Nvidia GTC 2026 starts March 16 with Jensen Huang keynote

In short: Nvidia’s GTC 2026 conference starts next week, and CEO Jensen Huang is expected to share new products focused on running AI systems.

What happened

Nvidia’s GTC 2026 conference begins Monday, March 16 in San Jose, California. Jensen Huang’s keynote talk is set for 11:00 AM PT and will be livestreamed from the SAP Center. The keynote is expected to last about two hours.

Reports say Nvidia may use the keynote to show new hardware and software. A key theme is a shift in spending from training AI models to inference, which means using a trained model to produce answers (like using a finished cookbook to cook meals, rather than developing the recipes). Inference usually uses less power than training, but it is becoming a crowded market as companies like Google, Amazon, and Meta design their own chips.

There are also rumors about “NemoClaw,” described as an open-source platform for building “AI agents,” which are AI tools that can carry out multi-step tasks on their own (like a helper that can follow a checklist without being prompted for every step). Another reported item is a roughly $20 billion licensing deal tied to Groq, a company known for chips aimed at fast inference, and its team joining Nvidia to help scale that technology.

GTC runs March 16 to 19 and includes more than 1,000 sessions, demos, labs, and partner announcements. The agenda highlights AI uses in healthcare, robotics, self-driving car testing through simulation, climate research, and “digital twins,” which are virtual copies of real-world systems.

Why it matters

For most people, this matters because the cost and speed of running AI affects the tools you use every day, from search to customer support. Nvidia’s announcements can influence what AI services cost, how quickly they respond, and where companies invest next.

Source: Financial Times