Nvidia Launches Physical AI Models for Robots

2025-12-15 · Source: aibusiness · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Cloud Computing & IT Infrastructure · Depth: Intermediate, quick

Summary

Nvidia has launched a suite of new physical AI models, simulation frameworks, and edge computing hardware at CES 2026, aiming to accelerate the development of intelligent robots. Key releases include Nvidia Cosmos Transfer 2.5 and Cosmos Predict 2.5, open world models for simulating real-world physics and spatial dynamics, and Cosmos Reason 2, an open reasoning vision-language model for real-time decision-making. For humanoid robotics, Nvidia introduced Isaac GR00T N1.6, a vision-language-action model enabling full-body control. Additionally, the company unveiled open-source frameworks Isaac Lab-Arena for large-scale robot policy evaluation and OSMO for cloud-native robotic workflow orchestration. New Jetson Thor and IGX Thor platforms were also highlighted for humanoid and industrial edge computing, with partners like Neura Robotics and LG Electronics showcasing robots powered by these systems.

Key takeaway

For AI Architects and Machine Learning Engineers developing robotics solutions, Nvidia's new open physical AI models and simulation frameworks offer a significant opportunity to reduce development complexity and cost. You should explore the Cosmos models on Hugging Face and the Isaac Lab-Arena and OSMO frameworks on GitHub to accelerate your robot learning, reasoning, and deployment workflows, especially for safety-critical or humanoid applications.

Key insights

Nvidia's new physical AI models and frameworks aim to democratize intelligent robot development and deployment.

Principles

• Open models reduce robotics development cost.
• Simulation is crucial for safety-critical systems.
• Unified workflows accelerate development cycles.

Method

Nvidia's approach involves open physical AI models (Cosmos Transfer/Predict/Reason, Isaac GR00T), simulation/orchestration frameworks (Isaac Lab-Arena, OSMO), and edge hardware (Jetson Thor, IGX Thor) to enable real-world robot intelligence.

In practice

• Utilize Cosmos models for physics simulation.
• Deploy Isaac Lab-Arena for robot benchmarking.
• Integrate OSMO for cloud-native robot workflows.

Topics

• Physical AI
• Robotics
• Simulation Frameworks
• Edge Computing
• Vision-Language Models

Best for: AI Architect, Machine Learning Engineer, Investor, AI Engineer, Robotics Engineer, MLOps Engineer

Related on AIssential

• What’s Next in Robotics? — Physical AI, powered by NVIDIA's ecosystem, accelerates robotics development through advanced simulation and data generation.
• Open-Source Software Is Starting to Help Robots Think — Open-source AI is rapidly lowering barriers to advanced robotics development, accelerating innovation and accessibility.
• NVIDIA's Cosmos 3: The World's First Fully Open AI Omnimodel — NVIDIA's Cosmos 3 is an open, multimodal omnimodel integrating vision, world generation, and action prediction for physical AI applications.
• Nvidia Intros Data Factory, Robotics Models in Physical AI Push — Nvidia's new tools and models aim to overcome physical AI data collection challenges through synthetic data generation and advanced world models.
• Into the Omniverse: Physical AI Open Models and Frameworks Advance Robots and Autonomous Systems — Open-source physical AI models and simulation frameworks accelerate autonomous system development from design to deployment.
• As Open Models Spark AI Boom, NVIDIA Jetson Brings It to Life at the Edge — Edge AI platforms like NVIDIA Jetson enable real-time, private, and cost-efficient deployment of large AI models on physical systems.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by aibusiness.