit. Radiation can flip bits in a chip’s memory and corrupt data without warning. Google’s tests on the ground suggested this matters less for running models than for training them, according to TechCrunch.

If the chips hold up, Google plans to launch two linked satellites in 2027. They will make more power and test laser links between satellites, TechCrunch reported.

Google announced the launch in September. Startups also plan data centres in orbit, and Starcloud raised $250M in August. SpaceX has told investors its own orbital data centres may never work. Google’s plans also depend on SpaceX making its Starship rocket fully reusable, TechCrunch reported.

Get the most important tech news in your inbox each week.

Prev Swipe Story Next

Same Story from Other Publications

Compare how 6 other independent news outlets reported on this story.

6 Publications
Ars Technica
2d ago
Google DeepMind just mathematically verified key parts of Signal

Google DeepMind just mathematically verified key parts of Signal

The four-week project hints that AI could make formal verification far more practical.

The Verge
2d ago
Google is bringing Gmail’s “Help me write” AI tool to work accounts.

Google is bringing Gmail’s “Help me write” AI tool to work accounts.

Google says it’s rolling out the AI feature to all Workspace users 18 years and older over the next two weeks to help people manage their overflowing inboxes. It’s also expanding access to AI Overview conversation summaries to “synthesize key points and replies in the conversation.” [Link: Google Workspace Updates: Expanding Gemini in Gmail to more Google Workspace users | https://workspaceupdates.googleblog.com/2026/11/expanding-gemini-in-gmail-to-more-Google-Workspace-users.html | Workspace Updates Blog]

TechCrunch
3d ago
OpenAI rolls out new business profiles, making ChatGPT just a little bit more like Google | TechCrunch

OpenAI rolls out new business profiles, making ChatGPT just a little bit more like Google | TechCrunch

OpenAI is looking to make ChatGPT a more robust consumer discovery engine, and a host of new features seek to empower businesses to help it do that.

CNBC World
3d ago
Ex-Google recruiter: You're answering 'Tell me about yourself?' all wrong, and other job interview mistakes—do this instead

Ex-Google recruiter: You're answering 'Tell me about yourself?' all wrong, and other job interview mistakes—do this instead

Former Google recruiter Farah Sharghi shares her best advice for how to turn coffee chats into real introductions and interviews.

Wired
3d ago
A New AI Tool From Google DeepMind Proved the Signal Protocol Is Secure

A New AI Tool From Google DeepMind Proved the Signal Protocol Is Secure

DeepMind researchers were able to formally verify that the widely used encryption protocol works as intended. Their stretch goal? Do the same for “all the software of the world.”

The Guardian
4d ago
Parents and teachers decry Google pledge to yank Search if schools block AI: ‘This threat is nothing but extortion’

Parents and teachers decry Google pledge to yank Search if schools block AI: ‘This threat is nothing but extortion’

Company says schools should control student AI use, but if they block its AI tools, they’ll lose access to search engine

Related Stories

Technology
Build #2611