nvidianews.nvidia.com via Reddit

NVIDIA Releases Cosmos 3, Open Model for Physical AI

6 sources tracking this story
nvidia robotics open source physical-ai open-source robotics

Key insights

  • Cosmos 3 ships as 16B Nano (RTX PRO 6000 edge-compatible) and 64B Super (Hopper/Blackwell), with a Cosmos 3 Edge variant arriving soon.
  • Cosmos 3 uses the Linux Foundation's OpenMDW 1.1 license, explicitly permitting commercial use and targeting enterprise adopters wary of restrictive model terms.
  • Six open synthetic datasets accompany the weights on Hugging Face, covering robot manipulation, autonomous driving, warehouse automation, spatial reasoning, physics simulation, and human motion.

Why this matters

NVIDIA's Cosmos 3 release is structured as an ecosystem capture play: open model weights paired with the Linux Foundation's OpenMDW 1.1 commercial license, complete training recipes on GitHub, and six synthetic datasets lower the adoption cost to near zero for labs that cannot build world models in-house. The Cosmos Coalition (Agile Robots, Runway, Black Forest Labs, Skild AI) formalizes a software partner ring around NVIDIA hardware, while named industrial adopters including LG Electronics, Samsung, Li Auto, and Doosan Robotics confirm the strategy is reaching production commitments, not just developer interest. The 16B Nano variant compatible with RTX PRO 6000 extends the physical AI stack from data center training to edge deployment, tightening the hardware dependency at both ends of the inference pipeline.

Summary

NVIDIA released Cosmos 3 on May 31, an open foundation model for physical AI. Built on a mixture-of-transformers architecture, it processes text, images, video, ambient sound, and robot actions in one system, with NVIDIA claiming training cycles drop from months to days. The architecture pairs a reasoning module with expert generation components that understand object interactions and spatial-temporal relationships before producing video and action outputs. Essentially: (LG Electronics, Samsung Electronics, Li Auto) are named early adopters spanning robotics, autonomous vehicles, and vision AI. - Cosmos 3 Super (physics accuracy), Nano (fast reasoning), Edge (forthcoming, real-time inference). - Cosmos Coalition includes Agile Robots, Black Forest Labs, Generalist, LTX, Runway, Skild AI. An open model spanning sensing, reasoning, and physical action is a wider bet than prior open-source AI releases.

Potential risks and opportunities

Risks

  • LG Electronics and Samsung Electronics face liability exposure if Cosmos 3-powered robots cause physical harm before safety standards catch up to the model's shortened development cycles
  • Li Auto and other AV adopters risk regulatory scrutiny if Cosmos 3 world generation outputs contain spatial-temporal errors that propagate into autonomous vehicle decisions
  • Open distribution of a model generating robot action sequences creates dual-use risks if adversarial actors fine-tune Cosmos 3 Nano for physical systems outside intended parameters

Opportunities

  • Agile Robots and Skild AI, as Cosmos Coalition members, gain preferential access to model updates and can ship validated robotics products ahead of non-coalition competitors
  • Vision AI companies named as early adopters (Milestone Systems, Fogsphere, Linker Vision) can consolidate on Cosmos 3 as a shared perception backbone, reducing parallel model maintenance costs
  • Black Forest Labs, LTX, and Runway gain NVIDIA ecosystem exposure as coalition partners, positioning their generation platforms at the intersection of media AI and physical AI tooling

What we don't know yet

  • Whether Cosmos 3 Edge has a confirmed release timeline beyond 'forthcoming,' and which edge hardware platforms it targets
  • What commercial licensing terms govern Cosmos 3 model weights, given that 'open-source' covers a wide range of actual usage rights
  • How Cosmos Coalition members Generalist and LTX contribute technically versus serving primarily as adoption validators for the announcement

What others are reporting

Coverage cluster as of 8h after publish

  1. NVIDIA Blog Read →

    First-party post by NVIDIA researcher Ming-Yu Liu details real deployments: Agile Robots' Thor 3 and FR3 humanoids, Linker Vision smart city camera analysis, and NVIDIA GEAR team's embodied agent work.

  2. NVIDIA Technical Blog Read →

    Developer-facing post covers NIM microservices with BF16, FP8, and NVFP4 quantization, vLLM integration, and the HUE benchmark for evaluating physical reasoning quality.

    Cosmos 3 is a frontier foundation model for physical AI that combines physical reasoning, world generation, and action generation.
  3. GamesBeat Read →

    Captures Jensen Huang's 'big bang of physical AI' framing and details all six Cosmos Coalition founding members alongside industrial commitments from Doosan, Samsung, and Li Auto.

    The big bang of physical AI is just around the corner thanks to breakthroughs in multimodal reasoning language, vision and world models.
  4. CryptoBriefing Read →

    Frames the release through market accessibility: open Cosmos 3 removes the synthetic data generation bottleneck that has kept physical AI confined to well-funded labs.

    Cosmos 3 can generate predictive video sequences of up to 30 seconds based on text, image, or video inputs.
  5. Gizchina Read →

    Notes that the release bundles model weights, training scripts, deployment tools, and datasets simultaneously, framing Cosmos 3 as a reproducible ecosystem rather than a model drop.

    A robotic system trained on Cosmos 3 doesn't just see the world; it simulates what the world will look like a second from now.

Originally reported by nvidianews.nvidia.com

Read the original article →

Original headline: NVIDIA Launches Cosmos 3, World's First Fully Open Physical AI Omnimodel, at GTC Taipei — Unifies Vision Reasoning, World Generation, and Action Prediction