CV72S is Ambarella's 5 nm CVflow AI vision SoC for edge image processing, video encoding, and computer-vision workloads in cameras, drones, and robots.
Target environment: Embedded cameras, compact drones, industrial robots, smart cameras, and edge devices requiring local video and AI processing.
Workflow context: Image signal processing, neural-network inference, video encoding, perception preprocessing, and low-power computer vision at the sensor edge.
Customer context: Camera OEMs, drone makers, robotics developers, industrial-vision suppliers, and embedded-device manufacturers.
Deployment model: Silicon-level integration into customer boards and camera products with Ambarella software tools and model support.
Commercial maturity: Commercial SoC family with public Ambarella product positioning and edge-vision use cases.
Adoption constraints: Adoption depends on software-toolchain fit, model support, board design, thermal budget, video pipeline needs, and production volume.
Market position: AI vision silicon layer for robotics, cameras, drones, and embedded perception hardware.
Adjacent products: MIC-711-OX Edge AI System, OAK-D Pro AI Stereo Camera, BHI360 Smart IMU
Audience
Camera OEMs
Workflow
Image signal processing, neural-network inference, video encoding, perception preprocessing, and low-power computer vision at the sensor edge.
Deployment environment
Embedded cameras, compact drones, industrial robots, smart cameras, and edge devices requiring local video and AI processing.
Specifications
Process Node: 5 nm
Video Performance: 4KP60+ image processing, video encoding and decoding
AI Processing: CVflow computer-vision processing
Supported Ml Frameworks: Caffe, PyTorch, TensorFlow and ONNX CNN toolchain support
Target Use Cases: IP cameras, consumer and industrial robots, sports/action cameras, video conferencing cameras and drones