Nvidia CEO Jensen Huang introduced ‘Blue,’ an AI-driven robot developed in collaboration with Disney Research and Google DeepMind, at the GTC 2025 AI conference. Drawing inspiration from Star Wars, the small yet sophisticated robot made its way onto the stage and engaged in an interactive exchange with Huang.
“Hi Blue!” Huang greeted the robot before turning to the audience and revealing that what they were witnessing was an entirely real-time simulation. “This is how we are going to train robots,” he stated, emphasizing Blue’s cutting-edge technology, which includes two Nvidia computers housed within its compact frame.
Addressing the Global Labour Shortage with Robotics
Huang positioned AI-driven robotics as a game-changing solution to the growing labour shortage. He noted that by the end of the decade, the global workforce could face a deficit of at least 50 million workers, highlighting the critical role AI-powered robots could play in addressing this gap.
The CEO kicked off the keynote with an enthusiastic reflection on Nvidia’s breakthroughs. “What an amazing year it was, and we have a lot of incredible things to talk about. And I just want you to know that I am up here without a net. There are no scripts, there’s no teleprompter, and I have got a lot of things to cover. So let’s get started,” he remarked, as quoted by Business Insider.
Nvidia’s Next-Gen AI Hardware: Blackwell, Rubin, and More
While Blue captured attention, Nvidia’s latest AI hardware announcements remained a key focus for investors. The company is preparing for a major transition from its Blackwell AI chip architecture to the more advanced Blackwell Ultra, slated for launch later this year.
Additionally, Huang introduced two new AI supercomputers — DGX Spark and DGX Station — both powered by Nvidia’s Grace Blackwell platform. These systems aim to provide users with the ability to train and deploy large AI models with or without a data center connection.
Beyond Blackwell, Nvidia is set to launch its next-generation AI superchip platform, Rubin, named after astronomer Vera Rubin, whose research provided key evidence for dark matter. First previewed at Computex last year, the Rubin platform is scheduled to arrive in the second half of 2026, marking another significant leap in Nvidia’s AI computing roadmap.
