![GPUs]()
Oracle has launched its first wave of liquid-cooled NVIDIA GB200 NVL72 racks in its data centers, now available for customer use. These advanced systems are designed to power NVIDIA DGX Cloud and Oracle Cloud Infrastructure (OCI) for developing next-generation AI reasoning models and agents.
High-Speed Networking for Scalable Performance
The GB200 NVL72 racks are equipped with high-speed NVIDIA Quantum-2 InfiniBand and Spectrum-X Ethernet networking, enabling scalable, low-latency performance. This setup ensures efficient AI reasoning model development and accelerated computing tasks.
OCI’s Ambitious Plans for AI Growth
OCI is among the first to deploy NVIDIA GB200 NVL72 systems and plans to build one of the world’s largest Blackwell clusters, scaling beyond 100,000 NVIDIA Blackwell GPUs. This expansion aims to meet the rapidly growing demand for inference tokens and accelerated AI computing.
Transforming Data Centers into AI Factories
OCI’s installation of NVIDIA Grace Blackwell systems is a key part of transforming its data centers into AI factories. These new AI factories leverage the GB200 NVL72 platform, combining NVIDIA Grace CPUs and Blackwell GPUs to provide exceptional performance and energy efficiency for agentic AI.
Flexible Deployment Options for Diverse Customers
OCI offers flexible deployment options for customers, providing Blackwell-powered systems across public, government, and sovereign clouds, as well as customer-owned data centers. This ensures scalability and accessibility for a wide range of enterprise and government organizations.
![Oracle]()
Use Cases for the GB200 Systems
A variety of customers, including major tech companies, government agencies, and cloud providers, are already planning to deploy workloads on OCI’s GB200 systems. Use cases include AI model training, autonomous vehicle development, chip design acceleration, and more.
Now Live on DGX Cloud and OCI
The GB200 NVL72 racks are now live and available through NVIDIA DGX Cloud and OCI, providing organizations with cutting-edge hardware and support for AI-driven projects.