Premio adds JetPack 7.2 support to rugged JCO edge AI computers

Premio said on Sept. 22 it added NVIDIA JetPack 7.2 support across its JCO Series rugged edge AI computers built on Jetson Orin. The company is showing agentic AI and vision-language AI demos at embedded world North America 2026 in Anaheim, California.

Why it matters: - JetPack 7.2 gives robotics and physical AI developers new software tools for edge deployment on rugged, fanless systems. - The update is aimed at reducing platform setup work and improving performance for workloads running close to robots, cameras, sensors and machines. - Premio is pairing the software update with live demos to show how edge AI can work without relying on constant cloud connectivity.

What happened: - Premio announced support for NVIDIA JetPack 7.2 SDK across its JCO Series of Rugged Edge AI Computers powered by NVIDIA Jetson Orin. - The announcement came on Sept. 22, 2026, alongside a product showcase at embedded world North America 2026. - Premio is demonstrating the platform at Booth #6100 from Sept. 22-24 at the Anaheim Convention Center in Anaheim, California. - The booth demos focus on two use cases: agentic AI using NVIDIA NemoClaw and live vision-language AI using NVIDIA Jetson AGX Orin 32GB Super Mode.

The details: - The JCO Series includes the entry-level JCO-1000-ORN Series and midrange JCO-3000-ORN Series, both powered by NVIDIA Jetson Orin Nano and Orin NX. - The high-performance JCO-6000-ORN Series uses NVIDIA Jetson AGX Orin. - Premio says the fanless systems are built for real-world environments that need reliable local compute. - The portfolio is positioned for autonomous mobile robots, machine vision and industrial automation. - JetPack 7.2 adds NVIDIA Jetson Agent Skills for Jetson Linux customization, memory optimization and model benchmarking. - NVIDIA Jetson Agent Skills are designed to help developers streamline system bring-up, use memory more efficiently and identify suitable models and runtime configurations. - JetPack 7.2 also includes official Yocto Project support for leaner, reproducible embedded Linux builds in production. - One demo shows NVIDIA NemoClaw running locally on the JCO-1000-ORN Series. - That demo is designed to show AI agents processing and responding to tasks at the edge while reducing reliance on cloud connectivity. - The same demo shows agent orchestration on NVIDIA Jetson Orin in a controlled runtime environment. - A second demo shows a vision-language model analyzing continuous camera input on the JCO-6000-ORN Series for real-time scene understanding and reasoning. - Customizable prompts in that demo are set to show object recognition, spatial reasoning, activity detection, safety monitoring and OCR at the edge. - NVIDIA Jetson AGX Orin 32GB Super Mode raises performance from 200 TOPS to 241 TOPS. - The added headroom is intended for demanding multimodal and robotics workloads before developers move to the AGX Orin 64GB configuration. - Premio says the multimodal demo combines vision and language processing for robotics, machine vision and physical AI applications. - The company’s booth demos align with embedded world North America 2026’s focus on autonomous, agentic and physical AI. - On Sept. 22, NVIDIA Vice President of Robotics and Edge AI Deepu Talla is scheduled to deliver the opening keynote, “The Next Evolution of Embedded: Open Platform for Autonomous, Agentic Systems.”

Between the lines: - Premio is trying to show that edge AI software and rugged hardware are converging into one deployment stack for industrial robotics. - The company is also signaling a broader shift from model experimentation to production-oriented workflows such as benchmarking, memory tuning and reproducible Linux builds. - The Super Mode demo suggests NVIDIA and Premio are targeting developers who need more compute before stepping up to higher-memory hardware.

What’s next: - Premio said it is expanding its NVIDIA Jetson portfolio with the upcoming WCO-3000-ORN Series and WCO-6000-THR Series. - The WCO-3000-ORN Series will use NVIDIA Jetson Orin Nano and Orin NX. - The WCO-6000-THR Series will use NVIDIA Jetson AGX Thor. - Both upcoming IP66 platforms are intended for robotics, vision AI and autonomous systems in outdoor and industrial environments. - Existing JCO Series customers can obtain the JetPack 7.2 BSP by contacting sales@premioinc.com for support. - More information is available through Premio's LinkedIn page and Premio's YouTube channel.

The bottom line: - Premio is using JetPack 7.2 to position its rugged Jetson Orin computers as a ready-made edge AI platform for robotics teams that want faster deployment and more capable on-device inference.