NVIDIA Revolutionizes Robotics with Omniverse NuRec 3D Gaussian Splatting Libraries and Advanced AI Technologies

At the SIGGRAPH conference, NVIDIA announced significant advancements in its Omniverse platform, which aims to transform robotics development and deployment. The company introduced new libraries and models that enable developers anywhere to create physically accurate simulations, reconstruct the real world, and train AI agents that understand the physical world. This marks a crucial step towards developing robots and autonomous vehicles that can transform trillions of dollars in industries.

Key Takeaways:

• The new NVIDIA Omniverse NuRec 3D Gaussian Splatting Libraries enable developers to capture, reconstruct, and simulate the real world in 3D using sensor data, accelerating world composition applications.
• The Omniverse SDKs introduce data interoperability between MuJoCo (MJCF) and , allowing over 250,000 MJCF developers to seamlessly simulate robots across platforms.
• New Omniverse NuRec introduces Omniverse RTX ray-traced rendering, a technique that lets developers capture, reconstruct, and simulate the real world in 3D using sensor data.
• The Isaac Sim and Omniverse libraries are being adopted by companies such as Boston Dynamics, Figure AI, , RAI Institute, Lightwheel, and Skild AI to accelerate their AI robotics development.
• The NVIDIA Cosmos Models, downloaded over 2 million times, enable developers to generate diverse data for training robots at scale using text, image, and video prompts.
• The new Cosmos Transfer accelerates photorealistic synthetic data generation from ground-truth 3D simulation scenes or spatial control inputs like depth, segmentation, edges, and high-definition maps.
• The Cosmos Reason Breaks Through World Understanding and enables robots and vision AI agents to reason like humans, using prior knowledge, physics understanding, and common sense to understand and act in the real world.

Statistics:

• Over 2 million times, developers have downloaded the NVIDIA CosmosWFMs to generate diverse data for training robots at scale.
• The OpenUSD Curriculum and Certification program has been launched to address demand for USD expertise, with support from industry leaders such as Adobe, Amazon Robotics, and Ansys.
• The new Open-source collaboration with Lightwheel will integrate robot policy training and evaluation frameworks into the NVIDIA Isaac Lab, featuring parallel reinforcement learning training capabilities, benchmarks, and simulation-ready assets.
• The NVIDIA AI Infrastructure offers a single architecture for every robot development workload across training, simulation, and runtimes.

Sources:

*

*

*

*

*