Recently, GPU inventor Nvidia has been all-in on its generative AI advancements which played a big role in helping the company become a trillion-dollar behemoth. In an ever more digital and hyperconnected world, generative AI is making its presence felt. This has forced many tech giants to make significant changes to their products with AI offerings.
Jensen Huang, CEO and founder of Nvidia has appeared several times at this year’s annual conference on computer graphics SIGGRAPH. “The generative AI era is upon us; the iPhone moment if you will,” Huang told an audience during an in-person special address in Los Angeles.
AI advances
A lot of the company’s future is moving towards generative AI, bringing together all of the past decade’s innovations from AI, virtual worlds, acceleration, simulation, collaboration and more.
“Graphics and artificial intelligence are inseparable; graphics needs AI and AI needs graphics,” Huang said. The CEO went on to explain that AI will not only learn skills in virtual worlds but will also help create these virtual worlds.
To sustain the progress of AI, Nvidia introduced the Grace Hopper Superchip, denoted as the NVIDIA GH200. This superchip brings together a 72-core Grace CPU along with a Hopper GPU, and it went into full-scale production in May. Nvidia says it is, “Built to handle the world’s most complex generative workloads, spanning large language models, recommender systems and vector databases, the new platform will be available in a wide range of configurations.”
The company adds that manufacturers are expected to release systems based on the platform in the second quarter of 2024.
AI workbench
The company’s Workbench is a unified toolkit that helps developers quickly create, test and fine-tune generative AI models on a PC or workstation. Once the models are ready, they can be scaled to virtually any data centre, public cloud or Nvidia DGX Cloud, making it easier for enterprises to adopt generative AI for their applications.
“With AI Workbench, developers can customise and run generative AI in just a few clicks. It allows them to pull together all necessary enterprise-grade models, frameworks, software development kits and libraries into a unified developer workspace,” said Nvidia.
“Everybody can do this,” adds Huang.
“Leading AI infrastructure providers – including Dell Technologies, Hewlett Packard Enterprise, HP Inc., Lambda, Lenovo and Supermicro – are embracing AI Workbench for its ability to bring enterprise generative AI capability to wherever developers want to work,” the company said.
OpenUSD advances
To offer new foundation applications and services for developers and industrial enterprises to optimise and enhance their 3D pipelines with Nvidia’s OpenUSD framework and generative AI, CEO Huang shared the latest info on the Nvidia Omniverse, an OpenUSD-native development platform for building, simulating, and collaborating across tools and virtual worlds.
The company says its Omniverse platform has been updated with new features and capabilities, including improvements to Omniverse Kit, the engine for developing native OpenUSD applications and extensions.
“Omniverse users can now build content, experiences and applications that are compatible with other OpenUSD-based spatial computing platforms such as ARKit and RealityKit,” said Nvidia.
Huang also revealed four new Omniverse Cloud APIs built by the company which developers can use to implement and deploy OpenUSD pipelines and applications. The Cloud APIs include:
- ChatUSD – An LLM copilot for developers that can answer USD knowledge questions or generate Python-USD code scripts
- RunUSD – A cloud API that translates OpenUSD files into fully path-traced rendered images by checking the compatibility of the uploaded files against versions of OpenUSD releases
- DeepSearch – An LLM agent enabling fast semantic search through massive databases of untagged assets
- USD-GDN Publisher – A service that enables enterprises and software makers to publish high-fidelity, OpenUSD-based experiences to the Omniverse Cloud Graphics Delivery Network (GDN)
Nvidia says the updates showcased in Omniverse foundation applications, are fully customisable reference applications that creators, enterprises and developers can copy, extend or enhance.
The improved Omniverse apps include Omniverse USD Composer, enabling 3D users to construct extensive scenes using OpenUSD. Omniverse Audio2Face now offers generative AI APIs for lifelike facial animations from audio files, with added multilingual support and a new female base model.
Building out the Omniverse platform
Nvidia’s extended reality tool allows users to integrate spatial computing capabilities directly into their Omniverse-driven applications. It also gives users the versatility to engage with their 3D projects and virtual environments in their preferred manner.
“New application and experience templates provide developers getting started with OpenUSD and Omniverse a major headstart with minimal coding,” said Nvidia. The company adds that developers and industry enterprises now have new foundation apps and services to use in optimising and enhancing 3D pipelines using the OpenUSD framework and generative AI.
In addition, Omniverse users have the opportunity to leverage the recently introduced NVIDIA L40S GPU to enhance the performance of demanding graphics. Nvidia’s Omniverse is currently accessible as a beta version and is free to download.
Expanding partnerships
The GPU manufacturer has partnered with AI startup Hugging Face to bring generative AI supercomputing to the fingertips of developers building large language models and other advanced AI applications.
Nvidia says developers can access its DGX Cloud AI supercomputing within the Hugging Face platform in training and advancing AI models. “This is going to be a brand new service to connect the world’s largest AI community to the world’s best training and infrastructure,” Huang said.
The company has also collaborated with luxury EV brand Denza to build and deploy its next generation of car configurators, Huang announced. Shutterstock is also incorporating generative AI into 3D scene backgrounds by utilising a foundational model trained using Nvidia Picasso. This cloud-based platform serves as a creative space for constructing visual generative AI models.
Isa Muhammad is a writer and video game journalist covering many aspects of entertainment media including the film industry. He's steadily writing his way to the sharp end of journalism and enjoys staying informed. If he's not reading, playing video games or catching up on his favourite TV series, then he's probably writing about them.
