TL;DR:
- NVIDIA unveils the Gecko foundation model at GTC 2026, specifically designed for humanoid robotics.
- Partnerships with ABB, FANUC, and Kuka aim to bridge the ‘reality gap’ in industrial automation.
- The initiative focuses on Physical AI, enabling robots to perceive and interact with complex environments in real-time.
The Leap Toward Physical AI
At the GTC 2026 conference, NVIDIA CEO Jensen Huang introduced a suite of technologies aimed at accelerating the deployment of humanoid robots. The centerpiece of this announcement is Gecko, a new AI foundation model trained on massive datasets of human movement and robotic simulations. Unlike previous models that focused on digital tasks, Gecko is optimized for physical interaction, allowing robots to learn complex maneuvers—such as navigating cluttered warehouses or performing delicate assembly—with unprecedented speed.
The move marks a significant shift from ‘Cyber AI’ to Physical AI. By leveraging NVIDIA’s Isaac robotics platform and Omniverse simulation tools, developers can now train robots in high-fidelity virtual worlds before deploying them to the physical realm. This ‘simulation-to-reality’ (Sim2Real) pipeline is critical for reducing the costs and risks associated with traditional robotic training.
Global Leaders Unite for Automation
To bring these capabilities to the factory floor, NVIDIA has strengthened its alliances with the world’s leading robotics manufacturers. ABB, FANUC, and Kuka have all announced plans to integrate NVIDIA’s AI chips and foundation models into their next generation of industrial robots. This collaboration is expected to transform sectors ranging from logistics to healthcare, where robots must operate safely alongside humans.
Background: NVIDIA and the Evolution of Robotics
Founded in 1993, NVIDIA has evolved from a graphics chip manufacturer into the primary engine of the global AI revolution. Under the leadership of Jensen Huang, the company has consistently pushed the boundaries of high-performance computing. Its entry into the robotics space began in earnest with the launch of the Isaac platform in 2018, which provided a comprehensive suite of software tools for robotic development.
By 2026, NVIDIA has positioned itself not just as a hardware provider, but as the foundational software layer for the ‘next wave’ of AI: autonomous machines. The robotics industry, traditionally dominated by rigid, pre-programmed systems from companies like ABB (Switzerland) and FANUC (Japan), is currently undergoing a paradigm shift. ABB, with its century-long history in power and automation, and FANUC, the world’s largest maker of industrial robots, are increasingly adopting AI-driven software to make their machines more flexible. This convergence of Silicon Valley’s AI prowess and the precision engineering of global industrial giants represents a pivotal moment in the quest for truly intelligent, autonomous robotics.
According to reports from NVIDIA Newsroom, these partnerships are not just about hardware, but about creating a unified ecosystem for robotic intelligence.
