Google's first orbital data center satellite successfully runs AI model in space
Google's Project Suncatcher prototype satellite launched on October 1 and is now operational, marking an early milestone in space-based AI computing research. The satellite, built by Planet Labs, carries four Tensor Processing Units and generates up to one kilowatt of power. Engineers successfully ran Google's Gemma 4 AI model in orbit and had it generate a haiku about being in space. The satellite can currently operate for only about twenty minutes at a time due to power constraints. The test campaign will examine how chips perform under thermal stress and space radiation. If successful, Google plans to launch two more satellites in 2027 with improved power generation and inter-satellite optical links. The project addresses growing concerns about AI's enormous energy demands by exploring whether solar-powered satellites could eventually supplement Earth-based data centers.
Key facts
- Google's Suncatcher satellite carries four TPUs generating 1 kilowatt of power and launched on October 1 via SpaceX Falcon 9
- The satellite can operate only one TPU at a time for approximately 20 minutes due to power constraints
- Engineers prompted the onboard Gemma 4 AI model to write a haiku about being in space
- Google plans to launch two additional satellites with advanced chips and inter-satellite optical links in 2027
- Environmental groups are petitioning the FCC to require full environmental reviews before approving orbital data center proposals
- Google research estimates launch costs must fall to approximately 200 dollars per kilogram for space-based computing to become economically competitive with terrestrial data centers
How different outlets framed it
Left-leaning coverage: Environmental groups and progressive outlets highlight significant concerns about orbital data centers, including potential atmospheric pollution from rocket launches, the brightening of the night sky affecting astronomy and wildlife, and the need for comprehensive environmental review before approving orbital infrastructure. Advocates argue that moving data centers to space could relocate rather than solve environmental damage, and warn of potential effects on the ozone layer and climate if large satellite constellations are deployed.
Right-leaning coverage: Right-leaning and skeptical voices question the cost-effectiveness and prioritization of orbital solutions, arguing the resources could be better spent on terrestrial solutions like solar farms. Some commenters note the enormous launch costs relative to the minimal power output of the prototype and question whether space-based infrastructure represents sound resource allocation.
Limits of this briefing
The poetry aspect, while capturing attention, is primarily a validation test rather than the core research goal. Coverage heavily skews toward tech and centrist outlets with underrepresentation of sustained environmental analysis. The long-term viability remains unproven, and technical challenges around cooling, radiation hardening, and launch economics are substantial.
Sources
- TechCrunch: Google's first data center satellite is online and writing poetry Centre
- The Register: Google's AI satellite sends a very expensive haiku to Earth Centre
- Christian Science Monitor: Faced with earthly opposition, tech leaders want to put data centers in orbit Left
- CNN: AI data centers in space? Google kicks off the race with a key test Centre
- Scientific American: Google tests its plan for AI data centers in space with Project Suncatcher Centre