Abstract：A new AI vision platform, Darsi Pro, will debut at CES 2026, offering industrial-grade performance and multi-sensor connectivity for autonomous systems. The platform aims to simplify edge AI deployment for robotics, intelligent transportation, and other advanced vision applications.
A leading embedded vision solutions provider announced it will unveil its first AI vision compute box, Darsi Pro, at CES 2026. Built on the NVIDIA Jetson platform, the system integrates cloud-based device management, multi-sensor connectivity, and industrial reliability to streamline edge AI deployment.
The Darsi Pro vision platform is built on NVIDIA Jetson and delivers up to 100 TOPS of AI performance.
The developer highlighted that Darsi Pro represents a shift from providing cameras alone to offering a complete vision and compute platform. The system is designed to accelerate production-ready solutions for autonomous mobile robots, delivery robots, warehouse systems, and intelligent transportation applications.
Darsi Pro delivers up to 100 TOPS of AI performance and supports synchronized inputs from cameras, lidar, radar, and IMUs using Precision Time Protocol. Its GMSL variant can connect up to eight cameras and is compatible with NVIDIA JetPack 6 and higher versions.
The platform offers extensive camera compatibility, flexible connectivity options including dual GbE with PoE, USB 3.2, HDMI, CAN, and wireless modules, and a rugged enclosure for industrial-grade reliability. Cloud management allows over-the-air updates, remote configuration, and device monitoring.
Drawing on more than 20 years of embedded imaging experience, the developer ensures consistent performance across a broad portfolio of cameras, including MIPI, GMSL, USB, HDR, depth, and low-light models. End-to-end integration and customization support are also provided to meet unique application requirements.
At CES 2026, the platform will be demonstrated in real-world scenarios, such as automated license plate recognition and delivery robots. Additional showcases will feature ultra-low-light, high-resolution cameras integrated into inventory robots for navigation, obstacle detection, and automated shelf analytics.
