What’s Next in Robotics?
A representative of unknown name, role, and company discusses how NVIDIA tools support physical AI and robotics development, with examples from industrial machinery and field operations.
In this talk
- 0:02
Physical AI for machines
Advances in large AI models are bringing new capabilities into physical settings. The discussion points to roughly a century of operational data from job sites as a resource for scaling robotics.
- 0:39
NVIDIA infrastructure in the field
NVIDIA inference services, AI Factory, and its broader ecosystem help deliver AI capabilities to customers operating machines in the field. NVIDIA infrastructure is also part of the development process.
- 0:59
Simulation speeds development
Isaac Sim, Omniverse, and Isaac Lab are used to build digital twins and evaluate designs before constructing physical systems. Simulation also helps generate training data and improve development efficiency.
- 1:42
Training for unfamiliar scenarios
Cosmos supports testing and validation against situations not yet encountered in real operations. Cosmos Transfer can create varied synthetic training examples based on data recorded from human operators.
- 2:09
From simulation to production
Development can happen largely in simulation, followed by real-world adjustment of the deployed machine. AI can also help operators become productive sooner, while the next year may bring changes to manufacturing and robotics.
Briefs summarise talks and interviews; timestamps open the video at that moment. Claims in a talk are the speaker's own.