The ultra-low power draw of Volga AMP™ makes them ideal for onboard AI in unmanned aerial vehicles.
Range, endurance, and payload capacity define the competitive landscape for commercial and defense UAVs. Every milliwatt added to the avionics budget is a milliwatt stolen from propulsion or sensor payload. Volga AMP™ chips change the trade-off calculus by delivering NPU-class neural inference at a fraction of the power budget, opening the door to always-on AI situational awareness aboard small platforms that previously had to choose between flying and thinking.
Obstacle avoidance and object tracking
A Volga AMP™-equipped drone can run simultaneous depth-estimation and object-classification networks at 60 fps using less than 1 W total. This enables smooth obstacle avoidance in GPS-denied environments and real-time target tracking without an uplink to a ground station — capabilities that were previously only achievable with heavy, power-hungry mobile GPUs.
Looking ahead
AI Volga is collaborating with airframe manufacturers and avionics integrators to certify AMP-based inference modules for regulated aerospace applications. Radiation-hardened variants of the V1076 are also in development for stratospheric and space applications.
