Key insights

• ⚙️ Introducing Blackwell platform with 28 billion transistors and 10 terabytes of data per second between two dies
• 🚀 Efficient ramping for systems like hgx configuration, designed to be form fit function compatible with Hopper
• 💻 Development of processor for generative AI and introduction of MVY Link Switch with high-speed communication capabilities
• 🔌 Enabling an exaflops AI system in one rack and forming partnerships with various companies to accelerate AI
• 🔗 Collaboration with Google and Microsoft to accelerate various aspects including data processing, AI services, and hosting Nvidia's ecosystem
• ⚙️ Nvidia AI Foundry offers optimized and packaged software for AI with collaborations with major companies like SAP, ServiceNow, and Cohesity
• 🏗️ Collaboration with Snowflake, NetApp, and Dell in building co-pilots, AI factories, and end-to-end systems for large-scale Enterprise AI
• 🤖 Introduction of Nvidia Project Groot, an AI-powered humanoid robot model for everyday tasks and interconnected tools like Isaac Lab and Osmo

Q&A

• What is Project Groot by Nvidia?

  Project Groot is an AI-powered humanoid robot model for everyday tasks, equipped with interconnected tools like Isaac Lab and Osmo, powered by Jetson Thor robotics chips. The project also showcases special guests, Disney's bdx robots, demonstrating learning and interaction. Blackwell is introduced as a powerful processing and networking system.

• What collaborations are taking place with Snowflake, NetApp, Dell, and Azure Cloud?

  Snowflake collaborates with Nvidia AI Foundry and NetApp to build co-pilots and AI factories, while Dell offers end-to-end systems for large-scale Enterprise AI. Additionally, Omniverse and OVX in Azure Cloud represent the future of digital twins for heavy industries, enabling connected productivity and unified ground truth.

• What does Nvidia's AI Foundry offer?

  Nvidia's AI Foundry provides optimized and packaged software for AI, including services such as Nvidia Inference Microservice (Nims), Nemo microservice, and DGX Cloud. Moreover, they are collaborating with major companies like SAP, ServiceNow, and Cohesity to build AI solutions for various applications.

• How is Nvidia collaborating with Google and Microsoft?

  Nvidia is working with Google and Microsoft to accelerate various aspects of their platforms, including data processing, AI services, and hosting Nvidia's ecosystem. This involves optimizing and packaging pre-trained models to run across Nvidia's large install base.

• What is the MVY Link Switch?

  The MVY Link Switch is a powerful chip developed by Nvidia to enable high-speed communication between GPUs, leading to an exaflops AI system in one rack.

• What is Blackwell?

  Blackwell is a platform designed with 28 billion transistors and 10 terabytes of data per second, form fit function compatible with Hopper, and provides efficient ramping for systems like hgx configuration.

• 00:01 Introducing Blackwell, a platform with 28 billion transistors and 10 terabytes of data per second, designed to be form fit function compatible with Hopper. Blackwell provides efficient ramping for systems like hgx configuration.
• 02:44 Nvidia has developed a new processor for generative AI and a powerful chip called the MVY Link Switch. This chip enables high-speed communication between GPUs, leading to an exaflops AI system in one rack. Nvidia has partnered with various companies to accelerate AI and integrate its technology into their systems.
• 05:11 Google and Microsoft are working with Nvidia to accelerate various aspects of their platforms, including data processing, AI services, and hosting Nvidia's ecosystem. This collaboration involves optimizing and packaging pre-trained models to run across Nvidia's large install base.
• 07:09 Nvidia's AI Foundry offers optimized and packaged software for AI, including Nvidia Inference Microservice (Nims), Nemo microservice, and DGX Cloud. They are collaborating with major companies such as SAP, ServiceNow, and Cohesity to build AI solutions for a variety of applications.
• 09:12 Snowflake collaborates with Nvidia AI Foundry and NetApp to build co-pilots and AI factories, while Dell offers end-to-end systems for large-scale Enterprise AI. Omniverse and OVX hosted in Azure Cloud represent the future of digital twins for heavy industries, enabling connected productivity and unified ground truth.
• 11:49 Nvidia presents Project Groot, an AI-powered humanoid robot model for everyday tasks, with interconnected tools like Isaac Lab and Osmo, powered by Jetson Thor robotics chips. Special guests, Disney's bdx robots, demonstrate learning and interaction. Blackwell is introduced as a powerful processing and networking system.