NVIDIA AI Updates: June 2, 2026

1. NVIDIA Jetson Brings Agentic AI to the Physical World With JetPack 7.2

NVIDIA. NVIDIA announced that Jetson edge AI platforms are gaining expanded agentic AI capabilities, positioning the hardware for autonomous robots, industrial inspection, and physical AI systems that must act on sensor data without cloud round-trips. Alongside the platform announcement, NVIDIA published JetPack 7.2, which delivers memory-efficiency improvements and agentic-ready AI deployment tools optimized for edge inference. The release supports NVIDIA NIM microservices at the edge and is aimed at developers building perception, reasoning, and action pipelines on embedded hardware. Source

2. NVIDIA DGX Spark Gets Faster Models and Multi-Node Clustering for Local AI Agents

NVIDIA. NVIDIA detailed new software capabilities for DGX Spark that enable running local AI agents with faster models and multi-node clustering, allowing multiple DGX Spark units to be combined for larger workloads. The update is part of the broader RTX AI Garage initiative announced at Computex 2026, which expands local agentic AI across the RTX PC and DGX Spark ecosystem. NVIDIA is positioning DGX Spark as a personal AI supercomputer capable of running sovereign, on-premises agent workflows without external API dependencies. Source

3. NVIDIA DOCA In-Silicon Security Hardens Agentic AI Infrastructure

NVIDIA. NVIDIA published details on DOCA In-Silicon Security, a capability that embeds security enforcement directly into the silicon of NVIDIA BlueField data processing units rather than relying on host software. The technology is designed to protect AI factory infrastructure from east-west lateral movement attacks, a growing threat as interconnected agentic AI systems communicate across large clusters. NVIDIA frames this as a foundational infrastructure layer for the multi-tenant AI factories enterprises are building to run agentic workloads at scale. Source

4. NVIDIA AI Cloud Ecosystem Expands Globally to Meet AI Compute Demand

NVIDIA. NVIDIA announced a global expansion of its AI cloud ecosystem, adding new cloud service provider and colocation partners worldwide to increase availability of NVIDIA GPU compute. The program is designed to help enterprises and sovereign AI initiatives access NVIDIA-accelerated infrastructure without building their own data centers. The expansion is tied to NVIDIA’s broader Computex 2026 announcements around scaling AI factory deployments internationally. Source

5. NVIDIA Factory Operations Blueprint Gives Industrial AI a Software Brain

NVIDIA. NVIDIA introduced the Factory Operations Blueprint, a reference architecture that combines robotics, simulation, and AI orchestration software to give factories an AI-driven operational intelligence layer. The blueprint builds on NVIDIA Omniverse and the Isaac robotics platform, targeting manufacturers looking to deploy digital twins and autonomous inspection or assembly systems. It is part of NVIDIA’s broader push to deliver end-to-end AI factory software stacks as productized blueprints rather than individual tools. Source