Nvidia CEO Jensen Huang Unveils Rubin AI Chips at GTC 2025
Nvidia CEO Jensen Huang introduced the company’s next-generation AI chips at GTC 2025, highlighting advancements in artificial intelligence and robotics.

Speaking at the event, known as the "Super Bowl of AI," Huang said AI is at an "inflection point" and predicted Nvidia’s data centre infrastructure revenue will reach USD 1 trillion by 2028.
Huang announced two new graphics architectures: Blackwell Ultra, set for release in the second half of 2025, and its successor, the Rubin AI chip, expected in late 2026. Rubin Ultra will follow in 2027.
During his keynote, which lasted over two hours, Huang discussed AI’s evolution from perception and computer vision to generative AI and now agentic AI, which can reason and understand context.
He said the next phase of AI is robotics, powered by "physical AI" that can grasp concepts like friction, inertia, and object permanence. He emphasised that each AI advancement creates new market opportunities.
Huang highlighted the role of synthetic data generation in AI training, stating that AI learns at speeds that make human-led training obsolete. He described reinforcement learning as a major breakthrough in recent years.
To support AI-driven robotics, Huang introduced Isaac GR00T N1, an open-source foundation model for humanoid robot development. It will work alongside an updated Cosmos AI model to generate simulated training data.
Benjamin Lee, a professor of electrical and systems engineering at the University of Pennsylvania, said simulated environments are crucial for reinforcement learning, as real-world data collection is costly and time-consuming.
Huang also discussed the Cosmos series of AI models, which generate cost-efficient, photo-realistic video for training robots and automated services. The models integrate with Nvidia’s Omniverse, a physics simulation tool.
General Motors plans to incorporate Nvidia’s technology into its new fleet of self-driving cars. The companies will collaborate on custom AI systems using Omniverse and Cosmos for AI manufacturing models.
Huang unveiled Halos, an AI system focused on automotive safety, particularly for autonomous driving. He claimed Nvidia is the first company to have every line of code safety assessed.
He concluded by introducing Newton, an open-source physics engine for robotics simulation, developed with Google DeepMind and Disney Research. A small robot named Blue joined him on stage, responding to his commands.
"The age of generalist robotics is here," Huang said.
Nvidia CEO Jensen Huang introduced new AI chips at GTC 2025
Blackwell Ultra launches in 2025, followed by Rubin AI in 2026 and Rubin Ultra in 2027
AI’s next phase is robotics, powered by synthetic data and reinforcement learning
Source: AP NEWS