Join top executives in San Francisco on July 11-12, to hear how leaders are integrating and optimizing AI investments for success. Learn More
Nvidia announced on Tuesday at its annual developer conference — GPU Technology Conference (GTC) in Las Vegas — that it’s offering a new set of cloud services that will allow businesses to create and use their own AI models based on their proprietary data and specific needs.
The new services, called Nvidia AI Foundations, include three major components and are meant to accelerate enterprise adoption of generative AI: Enterprises can use Nvidia NeMo language service or Nvidia Picasso image, video and 3D service to gain access to foundation models that can generate text or images based on user inputs. The company also announced new models for its Nvidia BioNeMo, which was launched last year for the life-science industry. The service allows users to generate scientific texts based on biological data.
>>Follow VentureBeat’s ongoing Nvidia GTC spring 2023 coverage<<
Enterprises can now customize these three foundational models with their own data and domain knowledge to build applications for various purposes, such as intelligent chat and customer support, professional content creation and digital simulation.
Join us in San Francisco on July 11-12, where top executives will share how they have integrated and optimized AI investments for success and avoided common pitfalls.
The company said that several enterprises were already using its cloud services to build generative AI applications for various purposes such as intelligent chat and customer support, professional content creation and digital simulation. Among the brands using these new tools are Adobe, Getty Images, Morningstar, Quantiphi and Shutterstock.
“It’s a collection of services that are all the same architecture, the same building blocks, but for different modalities,” Manuvir Das, VP of enterprise computing at Nvidia, said during a press briefing detailing the AI Foundations effort.
Accelerating enterprise adoption of generative AI
Nvidia AI Foundations services are a set of cloud services hosted in the public cloud.
Das explained that the AI Foundations services provide a platform that helps enterprise companies create, customize, optimize and deploy custom models for AI. He emphasized the AI Foundations offering is not a new generative AI offering from Nvidia; rather, he referred to it as a foundry that allows companies to create their own models that they can then use in their generative AI services.
The three AI Foundations offerings are all making use of the Nvidia AI Enterprise software suite, as well as Nvidia’s GPU hardware running in the cloud.
Nvidia builds gen AI services with key partnerships
While Nvidia has spoken at length about NeMo and BioNeMo in the past, Picasso is an entirely new release. Das said Picasso can be used to produce images and videos, as well as 3D models. The 3D models can be generated in the Universal Scene Description (USD) format that enables component portability. USD is also supported in the Nvidia Omniverse metaverse platform, where the models created by Picasso could be used as assets to create any kind of environment.
Though Picasso might sound like it’s competitive with other generative AI services for image creation, like Open AI’s DALL-E, or Stable Diffusion, Das stressed that’s not the intention. Rather, the goal is to enable organizations to build their own custom models and train them on the data they already have.
To that end, as part of the AI Foundations launch, Nvidia also announced partnerships with Getty Images, Shutterstock and Adobe to help with generative AI.
Das commented that Getty Images has a large corpus of imagery that it has been properly sourced from creators. He said Getty is working with Nvidia to train their own models using all stock imagery to help build 2D images and video.
Shutterstock similarly has a large repository of images as well as 3D models. Das said that Shutterstock is working with Nvidia now specifically to help generate 3D models.
The partnership with Adobe is intended to help Adobe build out a new generation of AI models as it grows its own generative AI services including Firefly, which is a multimodel approach to image generation.
“The emphasis with our service is customization,” Das said. “Any customer can either start from the ground up, or they can take one of our pretrained models in order to get a head start.”
VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.