Nvidia, a leading technology company in the field of artificial intelligence (AI), has announced new services aimed at accelerating humanoid robot development globally. The company is providing a suite of services, models, and computing platforms to support the development, training, and building of the next generation of humanoid robots. These offerings include the Nvidia NIM microservices and frameworks for robot simulation and learning, the Nvidia Osmo orchestration service for running multi-stage robotics workloads, and an AI- and simulation-enabled teleoperation workflow for training robots.
Nvidia CEO Jensen Huang made these announcements during a talk at the Siggraph computer graphics conference in Denver, Colorado. Huang emphasized the importance of humanoid robots in the next wave of AI and expressed his excitement about advancing the entire Nvidia robotics stack. He believes that opening access to the platforms, acceleration libraries, and AI models will benefit worldwide humanoid developers and companies.
To accelerate development, Nvidia is introducing the NIM microservices and Osmo. NIM microservices are pre-built containers powered by Nvidia inference software that reduce deployment times significantly. They enhance simulation workflows for generative physical AI in Nvidia Isaac Sim, a reference application for robotics simulation. The MimicGen NIM microservice generates synthetic motion data based on recorded teleoperated data, while the Robocasa NIM microservice generates robot tasks and simulation-ready environments. On the other hand, Osmo is a cloud-native managed service that simplifies robot training and simulation workflows, reducing deployment and development cycle times from months to under a week.
Another challenge in humanoid robot development is the need for an enormous amount of data for training foundation models. Nvidia addresses this issue with an AI- and Omniverse-enabled teleoperation workflow. This workflow allows researchers and AI developers to generate synthetic motion and perception data from a minimal amount of teleoperated demonstrations. By using Apple Vision Pro to capture a small number of demonstrations, simulating the recordings in Nvidia Isaac Sim, and using the MimicGen NIM microservice to generate synthetic datasets, developers can train the humanoid foundation model with real and synthetic data, saving time and reducing costs.
Companies like Fourier, a general-purpose robot platform company, recognize the benefits of using simulation technology to synthetically generate training data. Fourier’s CEO, Alex Gu, believes that Nvidia’s new simulation and generative AI developer tools will help bootstrap and accelerate their model development workflows.
Nvidia provides three computing platforms to support humanoid robotics development: Nvidia AI supercomputers for training models, Nvidia Isaac Sim for learning and refining skills in simulated worlds, and Nvidia Jetson Thor humanoid robot computers for running the models. Developers can access and use all or part of these platforms based on their specific needs.
To expand access to Nvidia humanoid developer technologies, the company has launched the Nvidia Humanoid Robot Developer Program. This program allows developers to gain early access to the new offerings and the latest releases of Nvidia Isaac Sim, Nvidia Isaac Lab, Jetson Thor, and Project GR00T general-purpose humanoid foundation models. Several companies have already joined the early-access program, including Boston Dynamics, ByteDance Research, Field AI, and RobotEra.
In conclusion, Nvidia’s new services and developer program aim to accelerate humanoid robot development by providing advanced tools, models, and computing platforms. By simplifying and enhancing various aspects of the development process, Nvidia is poised to drive innovation and progress in the field of AI and robotics. Developers now have access to powerful resources that can significantly reduce deployment times, streamline workflows, and generate synthetic data for training models. With the support of Nvidia, the future of humanoid robots looks promising.