NVIDIA Unveils Physical AI Models, Partners Launch Next-Gen Robots

NVIDIA today announced new open models, frameworks, and AI infrastructure for physical AI at CES 2026, alongside global partners including Boston Dynamics, Caterpillar, Franka Robotics, Humanoid, LG Electronics, and NEURA Robotics debuting next-generation AI-driven robots.

The company released NVIDIA Cosmos Transfer 2.5 and Predict 2.5 world models for synthetic data generation and robot policy evaluation, alongside Cosmos Reason 2—a reasoning vision language model enabling machines to see, understand, and act in physical environments. NVIDIA also introduced Isaac GR00T N1.6, a purpose-built model for humanoid robots, and the OSMO edge-to-cloud compute framework to streamline robot training workflows.

NVIDIA's official announcement states these tools allow developers to bypass resource-intensive pretraining, focusing instead on creating next-generation AI robots. CEO Jensen Huang declared, "The ChatGPT moment for robotics is here. Breakthroughs in physical AI — models that understand the real world, reason and plan actions — are unlocking entirely new applications."

Integration with Hugging Face's LeRobot library accelerates open-source robotics development, while the Blackwell architecture-powered Jetson T4000 module delivers 4x greater energy efficiency and AI compute for edge deployment. Partners like Salesforce use Agentforce, Cosmos Reason, and NVIDIA Blueprint to analyze robot-captured video footage, reducing incident resolution times by 2x, and LEM Surgical employs Cosmos Transfer to train autonomous surgical arms for its Dynamis robot.

Industry leaders including ABB Robotics, FANUC, KUKA, and Universal Robots are adopting NVIDIA's simulation frameworks for virtual commissioning and edge AI inference, signaling a shift toward "generalist-specialist" robots capable of mastering multiple tasks with minimal retraining. As Huang emphasized, "NVIDIA's full stack of Jetson robotics processors, CUDA, Omniverse and open physical AI models empowers our global ecosystem of partners to transform industries with AI-driven robotics."

Artūras Malašauskas is an AI Systems Integrator with 20+ years of production-grade web engineering experience. He has designed, shipped, and scaled enterprise Python/PHP systems for logistics, SaaS, and public-sector clients. For the past year, he has focused exclusively on AI integrations: deploying open-source LLMs, building generative media pipelines (image, audio, video), and engineering multi-agent workflows for real production environments. His standard: reproducibility, security, cost-efficient inference—no vaporware. He documents and evaluates emerging AI tooling, separating verified capabilities from marketing noise. Technical editor at: muza-ai.eu, ai-verslas.lt, ai-naujinos.lt Connect on LinkedIn