Building a Physical AI model factory with NVIDIA Cosmos 3 on SageMaker HyperPod
AWS ML Blog · rss · 2026-09-05
AWS ML Blog publishes an end-to-end guide for building a Physical AI model factory with NVIDIA Cosmos 3 on SageMaker HyperPod. Cosmos 3 uses a single token stream across video/image/action/sound with a Mixture-of-Transformers design (per-layer dual-stream attention joining a reasoner and generator) and train/inference asymmetry. One architecture runs three modes — forward-dynamics world model for synthetic data, inverse-dynamics action labeler, and deployable policy — across Nano (16B), Super (64B), and Edge (4B on-device) tiers. The pipeline shares one persistent GPU node pool, making GPU goodput the governing cost metric; a full robot-policy post-training walkthrough on DROID plus open-source manifests is included.
More from Embodied
- One policy controlling 4 different robot hands for in-hand manipulation rejected by CoRL — YuXiang_IRVL · 2026-09-05
- Shenzhen collapses hardware iteration to hours — why physical tech builders go to Huaqiangbei — zakelfassi · 2026-09-05
- HARBOR trains robot locomotion policies from a single prompt, fully autonomously in 1.5 hours — breadli428 · 2026-09-05
- Anthrobotics x Tnkr article explores soft robotics, artificial muscles and engineering tradeoffs — IanPritchard · 2026-09-05
- Voxel Grid Obstacle Avoidance on Jetson Orin Nano with a RealSense D436 — chrismatthieu · 2026-09-05
- HTC's Vive Eagle smart glasses let users pick ChatGPT or Gemini as the AI backbone — subvisser · 2026-09-05