NVIDIA pitches agentic AI
At GTC 2026 Jensen Huang pushed an 'agentic AI' future and introduced enterprise tooling like NemoClaw to make autonomous agents secure and deployable in regulated environments. — the keynote framed AI as an enterprise operating layer, stressing engineering for security and reliability as critical to real adoption. (blogs.nvidia.com)
NVIDIA formally announced the NemoClaw stack on March 16, 2026, saying it bundles NVIDIA Nemotron models with a new OpenShell runtime and can be installed in a single command. (investor.nvidia.com) Jensen Huang framed OpenClaw as “the operating system for personal AI” during the GTC keynote and reiterated the company’s five-layer AI framework while addressing a conference that featured 450+ sponsors, 1,000 sessions and 2,000 speakers. (investor.nvidia.com) NemoClaw adds an isolated sandbox, policy-based security and a privacy router that lets agents combine locally hosted Nemotron models with frontier cloud models while enforcing network and data-privacy guardrails. (investor.nvidia.com) NVIDIA positioned NemoClaw for always-on, enterprise deployments by certifying it for dedicated hardware tiers including GeForce RTX PCs and laptops, NVIDIA RTX PRO workstations, NVIDIA DGX Station, and NVIDIA DGX Spark AI supercomputers. (investor.nvidia.com) Third-party tooling surfaced immediately: JetPatch announced an enterprise control plane for NemoClaw to add governance and patching controls, and NVIDIA showcased new or expanded partner ties with firms such as IBM and Adobe during the GTC program. (marketwatch.com) In the keynote Huang projected at least $1 trillion in cumulative AI chip revenue through 2027 driven by orders for new Blackwell and Vera Rubin platforms, extending a prior $500 billion two‑year outlook. (bloomberg.com) NVIDIA invited developers to an on‑site “build-a-claw” event held March 16–19 in San Jose to accelerate NemoClaw lab-to-production trials, underscoring the company’s push to move OpenClaw prototypes into managed, audited enterprise environments. (investor.nvidia.com)
