Understand this story
On October 30, 2026, Google announced that it had received the first AI response from the Project Suncatcher prototype satellite. The satellite's in-orbit data-center-class TPU successfully ran the Gemma 4 open model and returned a haiku about space. According to the Google Blog, the satellite was launched on October 1 and had previously completed power-on and early machine learning tests on the host and TPU trays. IT Home reported that the satellite carries four TPUs, can generate up to 1 kilowatt of power, and, limited by heat dissipation and energy load, can only be switched on for about 20 minutes at a time. Google will next test how the TPUs perform under orbital thermal stress and radiation exposure, paving the way for the goal of two interconnected satellites in 2027.
AI-generated from public reports
Latest development · Report
Google's in-orbit TPU completes AI inference for the first time, Gemma 4 generates a haikuDevelopments
1 confirmed developments
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Oct 31
English text is being prepared.
Google's in-orbit TPU completes AI inference for the first time, Gemma 4 generates a haiku
Google announced that a data-center-class TPU running in orbit aboard its "Project Suncatcher" prototype satellite has for the first time produced an AI response; the pre-installed DeepMind model Gemma 4, following a prompt from ground engineers, composed a haiku about being in space. The satellite carries four TPUs and can generate up to 1 kilowatt of electrical power; limited by heat dissipation and energy load, it can only be powered on for about 20 minutes at a time, and the test focuses on verifying the chip's stability and error-correction capability under microgravity, severe temperature swings and the space radiation environment.
IT Home · AIRelated discussion
IT Home · AI joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
TMTPost · AGI
Edge AI Daily Morning Brief: Alphabet stock falls 12.6% after earnings; NVIDIA open-sources Nemotron 3.5…
Alphabet reported third-quarter 2026 earnings per share of $2.12, far below Wall Street's expected $3.02; its stock closed down 12.6% after the earnings report, with market value evaporating nearly $500 billion in a single day. NVIDIA opened the weights of the Nemotron 3.5 Transcribe model on October 28, 2026; it uses a MoE architecture with about 34 billion total parameters, activating only about 4.3 billion per forward pass, supports transcription in 40 languages, and can distinguish up to 8 speakers. Since June 2026, Anthropic has run a "self-evaluation" safety mechanism in production that can catch 78% of violating conversations flagged by a heavy prompt classifier, with a 94% detection rate for high-risk abuse, at an additional inference compute cost of less than 2%.
David OndrejRelated discussion
David Ondrej joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
cnBeta
Google's first data center satellite goes online, writes haiku in orbit using Gemma 4
After about three weeks of in-orbit testing and calibration, Google's first data center satellite carrying a TPU has entered its formal operational phase. This week, engineers sent prompt instructions to the satellite for the first time, and the pre-installed Google DeepMind Gemma 4 model composed a haiku about space. The satellite can generate up to about 1 kilowatt of power; because the TPU has relatively high power consumption, the chip can only be powered on for about 20 minutes at a time. Google will next evaluate the TPU's reliability under temperature changes and cosmic radiation. If it is verified to be stable, the company plans to launch two more follow-up satellites in 2027 using more advanced chips.
Ayi AI Notes
TPU on Google Satellite Completes First Space AI Inference and Returns Results
Google's Project Suncatcher prototype satellite booted up a TPU in low Earth orbit, received a ground command to complete a model inference, and transmitted an AI-generated haiku back to Earth. Sundar Pichai called this the project's hello world milestone. The author believes that space-based solar power and deep-space radiative cooling can bypass the power and cooling bottlenecks of ground-based data centers.
Philipp Schmid
Philipp Schmid relays that Google runs Gemma 4 on a TPU in space
Philipp Schmid relayed Sundar Pichai's message, stating that Google's Project Suncatcher prototype satellite booted up a TPU in orbit, ran a prompt, and a complete response was received on Earth—a haiku. On that basis, he said Gemma 4 became the first LLM to run in space, but the original post did not name the model used.
The Verge AIRelated discussion
The Verge AI joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Carlos E. PerezRelated discussion
Carlos E. Perez joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Riley GoodsideRelated discussion
Riley Goodside joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Unite.AI
Google Project Suncatcher prototype satellite's on-orbit TPU returns a haiku generated by Gemma 4 for the first time
On October 30, Google said that the Gemma 4 open model running on the on-orbit TPU of the Project Suncatcher prototype satellite returned its first AI response, a haiku about being in space, and the team subsequently tested a space-themed limerick and telemetry prompts reporting payload status. The satellite was launched on October 1 on the SpaceX Transporter-18 rideshare mission, was built by Google in collaboration with Planet, and the mission goal is to collect on-orbit data on the TPU's performance under the physical stresses of spaceflight as well as radiation and extreme temperatures. Google plans to launch two interconnected satellites in 2027 to test inter-satellite laser links.
Rohan Paul
Google's in-orbit TPU completes first prompt inference from Earth
Google's Project Suncatcher prototype satellite booted up a TPU in orbit, received a command from Earth to run a prompt, and transmitted a complete haiku back to Earth. Sundar Pichai called it a hello world milestone for the project; the satellite launched several weeks ago.
Testing Catalog
Google Achieves Prompt Calls to TPU on Satellite in Project Suncatcher
Google reports that in Project Suncatcher it can now send prompts to a TPU running on a satellite. The news was relayed by TestingCatalog; the original article did not disclose the satellite model, TPU specifications, or details of the call results.
Elon MuskRelated discussion
Elon Musk joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Sundar Pichai
Google Project Suncatcher prototype satellite completes first prompt inference on in-orbit TPU
Google's Project Suncatcher prototype satellite started a TPU in orbit, ran a prompt and transmitted its first complete reply back to Earth, which was a haiku. Sundar Pichai called this a hello world milestone for the project; the satellite was launched a few weeks ago.
Oct 30
Google BlogFirst-partySelected
Google Receives First AI Reply from Project Suncatcher Prototype Satellite's On-Orbit TPU
Google received the first reply from the Project Suncatcher prototype satellite, with the on-orbit data-center-class TPU successfully running the open Gemma 4 model and returning a haiku about space. The satellite was launched on October 1, and the host and TPU tray have previously completed power-on and early machine learning tests. Next, Google will test how the TPU performs under orbital thermal stress and radiation exposure, paving the way for the 2027 two-satellite interconnection goal.
Heat trend for this story
- Current
- 52
- Period peak
- 103 Oct 31 06:00
- 24-hour change
- -45%
Move the pointer or select the chart to inspect hourly heat. Use the left and right arrow keys with a keyboard.