From Prompt to Real-World Robot Skills With NVIDIA AI Agents
1 Video View·Nov 1, 2026
Building robots is hard.
Simulation makes training and testing easier,
but building these worlds takes time and expertise.
GPT-6 Astra makes this easier than ever.
Developers describe the assets, environment, and task
with a simple prompt and data.
Astra builds
the training world
using NVIDIA Open Libraries
and models as tools.
It replays real-world motion and tunes the physics
to get the simulation ready for training.
Next, Astra trains and refines robot skills using Isaac Lab and Newton,
and evaluates them at scale using Isaac Lab Arena.
Every tool is running locally
on NVIDIA DGX Station for faster iteration.
This is just one skill.
Developers are training robots to learn tasks of every kind,
at home, in hospitals, stores,
factories, and beyond.
Then, these learned skills move from simulation to the real world.
Introducing Agile One SN, a full-stack reference robot built
on NVIDIA Isaac Groot and powered by Jetson Thor,
giving developers a starting point to train and deploy new skills.
What once took a team of specialists can be done
by any developer with Astra and NVIDIA Isaac tools for agents.