Creating a New World with NVIDIA's DGX Supercomputer: Generative Artificial Intelligence on Oracle Cloud

Nvidia announced that its advanced artificial intelligence supercomputing platform, Nvidia DGX Cloud, is now available on the Oracle Cloud Marketplace along with its Nvidia AI Enterprise software suite. Nvidia DGX Cloud is a cloud-hosted AI supercomputing service that allows customers to immediately access everything they need to start training powerful generative AI and other applications. Nvidia announced this service at the GTX conference in March. The service is based on Nvidia's popular DGX platform, which is a dedicated hardware product that enterprises can purchase and run in their own on-premises data centers. With DGX Cloud, enterprises no longer need to wait to purchase expensive and often scarce platforms, but can instead rent the infrastructure they need for their AI workloads on a monthly basis. In other words, Nvidia enables more enterprises, including those unable to deploy and manage their own systems, to use its AI supercomputers. Nvidia explains that each DGX Cloud instance has access to eight 80 GB Tensor Core GPUs, meaning each node has 640 GB of GPU memory. The platform is built on a high-performance, low-latency network architecture to ensure workloads can scale across interconnected clusters of systems. This allows multiple DGX Cloud instances to act as a massive GPU to handle the most demanding workloads. The DGX Cloud platform works in conjunction with Nvidia's AI Enterprise software, which provides customers with over 100 AI frameworks and pre-trained models to build, optimize, and run custom large language and other generative AI models trained on their proprietary data for specific domain tasks. Nvidia states that customers can now access Nvidia DGX Cloud AI through the Oracle Cloud Marketplace. They will use the Nvidia Base Command Platform to interact with the hardware, which provides AI supercomputing capabilities to developers through a web browser. By offering a single-window view of customers' AI infrastructure, the Base Command Platform simplifies the management of multi-node clusters. Nvidia states that customers will be able to quickly build their generative AI applications and easily deploy them on the Oracle Cloud Infrastructure or their own on-premises platforms, simplifying the adoption of hybrid cloud strategies. To coincide with this release, Nvidia shared a customer success story, stating that the University at Albany, State University of New York, recently used Nvidia DGX Cloud AI on Oracle's infrastructure as the foundation of its AI Plus program. The program integrates AI education and learning into the university's research and academic endeavors, driving innovation in areas such as cybersecurity, weather forecasting, health data analysis, drug discovery, and next-generation semiconductor design. The University at Albany states that it is building its own on-premises supercomputer while using Nvidia's platform on Oracle's cloud. Thenkurussi Kesavadas, Vice President for Research and Economic Development at the University at Albany, said, "We are accelerating our mission to integrate AI into almost every academic and research domain. We will advance healthcare, security, and economic competitiveness while providing students with skills for the ever-changing job market."