Nvidia GTC 2026 | Nvidia CEO Jensen Huang Keynote

NVIDIA founder and CEO Jensen Huang delivered the highly anticipated keynote address at GTC 2026 on Monday, March 16, 2026, at the SAP Center in San Jose, California. The nearly 2- to 3-hour presentation (described as packed with announcements) kicked off the company’s flagship AI developer conference, which ran March 16–19, 2026, both in-person and virtually, drawing over 30,000 attendees from 190+ countries. Huang’s central theme was the arrival of the agentic AI era and the shift from training-focused AI to inference, agentic systems, and physical AI. He declared that “the inference inflection has arrived,” positioning NVIDIA as the provider of the full-stack infrastructure — from chips and systems to AI factories — powering the next decade of computing. He repeatedly emphasized that NVIDIA is no longer just a chip company but the builder of the “AI factory floor” for the global economy. Major Announcements and Highlights - Explosive Demand Forecast — Huang revealed that NVIDIA has line-of-sight to $1 trillion in orders for Blackwell and the upcoming Vera Rubin systems through 2027 (double the previous outlook). He noted strong purchase orders, including for China, and highlighted 11 straight quarters of massive revenue growth. - Vera Rubin Platform — The star of the show was the next-generation Vera Rubin full-stack AI platform (successor to Blackwell), already in full production and shipping later in 2026. It includes: Vera Rubin GPUs (built on TSMC 3nm, dual-die design with ~336 billion transistors, massive HBM4 memory). - Vera CPU — A new purpose-built CPU for agentic AI workloads, delivering 2x efficiency and 50% faster performance than traditional CPUs. Multiple rack-scale systems (including NVL72) promising up to 10x better performance per watt and dramatically lower cost per token for inference. - Agentic AI Breakthroughs: OpenClaw — An open-source agentic AI framework described as “the operating system for AI agents.” NemoClaw — NVIDIA’s optimized stack for building, deploying, and running autonomous agents with a single command, including privacy sandboxes and integration with open models. Emphasis on moving agents from chatbots into real enterprise workflows, digital twins, and physical systems. - Physical AI & Robotics — Huang declared the “Big Bang of Physical AI” has begun. He showcased advancements in NVIDIA Cosmos (world models), Isaac GR00T (robotics foundation models), and partnerships with humanoid/robotics companies. Demonstrations included lifelike robots (including a Disney Olaf robot) and progress toward autonomous systems. Other Notable Reveals: DLSS 5 — Next-generation AI upscaling with “Neuro Rendering,” bringing photorealistic lighting and materials (coming fall 2026). Partnerships with Uber (autonomous fleets in 28 cities by 2028), OpenClaw, Groq (integration of Groq 3 LPUs with Vera Rubin for inference), Disney Imagineering, and others. Plans for space-based data centers and next-generation architectures like Kyber (preview for 2027). Huang’s signature energetic style, dramatic visuals, and optimistic outlook captivated the audience. He framed the moment as the beginning of a new computing paradigm where AI agents and physical intelligence reshape every industry. Subscribe: https://www.youtube.com/@techusiness Follow us on Social Media: Twitter / X https://x.com/Techusiness Threads https://www.threads.net/@techusiness Instagram https://www.instagram.com/techusiness/ TikTok https://www.tiktok.com/@techusiness LinkedIn https://www.linkedin.com/company/techusiness/ Substack https://techusiness.substack.com/ Contact Info Email techusiness@gmail.com