NVIDIA and Oracle Collaborate to Boost Enterprise AI with Zettascale10 Cluster

Rongchai Wang
Oct 14, 2025 15:50

NVIDIA and Oracle announce the OCI Zettascale10 cluster, promising enhanced AI capabilities for enterprises with NVIDIA’s GPUs and Spectrum-X Ethernet platform.

NVIDIA and Oracle have joined forces to launch the OCI Zettascale10 computing cluster, a cutting-edge solution aimed at enhancing enterprise AI capabilities. Announced at Oracle AI World, the cluster is designed to handle high-performance AI inference and training workloads, according to NVIDIA’s blog.

Revolutionizing AI Infrastructure

The Zettascale10 cluster, powered by NVIDIA’s GPUs, promises to deliver up to 16 zettaflops of peak AI compute performance. This advancement is set to revolutionize enterprise AI by integrating NVIDIA Spectrum-X Ethernet, the first Ethernet platform specifically built for AI, enabling the interconnection of millions of GPUs efficiently and at scale.

Oracle’s Executive Vice President of Cloud Infrastructure, Mahesh Thiagarajan, emphasized the significance of partnerships in the AI market, stating, “OCI Zettascale10 delivers multi-gigawatt capacity for the most challenging AI workloads with NVIDIA’s next-generation GPU platform.” This collaboration aims to provide customers with robust AI toolsets alongside OCI’s 200+ cloud services.

Enhancements in Database and AI Platforms

Oracle’s Database 26ai is set to incorporate NVIDIA NIM microservices, facilitating the acceleration of high-volume AI vector workloads. This integration allows developers to utilize NVIDIA NeMo Retriever for efficient data processing, enhancing the overall accuracy of AI responses.

The new Oracle AI Data Platform, featuring NVIDIA’s accelerated computing, offers a comprehensive ecosystem for uniting enterprise data with AI models. It includes the NVIDIA RAPIDS Accelerator for Apache Spark, which utilizes GPUs to enhance analytics and machine learning pipelines, significantly speeding up data processing tasks.

Empowering AI Applications

NVIDIA’s NeMo Curator library, combined with Oracle’s Nemotron vision language model, will power Oracle Media and Entertainment’s video understanding capabilities. This integration automates video-centric AI workflows, improving efficiency and scalability in video processing tasks.

Furthermore, Oracle AI Hub now provides a centralized platform for enterprises to build, deploy, and manage custom AI solutions. With NVIDIA NIM microservices, users can deploy models swiftly, enabling a seamless, no-code experience for various business applications.

NVIDIA AI Enterprise on OCI

Enterprises can now access NVIDIA AI Enterprise directly within the OCI Console, streamlining the development and deployment of AI solutions. This integration offers a full suite of AI tools with flexible pricing, enterprise support, and security updates, enhancing the ease of building AI applications at scale.

This collaboration marks a significant step in transforming the enterprise AI landscape, with NVIDIA being recognized as a 2025 Oracle Partner award winner, highlighting its contributions to innovation within the OCI ecosystem.

Image source: Shutterstock