NVIDIA and Google Cloud have announced a collaboration to create what they call 'agentic physical AI factories' — systems that bring artificial intelligence into real-world industrial environments. The partnership aims to use simulation and digital twins to train AI agents for tasks in manufacturing, logistics, and other physical settings.
The companies describe the initiative as a step toward bridging the gap between digital AI models and physical execution. By combining NVIDIA's hardware and simulation platforms with Google Cloud's AI infrastructure, the project seeks to enable robots and automated systems to learn and operate in complex, dynamic environments.
Key aspects include:
- Use of digital twins to create virtual replicas of factory floors for training AI agents.
- Deployment of 'agentic' AI — systems that can perceive, decide, and act autonomously in the physical world.
- Focus on industrial automation, smart factories, and robotics transformation.
The announcement underscores a broader push by both companies to embed AI directly into physical operations, moving beyond purely digital applications.