Nvidia has teamed up with major cloud service providers, including Oracle Cloud, Microsoft Azure, and Google Cloud, to introduce its AI supercomputers as a cloud service. Nvidia DGX Cloud will enable enterprises to use a web browser to instantly access Nvidia’s DGX AI supercomputers and AI software necessary to train models for generative AI and other AI applications.
Oracle has already launched Nvidia’s DGX Cloud service, while Azure is expected to follow suit in the next quarter. Nvidia executives also announced three cloud services, collectively called Nvidia AI Foundations, to allow businesses to construct and operate large language models and generative AI models.
The NeMo cloud service facilitates custom language text-to-text generative models, the Picasso cloud service is for generating images and video models, and BioNeMo accelerates life sciences research and drug discovery. All three services operate on DGX Cloud.
DGX Cloud and its AI software, including the AI Enterprise software suite, offer businesses the chance to rent Nvidia’s DGX AI supercomputers on a monthly basis, allowing them to scale their AI projects without the complexity and expense of deploying and managing their own infrastructure.
Alexander Harrowell, Omdia’s principal analyst of advanced computing for AI, commented that Nvidia’s launch of an AI supercomputing cloud service is a smart move. Nvidia’s new cuLitho software library for computational lithography, which uses less power and enables faster design, was also launched during the GTC AI developer conference.
It will allow the industry to build more powerful and energy-efficient next-generation processors that are “2nm and beyond.” Additionally, Nvidia introduced two new GPUs for specific generative AI inference workloads, including the L4 GPU and H100 NVL GPU.
The sources for this piece include an article in DataCenterKnowledge.