Google has officially taken its Project Suncatcher experiment into orbit. The company’s prototype satellite launched on October 1 aboard SpaceX’s Transporter-18 rideshare mission, marking the first in-orbit test of Google’s idea for running scalable machine-learning infrastructure in space.

Built in partnership with Planet, the prototype is already communicating with Google’s team and is operating as expected, according to the company. Google will now collect data on how its Tensor Processing Units (TPUs) handle the physical stresses of spaceflight, as well as the radiation and extreme thermal conditions found in space.

Google describes Suncatcher as a long-term research project exploring whether space could eventually support scalable machine-learning infrastructure. This first satellite is therefore less about running a full AI data center in orbit and more about finding out whether the underlying hardware can actually operate reliably in space.

What is Project Suncatcher?

Project Suncatcher is Google’s research effort to explore the possibility of putting AI computing infrastructure in space.

The concept is based around satellites equipped with Google’s TPUs, the company’s custom processors designed for machine-learning workloads. In a future version of the idea, multiple satellites could potentially work together as an orbital computing system.

That vision is still a long way from becoming reality. Google’s newly launched prototype is intended to provide real-world data that can help the company understand the engineering challenges involved.

And that is why actually putting the hardware in orbit matters.

Google is putting its TPUs through a real space test

Computer hardware can be tested extensively on Earth, but some conditions are difficult to reproduce outside space.

For Project Suncatcher, Google wants to understand how its TPUs respond to the physical stresses of launch and operation in orbit, along with the radiation and thermal extremes of the space environment.

The company says it will spend the coming weeks gathering this in-orbit data. Those results will then be used to refine the project’s designs as the research continues.

Google has also published a peer-reviewed paper in Joule detailing the research behind the mission.

Why does Google want AI infrastructure in space?

The broader Suncatcher project is based on an intriguing idea: instead of keeping all AI computing infrastructure on Earth, some of it could eventually operate in space.

Satellites could potentially use solar power and communicate with one another to form a distributed computing system. For Google, the growing demand for AI computing makes alternative approaches to infrastructure worth investigating.

But Project Suncatcher is still a research project, not a replacement for conventional data centers.

The current mission is essentially Google asking a very practical question before it goes any further: what happens when you actually put its AI computing hardware in space?

The first Suncatcher satellite is just the beginning

Google calls the mission the first step in a long-term research moonshot. The company plans to use what it learns from the prototype to refine its designs and continue exploring the possibility of scalable machine-learning infrastructure in space.

For now, the milestone is much simpler: Google has successfully put its Project Suncatcher prototype into orbit, established contact with it, and started collecting data.

The next few weeks should reveal how well Google’s AI hardware handles life beyond Earth.