Oracle just dropped a bombshell at Oracle AI World with its new Zettascale10 computing cluster, powered by NVIDIA GPUs and capable of delivering 16 zettaflops of peak AI performance. The announcement signals a massive escalation in the enterprise AI arms race, with Oracle positioning itself to handle the most demanding AI workloads while streamlining database operations and making enterprise AI deployment dramatically easier.
Oracle just threw down the gauntlet in enterprise AI with a supercluster announcement that has the cloud computing world buzzing. At Oracle AI World today, the company unveiled its Zettascale10 computing cluster, a NVIDIA GPU-powered behemoth designed to deliver up to 16 zettaflops of peak AI compute performance.
The timing couldn't be more strategic. As enterprises scramble to deploy AI applications that demand speed, scalability, and bulletproof security, Oracle's betting big on partnerships to win the enterprise AI infrastructure war. "I believe the AI market has been defined by critical partnerships such as the one between Oracle and NVIDIA," Mahesh Thiagarajan, executive vice president of Oracle Cloud Infrastructure, told attendees.
What makes this announcement particularly interesting is the underlying technology. The Zettascale10 cluster harnesses NVIDIA Spectrum-X Ethernet - the first Ethernet platform purpose-built for AI - enabling hyperscalers to interconnect millions of GPUs with unprecedented efficiency. It's a direct shot at competitors who've struggled with GPU interconnectivity at scale.
But Oracle didn't stop at raw compute power. The company's real play here is integration across its entire stack. Oracle Database 26ai, the company's flagship database, now supports NVIDIA NeMo Retriever, making it dramatically easier for developers to run vector embedding models and implement retrieval-augmented generation (RAG) pipelines using NVIDIA NIM microservices.
"Through this latest collaboration, Oracle and NVIDIA are marking new frontiers in cutting-edge accelerated computing," said Ian Buck, vice president of hyperscale and high-performance computing at NVIDIA. The partnership addresses a critical bottleneck that's been plaguing enterprise AI deployments: the gap between database operations and AI inference.
The technical details reveal Oracle's sophisticated approach to enterprise AI challenges. The new Oracle Private AI Services Container service supports future use of NVIDIA GPUs for vector embedding and index generation using the NVIDIA cuVS open-source library. This addresses a real pain point - as data volumes explode and AI applications mature, vector index creation build times have become major bottlenecks.
