Project Suncatcher

Google's research moonshot to run TPU AI chips on solar-powered satellites in orbit, with a prototype satellite scheduled to launch October 1, 2026.

Google · Infrastructure and inference/serving · infrastructure

Published in Feed

This entry is marked for review. Check the original sources before relying on its details.

Overview

Google first disclosed Project Suncatcher in November 2025 as a long-term research effort exploring whether space could host scalable machine-learning infrastructure. A prototype satellite nicknamed MVP, built in partnership with satellite-imaging company Planet Labs and carrying four Google TPU chips (roughly the compute of one data-center server), is scheduled to launch October 1, 2026 aboard a SpaceX Falcon 9 -- the first real-world test of whether Google's AI chips can survive and operate in orbit. In low Earth orbit, solar panels on the satellite are expected to supply about a kilowatt of power to the chips, and Google says satellites there can generate up to eight times more solar power than on Earth. Google plans a two-satellite constellation by 2027.

This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.

Key features

  • Prototype satellite (MVP) carrying 4 Google TPU chips, built with Planet Labs
  • Scheduled launch October 1, 2026 aboard a SpaceX Falcon 9
  • ~1kW of onboard solar power for the chips in low Earth orbit
  • Long-term goal: a two-satellite TPU constellation in orbit by 2027

Use cases

  • Research into space-based scaling of AI compute infrastructure
  • Testing radiation/thermal survivability of TPU chips in orbit
  • Exploring solar-powered, off-Earth data center capacity

Pricing

not applicable -- research/infrastructure moonshot, no commercial product or pricing.

Pricing reflects the source record and may change. Confirm current terms with the provider.

Dates and sources

Released
1 Oct 2026
Record updated
29 Sept 2026
Record last checked
29 Sept 2026

Explore the Feed

Other entries filed under infrastructure and inference/serving; inclusion does not imply equivalent features.

  • NVIDIA PAIR (Personal AI Router)

    NVIDIA's free, open-source personal AI router that spreads local model workloads across PCs on your network; out of beta Oct 13, 2026.

  • Agent37

    YC F26 startup selling low-cost agent sandboxes (gVisor or full VM) with prebuilt harnesses for Claude Code, Hermes and opencode.

  • Microsoft-Decision-1

    Microsoft's Qwen3.5-9B-based decision-scoring model returns calibrated option probabilities in one pass, at $0.042 per million input tokens.

Browse the Feed