Nvidia closed out GTC 2026 in San Jose with a sweeping agenda centered on agentic Artificial Intelligence, physical Artificial Intelligence and robotics. CEO Jensen Huang delivered a nearly three-hour-long keynote and outlined a vision that stretched from data centers and gaming software to telecom networks and autonomous machines. New announcements included the Vera CPU, DLSS 5 for gaming graphics, and NemoClaw, a reference stack built for the OpenClaw agent platform. Partnerships with T-Mobile, Adobe and Disney underscored Nvidia’s effort to position its hardware and software as the foundation for the next phase of accelerated computing.
Huang said he expects to see $1 trillion in orders for its Blackwell and Vera Rubin systems through 2027, raising previous estimates. He also described a future Nvidia workforce of 75,000 employees working with 7.5 million AI agents, a vision that points to extremely heavy reliance on agentic systems inside the company itself. Nvidia only employs 36,000 people, according to a 2025 estimate. The conference repeatedly emphasized inference as the new center of demand, with Huang arguing that operating models in real time is now more important than training them. Nvidia said the new Vera CPU “delivers results with twice the efficiency and 50% faster than traditional CPUs.”
Robotics emerged as one of the event’s most visible themes. Nvidia and Disney demonstrated an Olaf droid based on the Frozen character, powered by Nvidia chips and simulations using the Newton Physics Engine developed with Google DeepMind and Disney Research. Nvidia also received attention for self-driving technology after MotorTrend named Huang its 2026 Person of the Year, citing the company’s work in software-defined vehicles. Nvidia’s Alpamayo models are designed for self-driving car software that can perceive, reason and act in real-world situations. The company also floated a more speculative idea with Vera Rubin Space-1, a computer for space, while acknowledging that cooling systems without conduction or convection remains a major obstacle.
The event also exposed tensions around Nvidia’s broader Artificial Intelligence agenda. NemoClaw was presented as a major step for autonomous agents, with Huang saying the OpenClaw movement is “as big of a deal as HTML” and “as big of a deal as Linux.” At the same time, concerns over privacy and security around highly capable agents remained unresolved. In gaming, DLSS 5 drew backlash from critics who said generative tools could flatten the distinctive visual style of games, though Huang rejected that criticism. Labor impacts remained another persistent concern, as Nvidia pushed a future where robots and agents fill shortages and augment workers, even as layoffs, workplace disruption and fears of an Artificial Intelligence bubble continue to shadow the industry.
