Google’s Project Suncatcher just crossed a threshold that sounded like science fiction. The company’s prototype satellite, launched on October 1, successfully received a command from Earth, booted up its onboard Tensor Processing Unit (TPU), and generated a complete response. Ceo Sundar Pichai called it a "hello world" moment for the initiative, confirming the system is alive and responsive.
The specific prompt? A request for a haiku about space. The satellite processed this using the Gemma 4 open model while passing over Svalbard, Norway. Pichai posted the confirmation on X, detailing the sequence: command sent, hardware activated, text returned. It’s a small output, but for a team testing if silicon can survive the vacuum, it’s a massive win.
The satellite has been in orbit for weeks, running commissioning checks. This included powering up host systems and running early machine learning tests. Google engineers didn't stop at poetry. They tested the system with space-themed limericks and prompts asking the AI to report its own technical status. The hardware held up.
Now comes the hard part. The next phase involves subjecting the TPUs to thermal stress and radiation exposure. These are the killers of electronics in low Earth orbit. If the chips fail here, the project stalls. If they survive, Google moves to its next milestone: two interconnected satellites in 2027. The goal is a constellation that leverages near-constant sunlight to generate up to eight times more solar power than ground-based servers.