Google’s Orbital AI Just Wrote a Haiku—Now Comes the Hard Part
Google’s Orbital AI Just Wrote a Haiku—Now Comes the Hard Part
Earlier this month, Google placed its Project Suncatcher prototype into orbit aboard a SpaceX Falcon 9 mission, alongside more than 100 other spacecraft. The satellite carries four Tensor Processing Units, though its roughly 1 kW power supply means it can run only one chip for about 20 minutes at a time.
After roughly three weeks of commissioning, ground engineers sent the spacecraft its first real instruction: ask the onboard Gemma 4 model to write a haiku about space. The reply came back intact: “Blue marble below / Data streams in silent grace / Starlight feeds my core.” Google also tested limericks and prompts requesting orbital telemetry.
Sundar Pichai cast the result as a “hello world” moment: a command sent to orbit, a TPU booted, a prompt run and “our first complete response back on earth.” Elon Musk amplified the more expansive interpretation, reposting a claim that the test showed “AI data centers in space actually work.”
That is the optimistic reading. The more cautious one is that Suncatcher has proved a narrow but important capability—orbital inference—not a scalable compute network. Google’s next tests will measure how the TPUs hold up under thermal stress and radiation, which can cause memory “bit flips” and silently corrupt data. The company says reliable workloads in orbit would clear the path toward two interconnected satellites planned for 2027.
The broader ambition is much larger: close-flying satellites linked optically to handle distributed AI workloads. But it depends on more power, dependable inter-satellite communications and, ultimately, launch costs falling sharply—conditions that remain well beyond one poetic transmission from space.
Write a comment