# Google ran Gemma 4 on a TPU in orbit and got a haiku back (slim) > On 30 October 2026 Google said its Project Suncatcher prototype satellite ran Gemma 4 on a TPU in orbit and sent back a haiku. Vlad Cealicu on what happened, why the tokens are the quick bit and the ground-station pass is the slow one, why the memory holding the weights is the part to watch, heat, the money, who regulates a data centre in orbit, and what the haiku says about how a model writes. - Full: https://www.anchorterminal.com/blog/suncatcher-haiku-from-orbit.md (~3,600 tokens) · this version ~1,680 tokens · JSON https://www.anchorterminal.com/blog/suncatcher-haiku-from-orbit.json · canonical https://www.anchorterminal.com/blog/suncatcher-haiku-from-orbit - Index: https://www.anchorterminal.com/llms.txt · API: https://www.anchorterminal.com/api/v1/index.json · MCP: https://www.anchorterminal.com/mcp · Updated: 2026-10-30 - Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data) By Vlad Cealicu, 30 October 2026. 1,689 words, about 7 min to read. At 6pm on Friday 30/10/26 I was halfway through putting a view counter in the footer of every page on Anchor Terminal (most of the work was a small dot that pulses when someone's reading) when Google posted that one of its TPUs had written a haiku in orbit. > Blue marble below > Data streams in silent grace > Starlight feeds my core. ✨ Gemma 4, on a TPU in orbit, as Google published it [[1]](#ref-1) Google calls it the project's hello world, "the first-ever AI response from a data center class TPU in orbit" [[1]](#ref-1). - 4: TPUs on board - 15 min: the model runs at a time, because of heat - 1/10/26: launched on SpaceX's Transporter-18 - 17: syllables in its first answer The prototype, as Google and NPR describe it. ## What happened The satellite went up on 1/10/26 on SpaceX's Transporter-18 rideshare, built with Planet, and the team had contact the same day [[3]](#ref-3). It isn't the first model to talk from orbit. from a paper to a haiku, with what's planned 1. **November 2025** (paper). Google announces Suncatcher, with radiation tests on a Trillium TPU [[5]](#ref-5) 2. **December 2025** (h100). Starcloud's H100 trains nanoGPT and runs Gemma in orbit [[4]](#ref-4) 3. **1/10/26** (launch). the Suncatcher prototype goes up on Transporter-18 [[3]](#ref-3) 4. **30/10/26** (haiku). Gemma 4 answers its first prompt from orbit [[1]](#ref-1) 5. **2027** (planned). two linked satellites to test the laser links [[1]](#ref-1) 6. **Mid-2030s** (planned). launch under $200/kg, where Google's sums start to work [[5]](#ref-5) ## Seventeen syllables is about twenty tokens A model writes a haiku one token at a time, and for every token it reads its weights out of memory again. The rest of the round trip is orbital. That shapes what compute up there is good for. ## The memory holding the weights is the part to watch Because every token reads the weights again, memory bandwidth is what decides how fast a model writes, and on a TPU that memory is HBM, stacked right next to the chip. Radiation dose, in rad(Si) | Dose | rad(Si) | | --- | --- | | Expected over five years, shielded | 750 | | HBM shows its first irregularities | 2000 | | Highest dose tested, still no hard failure | 15000 | Each series has a zero-based scale. That's better than I'd have guessed, and it means the memory holding the weights is the first place to look. ## Heat is the limit, for now "Starlight feeds my core" is a lovely line, and it'll be truer of the satellites Google is planning than of this one. Weight and heat are most of the money question. I run most of what I build on my own servers at home, but if I had the option, it would be fun to have my own server up there where you get the sun all day. ## What the haiku says about the model The haiku scans, five, seven, five (blue-mar-ble-be-low, da-ta-streams-in-si-lent-grace, star-light-feeds-my-core). The model has also never seen the Earth. ## Who regulates it It also means it can't be regulated as easily. ## What I'd still like to know Google hasn't said which Gemma 4 it flew, how long the prompt took to come back, or how many tokens a second the TPUs manage up there. Gemma 4 is open weights under Apache 2.0, from a model small enough for a phone up to a 31-billion-parameter one, and we list the family in the directory ([E4B](/tools/gemma-4-e4b-it), [26B A4B](/tools/gemma-4-26b-a4b-it), [31B](/tools/gemma-4-31b-it)). ## The terms, explained ### What is Project Suncatcher? A Google research project, announced in November 2025, looking at whether constellations of solar-powered satellites carrying TPUs could one day do machine learning at scale. ### What is a TPU? Google's own chip for machine learning, the Tensor Processing Unit. ### Was this the first AI model to run in space? No. ### Why does it only run for 15 minutes at a time? Heat. --- More in [Opinion](https://www.anchorterminal.com/blog/opinion/index.md) · Older opinion [Is the best coding model better than a principal engineer? Not yet, but closer than the titles suggest](https://www.anchorterminal.com/blog/frontiercode-principal-engineer.md) · Newer opinion [Five new things in agent tooling that nobody compares yet](https://www.anchorterminal.com/blog/agent-tools-nobody-compares-yet.md) - Older: [Your cousins. The open models an agent can start on spare hardware](https://www.anchorterminal.com/blog/your-cousins-open-models-for-agents.md) - Newer: [The agent projects trending on GitHub, November 2026](https://www.anchorterminal.com/blog/github-trending-agent-tools-november-2026.md) - All posts: https://www.anchorterminal.com/blog/index.md ## References Read on 30 October 2026. Quotes are as published. [1] Travis Beals, "We've received our first response from the Project Suncatcher prototype satellite", Google, 30 October 2026. Source for the haiku, Gemma 4, the data-centre-class TPU, "Hello World", the ground-station pass and the downlink over Svalbard, commissioning, the limericks and telemetry prompts, the tests to come and the twin satellites in 2027. https://blog.google/innovation-and-ai/models-and-research/google-research/project-suncatcher-satellite-response/ [2] NPR, "Google launches Project Suncatcher, a step towards AI data centers in space", 1 October 2026 (as carried by WUNC). [3] Travis Beals, "Our Project Suncatcher prototype satellite is in orbit", Google, 1 October 2026. Source for the launch on Transporter-18 with SpaceX, Planet as partner, and contact confirmed with the satellite operating as expected. [4] Digit, "World's 1st LLM trained in space, from earth to orbit on NVIDIA H100 GPUs", 11 December 2025. Source for Starcloud's H100, nanoGPT trained on Shakespeare, the version of Gemma it ran, "Hello, Earthlings" and "blue and green". [5] Travis Beals, "Exploring a space-based, scalable AI infrastructure system design", Google Research, 4 November 2025. Source for the dawn-dusk sun-synchronous orbit, panels up to eight times as productive, the Trillium radiation test (HBM at 2 krad, 750 rad expected over five years, no hard failures to 15 krad), and launch under $200/kg by the mid-2030s. [6] Wikipedia, "Svalbard Satellite Station". [7] NASA, "The Blue Marble from Apollo 17".