Uploaded November 2026 | Updated November 2026, 4 days ago
Why did Google shake its AI chips before sending them to space?
Google violently shook a satellite with its AI chips inside. On purpose.
You hear about putting a computer in space, and your mind goes straight to keeping it working up there. Google's chips had to survive about ten minutes of getting there first.
These are TPUs, Google's AI chips, built for data centers. A rocket ride puts that hardware through loads it was never designed for.
Google describes intense vibration and sustained acceleration loads on the spacecraft of up to ten times Earth's gravity during the trip to low Earth orbit.
And that number is for the spacecraft. Individual components, including the chips, can experience forces reaching fifty to a hundred times Earth's gravity.
So you can't take the number for the whole satellite and assume that describes the ride every chip gets inside it. The chips can have a much rougher trip.
That's why Google tested the assembled satellite. The team shook it intensely on all three axes, mimicking the frequencies of a rocket launch. They put the hardware through launch-like vibration while it was still on the ground.
If something gives way during that test, you get to find out before sending it into orbit.
Google's hardware held up. The team was pleasantly surprised by how it handled the force.
Google later confirmed contact with the prototype after launch and described it as operating as expected.
Keeping a computer working in space starts with getting a working computer there. So before handing those chips over to a rocket, Google did this.
Sources:
https://blog.google/innovation-and-ai/models-and-research/google-research/google-project-suncatcher-facts/
https://blog.google/innovation-and-ai/models-and-research/google-research/project-suncatcher-prototype/
thenextweb.com/news/starcloud-250m-orbital-data-centres-nvidia-launch
issnationallab.org/iss360/supercomputing-in-space-hpes-spaceborne-computer-returns
#Google #AI #TPU #Space #shorts #Shorts
Why did Google shake its AI chips before sending them to space?
Google violently shook a satellite with its AI chips inside. On purpose.
You hear about putting a computer in space, and your mind goes straight to keeping it working up there. Google's chips had to survive about ten minutes of getting there first.
These are TPUs, Google's AI chips, built for data centers. A rocket ride puts that hardware through loads it was never designed for.
Google describes intense vibration and sustained acceleration loads on the spacecraft of up to ten times Earth's gravity during the trip to low Earth orbit.
And that number is for the spacecraft. Individual components, including the chips, can experience forces reaching fifty to a hundred times Earth's gravity.
So you can't take the number for the whole satellite and assume that describes the ride every chip gets inside it. The chips can have a much rougher trip.
That's why Google tested the assembled satellite. The team shook it intensely on all three axes, mimicking the frequencies of a rocket launch. They put the hardware through launch-like vibration while it was still on the ground.
If something gives way during that test, you get to find out before sending it into orbit.
Google's hardware held up. The team was pleasantly surprised by how it handled the force.
Google later confirmed contact with the prototype after launch and described it as operating as expected.
Keeping a computer working in space starts with getting a working computer there. So before handing those chips over to a rocket, Google did this.
Sources:
https://blog.google/innovation-and-ai/models-and-research/google-research/google-project-suncatcher-facts/
https://blog.google/innovation-and-ai/models-and-research/google-research/project-suncatcher-prototype/
thenextweb.com/news/starcloud-250m-orbital-data-centres-nvidia-launch
issnationallab.org/iss360/supercomputing-in-space-hpes-spaceborne-computer-returns
#Google #AI #TPU #Space #shorts #Shorts










