ChipsSeptember 14, 2026via MarkTechPost
NVIDIA Open-Sources OSMO: One YAML Orchestrates Physical AI Training, Simulation, and Robot Testing
Why it matters
OSMO removes infrastructure friction from the physical AI buildout by letting robotics teams define cross-tier compute workflows declaratively. This is compute-stack enablement for the robot era — it matters because it accelerates how teams can move models from simulation to real hardware at scale.
Key signals
- OSMO: Kubernetes-native workflow orchestrator, Apache-2.0 licensed
- Routes jobs to GB200 clusters, Jetson AGX Thor, and intermediate tiers without infrastructure code
- Used internally for Project GR00T, Isaac Lab, Isaac Sim
- Single YAML definition for training, simulation, hardware-in-the-loop
- Latest release: 6.3.1
- Open-sourced from NVIDIA's internal stack
The hook
NVIDIA open-sources OSMO: one YAML file orchestrates training, sim, and hardware testing across GB200 clusters to edge Jetsons. Physical AI ops just got a lot simpler.
NVIDIA has open-sourced OSMO, the Kubernetes-native workflow orchestrator it uses internally for Project GR00T, Isaac Lab, and Isaac Sim. OSMO lets robotics teams define training, simulation, and hardware-in-the-loop tasks in a single YAML file and routes each one to the right compute tier, from GB2…