Nvidia and Alphabet’s Intrinsic have partnered to showcase innovations in robotic grasping and industrial scalability.
The companies are set to transform the robotics landscape with the integration of advanced AI and platform technologies. At this week's Automate trade show in Chicago, Intrinsic is revealing significant advancements in robotic grasping and scalability, powered by Nvidia’s Isaac Manipulator and its AI capabilities.
The Isaac Manipulator, launched by Nvidia at GTC 2024 in March, marks a significant leap in industrial automation. It features foundation models and GPU-accelerated libraries designed for scalable, repeatable workflows in dynamic manipulation tasks. These models leverage transformer deep learning architecture, enabling robots to perceive and make decisions autonomously, mimicking human-like understanding.
The collaboration between Nvidia and Intrinsic aims to create a broadly applicable robotic-grasping skill, capable of functioning across various grippers, environments, and objects. Wendy Tan White, CEO of Intrinsic, emphasized the transformative potential of foundation models in simplifying processing challenges, cutting development costs, and enhancing user flexibility.
“For the broader industry, our collaboration with Nvidia demonstrates how foundation models can profoundly impact processing challenges at scale, foster new applications, reduce development costs, and increase end-user flexibility,” Tan White stated.
Tan White will deliver a keynote address at Automate discussing the implications of AI for innovation and growth, on Friday, May 9, at 9 a.m. CT.
The primary focus of this partnership is improving robotic grip. Grasping is a crucial yet historically difficult skill to program and scale. By integrating Nvidia Isaac Sim with the Omniverse platform, Intrinsic utilizes simulation to generate synthetic data for vacuum grasping, leading to notable advancements in industrial automation.
The prototype employs Intrinsic Flowstate, a developer environment for AI-based robotics solutions, to visualize processes, perception, and motion planning. Utilizing the Isaac Manipulator, developers can generate grasp poses and CUDA-accelerated robot motions, first evaluating them through simulation with Isaac Sim, which minimizes costs before real-world deployment using the Intrinsic platform.
Nvidia and Intrinsic are committed to enhancing dexterity and modular AI capabilities in robotic arms, offering a comprehensive suite of foundation models and GPU-accelerated libraries to expedite new robotics and automation tasks.
On Tuesday, May 7, Nvidia senior research scientist Adithya Murali and Intrinsic chief science officer Torsten Kroeger will present their collaborative efforts in the session “Automating Smart Pick-and-Place With Intrinsic Flowstate and NVIDIA Isaac Manipulator” at Automate.