Deploy Agentic-Ready AI at the Edge with Memory Efficiency in NVIDIA JetPack 7.2

2026-06-02 · Source: NVIDIA Technical Blog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Internet of Things (IoT) & Connected Devices · Depth: Advanced, medium

Summary

NVIDIA JetPack 7.2 is a new software release designed to accelerate agentic AI deployment at the edge, optimizing memory and performance on NVIDIA Jetson platforms. This update introduces one-command deployment for NVIDIA NemoClaw, an open-source stack enhancing privacy and security for OpenClaw-based workflows. It also provides NVIDIA agent skills for Jetson, including device-side and BSP skills, to automate development tasks like Linux customization, memory optimization, and model benchmarking. Key features include Multi-Instance GPU (MIG) support on Jetson Thor for deterministic multi-workload execution, official Yocto Project support for custom Linux distributions, and a new Super Mode for Jetson AGX Orin 32 GB. Super Mode boosts AI performance from 200 TOPS to 241 TOPS by increasing GPU frequencies to 1.3 GHz and power envelopes to 60W, reducing module cost by 45% compared to the 64 GB version. These enhancements unify the Jetson software stack and improve efficiency for existing hardware.

Key takeaway

For AI Engineers deploying agentic workloads at the edge, JetPack 7.2 offers significant advantages. You can achieve higher performance and lower total cost of ownership by utilizing the new Super Mode on Jetson AGX Orin 32 GB, boosting AI performance to 241 TOPS. Leverage one-command NemoClaw deployment and agent skills to automate development, reducing time to market. Consider MIG on Jetson Thor for deterministic execution of mixed-criticality applications, ensuring reliable performance for your robotics and industrial automation projects.

Key insights

JetPack 7.2 enables efficient, agentic AI deployment at the edge through optimized software, hardware partitioning, and automated development.

Principles

• Software updates enhance existing hardware value.
• Agent skills automate complex development tasks.
• GPU partitioning ensures predictable workload performance.

Method

Deploy NemoClaw on Jetson with `curl -fsSL https://www.nvidia.com/nemoclaw.sh | bash`. Utilize agent skills for automated Linux customization, memory optimization, and model benchmarking.

In practice

• Use Super Mode on AGX Orin 32 GB for 241 TOPS.
• Partition Jetson Thor GPU for mixed-criticality tasks.
• Build custom Linux distributions with Yocto Project.

Topics

• NVIDIA JetPack 7.2
• Edge AI
• AI Agents
• NVIDIA Jetson
• Multi-Instance GPU
• Yocto Project
• NemoClaw

Code references

• NVIDIA-AI-IOT/jetson-device-skills
• NVIDIA-AI-IOT/jetson-bsp-skills
• NVIDIA-AI-IOT/DeepStream_Coding_Agent
• NVIDIA-AI-Blueprints/video-search-and-summarization

Best for: AI Architect, CTO, VP of Engineering/Data, AI Engineer, Machine Learning Engineer, Robotics Engineer

Related on AIssential

• How to Build In-Vehicle AI Agents with NVIDIA: From Cloud to Car — Agentic, multimodal AI systems are transforming automotive cockpits, requiring robust edge-cloud architectures for real-time performance.
• NVIDIA GTC Taipei at COMPUTEX: Live Updates on What’s Next in AI — NVIDIA's latest platforms advance AI, robotics, and autonomous vehicles with significant performance and efficiency gains.
• NVIDIA Partners With Microsoft on Unified Stack for Agentic AI Deployment, From Windows Devices to Cloud to Local — NVIDIA and Microsoft are building a full-stack, secure, and scalable ecosystem for agentic AI, from edge devices to cloud.
• 621: Going Deep on Nvidia GTC 2026, Agentic Era, Groq Strategy, OpenClaw Hype, The Android of Self-Driving, 1990 Video Store Simulator, Shingles Vaccine vs Cognitive Decline, and Rick Beato Interview — The AI industry is rapidly transitioning to an agentic era, driven by Nvidia's interconnected hardware and software innovations.
• The foundation for a governed agent workforce: DataRobot and NVIDIA RTX PRO 4500 — The NVIDIA RTX PRO 4500 provides a balanced hardware foundation for enterprise AI agent inference.
• Robots’ Holiday Wishes Come True: NVIDIA Jetson Platform Offers High-Performance Edge AI at Festive Prices — NVIDIA Jetson developer kits offer scalable edge AI compute for diverse robotics and physical AI applications.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by NVIDIA Technical Blog.