Primary source HardwareArticle·25 September 2026·4 min read

Project Suncatcher: Google Sends Its First TPUs Into Orbit

Before dreaming of space data centers, Google wants to answer one simple question: can its AI chips survive a launch, radiation, and the vacuum of space?

ShareWhatsApp X LinkedIn Facebook

Save to my selection

Project Suncatcher: Google Sends Its First TPUs Into Orbit

Source : Travis Beals · Google · 24 September 2026View original ↗

Who covers it

On TechmemeHacker News 159 pts ↗

Picked up by 28 outlets · 2 top-tier · 4 tech · 4 forums · 4 social posts

In brief

Google has launched the first in-orbit test of Project Suncatcher, a prototype satellite built with Planet and flown aboard SpaceX's Transporter-18 mission to evaluate TPUs in space. The post details the project's three technical challenges: hardware survival, cooling in a vacuum, and laser links between satellites. A modest but concrete step toward the idea of sun-powered AI computing in low Earth orbit.

🍺 Bar-stool version

Google decided that if data centers were running short on power on Earth, the answer was to go get it where the sun barely ever sets: in orbit. First test, send a few chips up there and check that they don't melt, don't get fried by cosmic rays, and don't rattle apart at liftoff—basically the spec sheet for a smartphone handed to a four-year-old. The funniest part is that space is very cold, yet you can't cool anything down, because there's no air to do it with. If this works, the AI energy race won't just be fought around power plants anymore, but above our heads too.

Key takeaways

1
A prototype satellite equipped with Google TPUs launches on SpaceX's Transporter-18 rideshare mission, developed in partnership with Planet.
2
In low Earth orbit, a satellite can generate up to eight times more solar power than on Earth thanks to near-constant sunlight.
3
At launch, loads reach up to 10 g for the vehicle and 50 to 100 g for certain components like the TPU chips; the hardware held up through three-axis vibration testing.
4
Tested under a proton beam at UC Davis's Crocker Nuclear Laboratory, Trillium TPUs withstand a total ionizing dose higher than that of a five-year mission.
5
Without air, cooling relies on heat pipes and radiators, already validated in a thermal vacuum chamber.
6
Future satellites will each carry dozens of TPUs, linked into clusters via very high-bandwidth lasers over very short distances.
7
Next step in 2027: two satellites in orbit to test the inter-satellite laser link.

A moonshot facing the test of reality

Announced the previous year, Project Suncatcher explores an idea simple to state and formidable to execute: hosting large-scale machine learning infrastructure in space.

Surviving launch and radiation

The trip to low Earth orbit takes about ten minutes, with intense vibrations and accelerations of up to 10 g.

Cooling without air

A TPU generates a lot of heat over a small surface area.

Lasers to build a cluster in orbit

Eventually, each satellite will carry dozens of TPUs and fly in tight formation with others.

Why it matters

AI's number one constraint is no longer just silicon—it's energy and cooling, and Suncatcher is the most radical answer any hyperscaler has put on the table.

Older than 30 days: the archive

This article is more than 30 days old

A free account opens articles from the last 30 days in full. The full archive is in the Pro plan.

Create my account

Already a member? Log in

Compare plans →

#google #tpu #space #data centers #energy #hardware

Original source

Behind Project Suncatcher, our moonshot to put AI in space

Travis Beals

Open the article ↗

For you

Put it to work on your sources.

Free: a month of articles and three sources of your own. Pro: the whole archive and your sources, from €8/month.

Create my account →

For your team

The same machine, on your topics.

A space in your colours, your watch angles, your curators. Pilot open to three companies.

Request pilot access →