GTC crowns agentic AI
NVIDIA’s GTC framed “agentic AI” — autonomous, multi‑step agents — as the new enterprise operating layer, with an emphasis on secure, scalable deployment and orchestration from chip to edge. The event also pushed security-first tooling and reference architectures to move agentic projects from pilots to production (bain.com).
NVIDIA unveiled an open-source Agent Toolkit at GTC that Jensen Huang said will underpin enterprise autonomous agents, and he named 17 initial adopters including Adobe, Salesforce, SAP, ServiceNow, Siemens, CrowdStrike, Atlassian, Cadence, Synopsys, IQVIA, Palantir, Box, Cohesity, Dassault Systèmes, Red Hat, Cisco and Amdocs. (venturebeat.com) The Agent Toolkit bundles Nemotron open models, an AI‑Q blueprint for perception–reason–act workflows, OpenShell as a runtime enforcing policy-based security, and cuOpt as an optimization skill library. (venturebeat.com) NVIDIA also published production-grade reference designs at GTC 2026, unveiling the Vera Rubin platform and a purpose‑built Vera CPU rack alongside a Vera Rubin DSX AI Factory blueprint intended for large-scale agentic deployments. (nvidianews.nvidia.com) Dynamo, NVIDIA’s inference operating system for AI factories, was announced as entering production to manage large fleets of models and runtimes across cloud and edge inference fabrics. (nvidianews.nvidia.com) Security vendors shipped integrations alongside the Agent Toolkit: analysts mapped five vendors to a five‑layer governance model — CrowdStrike for agent decisions/identity, Palo Alto Networks for cloud runtime enforcement, JFrog for supply‑chain provenance, Cisco for prompt‑layer inspection, and WWT for pre‑production validation. (venturebeat.com) NVIDIA cited multiple industry partners showing ecosystem support at GTC, and third‑party vendor lists highlighted 14 infrastructure and OEM partners including Anaconda, Dell, HPE, Lenovo, Microsoft, Supermicro and Vertiv on the show floor. (nvidianews.nvidia.com) (crn.com) Industry risk signals presented at GTC framed why security shipped with the platform: a VentureBeat roundup quoted that 48% of cybersecurity professionals rank agentic AI as the top attack vector heading into 2026 while only 29% of organizations feel fully ready to deploy these technologies securely. (venturebeat.com)
