ORBION INTELLIGENCE BRIEF · VERSION 1

Google Put AI Compute in Orbit. The Bigger Story Is the Industry That Could Form Around It.

Google’s Project Suncatcher prototype is now in orbit. Orbion maps the larger emerging stack behind orbital AI compute—from launch and solar power to thermal management, radiation tolerance, optical networking and distributed infrastructure.

Published

THE SIGNAL

On October 1, 2026, Google announced that its first Project Suncatcher prototype satellite, built with Planet, launched aboard SpaceX’s Transporter-18 rideshare mission. Google says it established contact and the spacecraft is operating as expected. Over the coming weeks, the team plans to collect in-orbit data on how TPUs handle launch stress, radiation and thermal extremes.

WHAT GOOGLE IS ACTUALLY TESTING

Project Suncatcher is not a deployed orbital data center. It is a research program exploring whether scalable machine-learning infrastructure could eventually operate in space.

Google’s published architecture envisions solar-powered satellites carrying TPUs and connected through free-space optical links. Its research identifies hard dependencies including high-bandwidth inter-satellite communications, tightly controlled satellite formations, radiation tolerance, thermal management, ground communications, on-orbit reliability and launch economics.

THE ORBION VIEW

The headline is that Google put AI hardware in orbit.

The intelligence question is: what new industrial stack would have to exist if orbital AI compute becomes viable?

Orbion maps the dependency chain:

AI demand → power demand → orbital solar power → launch → spacecraft platforms → thermal rejection → radiation-tolerant compute → optical inter-satellite links → precision formation and navigation → ground connectivity → distributed-compute software → operations → regulation, insurance and capital.

That chain matters because the economic opportunity may extend far beyond the company making the accelerator.

THE BOTTLENECKS MAY BECOME THE MARKET

Cooling is one example. Google describes thermal management as a crucial challenge because there is no airflow in vacuum; heat must ultimately be rejected through radiators. Google says it is testing approaches using heat pipes and radiators.

Networking is another. Large distributed AI workloads require enormous bandwidth. Google’s research discusses tens-of-terabits-per-second inter-satellite links and compact satellite formations, while reporting a bench-scale optical demonstration of 800 Gbps in each direction using a single transceiver pair.

Radiation is a third. Google reported proton-beam testing of its Trillium v6e TPU and found no hard failures attributable to total ionizing dose up to the maximum tested dose on a chip, although high-bandwidth memory was more sensitive. The orbital mission now provides a different class of evidence: real operating conditions.

THE ECONOMIC HINGE

Google’s research estimates that, if launch prices eventually fall below roughly $200 per kilogram in the mid-2030s, the cost of launching and operating space-based compute could approach reported terrestrial data-center energy costs on a per-kilowatt/year basis.

That is a modeled condition—not a demonstrated economic result.

WHAT TO WATCH NEXT

The strongest next signals are actual in-orbit TPU performance; thermal behavior; radiation-related faults; future optical crosslink results; changes in launch economics; new suppliers, patents, hiring and capital tied to orbital compute; and credible competing programs.

WHY THIS MATTERS

Search tells us that a satellite launched.

Intelligence asks what changed in the surrounding system, what depends on success, what could fail, who becomes strategically important, and what evidence would change the conclusion.

Project Suncatcher is still an experiment.

But if it works, the important story may not be a computer in space.

It may be the emergence of an entirely new infrastructure stack around computation itself.

SOURCES

[1] Google, October 1, 2026 — Our Project Suncatcher prototype satellite is in orbit.
https://blog.google/innovation-and-ai/models-and-research/google-research/project-suncatcher-prototype/

[2] Google Research, November 4, 2025 — Exploring a space-based, scalable AI infrastructure system design.
https://research.google/blog/exploring-a-space-based-scalable-ai-infrastructure-system-design/

[3] Google, September 24, 2026 — Behind Project Suncatcher, our moonshot to put AI in space.
https://blog.google/innovation-and-ai/models-and-research/google-research/google-project-suncatcher-facts/

[4] Planet, November 4, 2025 — Planet to Build and Operate Advanced Space Platform for Google’s Project Suncatcher Moonshot.
https://www.planet.com/pulse/planet-to-build-and-operate-advanced-space-platform-for-google-s-project-suncatcher-moonshot/

EVIDENCE

Sources

  1. Google, October 1, 2026 — Our Project Suncatcher prototype satellite is in orbit.
  2. Google Research, November 4, 2025 — Exploring a space-based, scalable AI infrastructure system design.
  3. Google, September 24, 2026 — Behind Project Suncatcher, our moonshot to put AI in space.
  4. Planet, November 4, 2025 — Planet to Build and Operate Advanced Space Platform for Google’s Project Suncatcher Moonshot.

Methodology · Corrections