Nvidia's Role in the Next Generation of Humanoid Robots

In this year’s NVIDIA GTC, the often repeated phrase is “physical AI”, which allows robots to operate effectively in the physical world.

To get there, NVIDIA offers its ISAAC GR00T N1, an open source basic model that has been pre-trained with some basics and broken down into fast systems for action and slow planning action systems.

To teach robots new tasks, many different data are needed. TV operations and video presentations are important, but are usually not enough. The synthetic data produced by training in the simulation helps fill the gap, which is the content of Nvidia’s Omniverse with cosmic content.

“Omniverse is a platform that can pool data between different sources and allows you to build a physically accurate digital twin,” said Akhil Docca, senior project marketing manager at Omniverse. “The universe is important because it allows not only data that enhances film and television realism, but also allows you to create an exponential large [amount] data. ”

This was illustrated in the process during the keynote address by NVIDIA CEO Jensen Huang, when the robot’s digital twins were displayed in a candy-like bright color in the Omniverse. With simple text prompts, the same image is displayed through several different iterations and loops through the lifetime and textures added by the universe.

In order to make this synthetic data suitable for the real world as much as possible, a detailed representation of the laws of physics is required. NVIDIA has collaborated with DeepMind and Disney to develop the open source physics engine Newton. During the keynote, Huang brought out a remotely controlled BDX robot to celebrate the news.

Much of the work on humanoid robots constitutes the answer to labour shortage anxiety. In his keynote, Huang predicted that by the end of the decade, there would be a shortage of 50 million workers worldwide and envisioned a future where we might go to work for robots “$50,000 a year.”

To make it clearer about Huang’s comments, I contacted NVIDIA, which told me that at the end of the decade, the estimated shortage of 50 million workers was based on a combination of estimates from various sources, industries and locations. Sources provided by NVIDIA include relevant manufacturing,,,,, Renewable energy,,,,, Truck Transport,,,,, Nursinghotel workers USA and Europedomestic workers us and Europeand construction workers us and Europe.

A NVIDIA representative also clarified the company’s expected pricing model for humanoid robots, saying: “It’s more like a full self-driving service to subscribe to self-driving cars. Customers buy cars and then pay for the services they want. We believe the robot will be the same person. Buy the robot and pay on the top.”

To see actions for these technologies, check out the videos in this article.