Practical Guidelines

The platform, called NVIDIA Cosmos, combines physics-based modeling with geographical accuracy to help industries test and improve their automation systems.

Developers can now use the platform to create digital testing grounds that mirror real-world conditions. Robots and automated systems can use these virtual environments for training prior to warehouse deployment.

The Stargate Project will invest $500 billion over four years to develop advanced computing systems. Work begins in Texas, creating jobs and aiming to keep the US competitive in AI technology.

"This initiative is a critical step on the path to realizing the full potential of AI,"

The technology gives robots what developers call a "God's-eye view," allowing machines to better understand and navigate complex environments.

This advancement tackles a major challenge in robotics: helping machines plan their movements effectively in changing environments.

Zhutian

"Skye"

Developers can use tools in the platform to create and test AI models using computer-generated data to cut down the need for resource-intensive real-world testing.

Major companies including Uber, XPENG, and Figure AI have already started using Cosmos in their development programs.

The robotics company uses NVIDIA's Cosmos platform to create realistic video clips of different environments. This helps robots learn tasks like moving drinks from carts to fridges without needing physical practice spaces.

Training robots usually requires real-world practice, which takes time and money. Companies often build multiple test areas to teach robots different tasks.

Major robotics companies have developed new training methods that help robots learn tasks more quickly using generative AI technology. Agility Robotics, Amazon, Covariant, Robust AI, and Toyota Research Institute are testing these systems.

The new approach combines different types of data - from sensors, human operators, and online videos - to teach robots complex tasks.

Developers can now customize their own AgiBot with the X1 robot, an open-source platform for the modular machine.

Thanks to the X1's customizable components, users can expand its uses and modify its features to meet specific demands.

The company aims to supply hardware to robot manufacturers rather than build robots itself.

The company currently provides technology for Tesla's Optimus humanoid robot project.

NVIDIA has just rebranded its Jetson Orin Nano Developer Kit into the Jetson Orin Nano Super Developer Kit - a move that will come with its slate of “super” upgrades in line with ongoing developments in the field of generative AI. Developers need powerful platforms to deploy these advanced models effectively, and NVIDIA aims to meet that demand.

The latest upgrade introduces a new power mode that increases GPU, memory, and CPU clock speeds. As a result, the Jetson Orin Nano Super can deliver up to 70% higher generative AI performance. This allows it to handle advanced Transformer-based models, including those with up to 8 billion parameters, like Llama-3.1-8B. The compact kit is now available for just $249, making it an attractive option for developers looking to implement scalable AI solutions.