NVIDIA announced today the launch of NVIDIA AI Foundry, a service that supports companies looking to enhance their artificial intelligence (AI) capabilities by building customized models tailored to specific industry needs.
The NVIDIA AI Foundry provides businesses with the necessary tools and infrastructure to create and deploy custom AI models. Through this service, companies can access NVIDIA’s DGX Cloud, foundation models, NVIDIA NeMo software, expert NVIDIA guidance, and ecosystem support tools. NVIDIA describes this service as similar to how TSMC manufactures chips designed by other companies, with the distinction that the NVIDIA AI Foundry focuses on creating customized AI models rather than semiconductor chips.
The AI Foundry enables businesses to adapt and enhance various NVIDIA and open-source models for their specific uses, including Meta’s new Llama 3.1 collection, Google DeepMind’s CodeGemma and Gemma, Meta’s Code Llama, Mistral, Mixtral, Microsoft’s Py-3, and Starcoder2.
Prominent companies such as Amdocs, Capital One, Getty Images, KT, Hyundai Motor Company, SAP, ServiceNow, and Snowflake are among the early adopters of the NVIDIA AI Foundry. These industry leaders utilize this service to drive innovation in enterprise software, technology, communications, and media sectors.
The core of NVIDIA AI Foundry’s offering includes the NVIDIA DGX Cloud, which provides a network of accelerated computing resources co-designed with major public cloud providers like Amazon Web Services (AWS), Google Cloud, and Oracle Corporation Cloud Infrastructure. This enables AI Foundry customers to develop and fine-tune generative AI applications with great efficiency and minimal upfront hardware investment. The service supports the scalable implementation of AI initiatives, providing a robust platform for rapidly tracking AI foundation models and model development within the AI community.
