Oracle recently announced the launch of OCI Zettascale10, a cutting-edge AI supercomputing infrastructure that is hailed as the largest of its kind in the cloud. This new generation of AI supercomputers connects hundreds of thousands of NVIDIA GPUs across multiple Oracle data centers, creating powerful superclusters capable of delivering up to 16 zettaFLOPS of peak performance.
The OCI Zettascale10 system forms the backbone of Stargate, a flagship supercluster developed in collaboration with OpenAI in Abilene, Texas. This collaboration marks a significant advancement in cloud-based AI performance, combining Oracle’s Acceleron RoCE networking architecture with NVIDIA’s AI infrastructure to set new benchmarks in scale, energy efficiency, and reliability in distributed AI computing.
With Zettascale10, Oracle is redefining enterprise-scale AI capabilities by offering unmatched AI capacity at an unprecedented scale. The system, which builds upon Oracle’s initial Zettascale cluster launched in 2024, is designed for extreme density within a two-kilometer radius to minimize GPU-to-GPU latency. This innovative architecture, showcased at the Abilene Stargate site, serves as a real-world testbed for next-generation AI infrastructure and sets the stage for further advancements in the field.
