Introduction 🚀

We have reached a tipping point for generative AI. From then on, every corner of the world will have computing needs.

Nvidia, whose share price has just skyrocketed by $200 billion, is already prepared for this moment.

「大家好！We’re back！」

640.gif

DGX GH200

One of the most impressive announcements at the COMPUTEX 2023 keynote was the DGX GH200, a new class of large-memory AI supercomputer that can handle terabyte-class models for massive recommender systems, generative AI, and graph analytics. The DGX GH200 is powered by 256 NVIDIA GH200 Grace Hopper Superchips, each combining an Arm-based NVIDIA Grace CPU with an NVIDIA H100 Tensor Core GPU in one package. The GH200 Superchips are connected by the NVIDIA NVLink Switch System, a new interconnect that enables all GPUs in a DGX GH200 system to work together as one. This provides an unprecedented 1 exaflop of performance and 144 TB of shared memory, nearly 500x more memory than the previous generation NVIDIA DGX A100. The DGX GH200 is designed to enable the development of giant, next-generation models for generative AI applications, such as natural language processing, computer vision, speech synthesis, and more. Some of the first customers expected to gain access to the DGX GH200 include Google Cloud, Meta, and Microsoft.

r.png

NVIDIA Omniverse

Another exciting announcement at the COMPUTEX 2023 keynote was NVIDIA Omniverse, a computing platform that enables individuals and teams to develop Universal Scene Description (USD)-based 3D workflows and applications. Omniverse connects industry-leading 3D tools such as Autodesk Maya, Blender, Adobe Photoshop, and more, allowing users to sync their 3D data in a unified view and collaborate in real time. Omniverse also leverages NVIDIA RTX technology to provide photorealistic rendering and simulation for 3D worlds and digital twins. One of the use cases of Omniverse showcased at the keynote was how WPP, the world’s largest advertising company, is using Omniverse to create digital ad content in real time, using generative AI and synthetic data. Omniverse is available both on premises with NVIDIA RTX systems and in the cloud with NVIDIA Omniverse Cloud.

12122.png

LLM NPC Characters

The last but not least announcement at the COMPUTEX 2023 keynote was NVIDIA ACE for Games, a custom AI model foundry service that aims to transform games by bringing intelligence to non-playable characters (NPCs) through AI-powered natural language interactions. NVIDIA ACE for Games provides developers of middleware, tools and games with optimized AI foundation models for speech, conversation and animation, such as NVIDIA NeMo, NVIDIA Riva, and NVIDIA Omniverse Audio2Face . These models can be customized and deployed in their software and games across the cloud and PC. One of the benefits of NVIDIA ACE for Games is that it enables NPCs with persistent personalities that evolve over time and accurate facial animations and expressions, all in the user’s native tongue. To demonstrate the power of NVIDIA ACE for Games, NVIDIA partnered with Convai, an NVIDIA Inception startup, to showcase a demo called Kairos, where players can interact with Jin, a realistic NPC who runs a ramen shop. The demo is rendered in Unreal Engine 5 using ray tracing and NVIDIA DLSS.

Conclusion

Generative AI is transforming the way we create and consume content, and NVIDIA is at the forefront of this innovation with its new platforms and technologies. Whether you are an enterprise looking for the best performance and scalability for your AI workloads, a content creator looking for the best tools and workflows for your projects, or a gamer looking for the most realistic and immersive experiences, NVIDIA has something for you.