01
Research
Google's Project Suncatcher satellite returns its first AI output from orbit
The test is an early step toward whether data-center-class chips can run AI workloads in space at all.
Tensor processing unit: Google's own chip designed for AI work, used inside Google and rented to others through its cloud as an alternative to Nvidia's GPUs.
The test is an early step toward whether data-center-class chips can run AI workloads in space at all.
Models can now target TPUs with far fewer source changes than before, though getting good performance still takes tuning.
Native PyTorch support for TPUs lowers the barrier to running production AI workloads on hardware outside Nvidia's ecosystem.
The tool lowers the shift from Nvidia GPUs to Amazon’s chips by changing configuration instead of application code.
The change is meant to keep future drives moving as SSD bandwidth and controller CPU demands keep climbing.
Google has confirmed contact with the satellite as it starts gathering data on whether machine learning infrastructure can operate in space.
The flight will put AI hardware in orbit for a real-world test, but a useful operational system remains years away.