From NVIDIA: 2024-09-11 09:00:51

At the Oracle CloudWorld conference, Oracle Cloud Infrastructure (OCI) announced the first zettascale OCI Supercluster, accelerated by NVIDIA Blackwell platform, with over 100,000 GPUs supporting enterprise AI workloads. Previewed instances include GB200 NVL72 for generative AI and HGX H200 for real-time inference. OCI also announced general availability of L40S GPU-accelerated instances for midrange AI workloads and visualization.

Oracle Autonomous Database now supports NVIDIA GPU to accelerate data processing workloads. NVIDIA and Oracle demonstrated capabilities for generative AI retrieval-augmented generation pipelines, including accelerating vector embeddings and index generation with NVIDIA GPUs. NVIDIA NIM microservices boost generative AI performance for text generation and translation, showcasing how NVIDIA GPUs can enhance Oracle Database capabilities.

NVIDIA and Oracle are collaborating to deliver sovereign AI infrastructure worldwide, addressing data residency needs. Wide Labs trained and deployed Amazonia IA using NVIDIA H100 in OCI’s Brazilian data centers, emphasizing data sovereignty for clients. Nomura Research Institute in Japan is enhancing its financial AI platform with LLMs on OCI’s Alloy infrastructure. Zoom and RSS-Hydro are utilizing NVIDIA GPUs on OCI to support compliance and drive domestic AI applications.

Enterprises can accelerate task automation on OCI with NVIDIA software like NIM microservices and cuOpt, enabling agentic workflows for complex tasks. NVIDIA AI Enterprise software platform, including cuOpt, NIM, and RAPIDS, is available on Oracle Cloud Marketplace. Learn more about the collaboration at Oracle CloudWorld 2024 and discover how AI and accelerated data processing are transforming organizations worldwide.



Read more at NVIDIA: NVIDIA and Oracle to Accelerate AI, Data Processing for Enterprises