TD SYNNEX Newsflash

Google tests AI data centres in Earth orbit with Suncatcher

Artificial Intelligence
By TD SYNNEX Newsflash 8th October 2026

Google has launched the first satellite in Project Suncatcher, a Google research moonshot designed to explore whether Earth orbit data centres could become a viable alternative to those on terra firma, or traditional land-based facilities.

Launched aboard a SpaceX Falcon 9 rocket on 1st October 2026, the satellite carries Google's Tensor Processing Units (TPUs) and will assess how the company's AI hardware performs during spaceflight and in the challenging conditions of low Earth orbit.

Google tests AI data centres in Earth orbit with Suncatcher

Project Suncatcher is part of Google's moonshot programme, a term the company uses for ambitious, long-term projects tackling major technological challenges. In this case, the challenge is meeting the rapidly growing demand for AI computing power.

AI demand fuels interest in orbital computing

The rise of AI is placing increasing pressure on data centre capacity, energy supplies and cooling infrastructure, prompting interest in orbital computing, where data is processed and stored using infrastructure in space rather than on Earth.

Google believes Earth orbit data centres could offer access to near-constant sunlight for power generation, while clusters of interconnected satellites could potentially handle large-scale computing workloads. Although several organisations have explored the concept of space-based data centres, Google is the first to launch a dedicated mission focused on testing its own AI accelerator hardware in orbit.

Testing TPUs for space

Before launch, Google's Trillium TPUs underwent extensive testing at the Crocker Nuclear Laboratory at the University of California to assess whether they could withstand the stresses of launch and long-term operation in space.

Testing focused on three key challenges:

  • Launch stresses: Simulations replicated the intense acceleration and vibration experienced during launch.
  • Radiation exposure: Engineers assessed how the chips would cope with solar and cosmic radiation, which can interfere with computing systems and cause data errors known as bit flips.
  • Cooling: Google evaluated whether its cooling systems could effectively manage the significant heat generated by AI chips in a vacuum environment.

According to Google, the TPUs performed well across all three areas.

What's next?

The first mission will provide real-world data on how Google's TPUs perform in orbit. Future missions could involve multiple TPU-equipped satellites operating in clusters and communicating through laser links, potentially laying the foundations for large-scale orbital computing infrastructure.

Today’s news was brought to you by TD SYNNEX – the UK’s number one solutions distributor.