Vision3h ago
NAVI-Orbital Debuts Onboard Vision-Language Inference
On April 16, 2026, NAVI-Orbital achieved the first in-orbit demonstration of a vision-language model performing autonomous inference. Deployed on a LEO spacecraft, the system uses Gemma 3 to classify scenes and generate text descriptions. This removes the downlink bandwidth bottleneck. Practitioners can now execute complex multi-modal analysis directly on satellite hardware.