Oracle today announced the launch of OCI Zettascale10, a new cloud-based supercomputer designed for large-scale artificial intelligence workloads. The system is engineered to connect hundreds of thousands of NVIDIA GPUs across multiple data centers, delivering up to 16 zettaFLOPS of peak computing performance. This architecture aims to provide efficient training and deployment for the largest AI models. The announcement positions Oracle to compete more intensely in the demanding AI infrastructure space. Mahesh Thiagarajan, Oracle's Executive Vice President of Cloud Infrastructure, highlighted the system's efficiency, stating, “Customers can build, train, and deploy their largest AI models into production using less power per unit of performance and achieving high reliability.” The technology is designed to offer consistent performance and stability for complex AI tasks. There is no information available regarding the market's immediate reaction to this announcement.