NVIDIA GTC 2027 San Jose

NVIDIA GTC is the AI conference run by NVIDIA, and the 2027 edition is set for 15 to 18 March in San Jose, California. The official site confirms the city and the dates. It has not yet named a venue or published any other logistics for the coming edition. NVIDIA describes GTC as a meeting point for thousands of developers, researchers and business leaders, reached through the main conference and a set of regional events. The organiser lists nine key themes for 2027. The engineering strand covers CUDA and its libraries, AI infrastructure and data processing. Applied topics include agentic AI, industrial AI, physical AI and robotics, AI for science, and content creation and rendering, while cybersecurity and AI safety form a theme of their own. One part of the 2027 programme is already moving: the call for poster submissions is open. The 2026 edition shows what the format looked like. Jensen Huang, NVIDIA's chief executive, gave the keynote, which was preceded by a GTC Live pregame show, and the featured speakers included Jeff Dean, Mira Murati and Chelsea Finn. Sessions that year centred on physical AI, AI factories, agentic AI and inference, and the programme also offered hands-on training, an exhibits area and time with NVIDIA experts. Replays of the 2026 sessions remain available on demand. Registration for 2027 has not opened, and no prices or pass types have been announced. The official site lets visitors sign up for a notice when sales begin. For reference, the 2026 edition sold four-day and one-day Conference passes, an Exhibits Only pass, full-day workshops on the Sunday and a Digital Only option. Only Conference pass holders could attend the keynote, and training labs cost extra on top of a pass. For affiliates and media buyers the fit is indirect. GTC is an engineering and enterprise event, and performance marketing, traffic buying and partner programmes are not among its themes. It suits teams that build AI into their own work, for example in creative production, data processing or security tooling, and publishers who cover AI software and want product news from the source. Anyone hoping to sign advertisers or meet affiliate networks will find little of that here.