NVIDIA's GTC 2026 conference begins Monday, March 16, in San Jose, with CEO Jensen Huang delivering the keynote at 11 a.m. PT. The event is expected to include major announcements on new AI chips, the NemoClaw agentic AI platform, Omniverse updates, and NVIDIA's full stack roadmap for 2026 and beyond.

What Happened

NVIDIA's premier AI conference runs March 16 to 19 at the San Jose Convention Center, with more than 30,000 attendees from over 190 countries registered, plus a virtual audience. The conference is already underway in pre-event form, with NVIDIA publishing live coverage on its blog. Huang's keynote on Monday is the anchor event.

Analysts from Bank of America flagged three specific areas to watch: updates on the Feynman GPU pipeline (expected 2028), progress on low-latency inference products, and networking and quantum computing roadmap updates. NVIDIA's Vera Rubin microarchitecture, the successor to Blackwell systems, is also expected to receive attention.

Why It Matters

GTC is where NVIDIA announces infrastructure that shapes what AI tools can do for the next 12 to 18 months. The hardware roadmap dictates the speed and cost of running models locally, in the cloud, and at the edge. For creative professionals, the key question is what comes after RTX 50 Series and whether local inference costs continue to fall fast enough to make on-device generation the default for studios.

The NemoClaw platform, which NVIDIA has previewed as an open-source system for building enterprise AI agents, is directly relevant to creative workflows. If the announcement delivers on what was previewed, it would mean a standard framework for building autonomous creative assistants that can navigate software, manage assets, and complete multi-step production tasks without constant human input.

Key Details

  • Keynote: Jensen Huang, Monday March 16, 11 a.m. PT at the SAP Center in San Jose. Streamed live.
  • Expected announcements: NemoClaw open-source agentic AI platform, Vera Rubin GPU microarchitecture update, AI inference chip for efficient model deployment, new Omniverse capabilities.
  • Physical AI focus: GTC 2026 has a heavy emphasis on physical AI, including robotics, digital twins, and autonomous systems.
  • Attendance: 30,000+ registered in-person, 190+ countries represented virtually.
  • Context: GTC follows directly after NVIDIA's announcements at GDC, where the company already revealed ComfyUI App Mode and NVFP4 for local creative AI workflows.

What to Do Next

Watch or stream Jensen Huang's keynote Monday, March 16, starting at 11 a.m. PT at nvidia.com/gtc. The keynote historically runs two-plus hours and includes live demos. For the creative AI highlights, follow the NVIDIA GTC live blog which is being updated in real-time throughout the week.

If you are a developer or technical artist, the GTC session catalog includes hands-on workshops on OpenUSD for physical AI simulation, agentic AI deployment, and inference optimization that are worth registering for even virtually.