NVIDIA partners with Oracle to accelerate AI inference

By combining NVIDIA’s accelerated computing and interference software with Oracles’ AI infrastructure and GenAI services, the collaboration aims to help businesses worldwide speed up and optimise the creation of agentic AI applications, which operate autonomously and make decisions based on goals rather than following pre-defined instructions.

Optimising large-scale AI workloads

The new integration between Oracle Cloud Infrastructure (OCI) and the NVIDIA AI Enterprise software platform will make 160+ AI tools available through the OCI Console. As Oracle is a platform that enables AI training and inferencing, the company aims to help customers get better business results and get more value from their data.

NVIDIA AI Enterprise will enable customers to access AI capabilities such as NVIDIA NIM, a set of 100+ cloud-native inference microservices for AI models, including the latest NVIDIA Llama Nemotron models for advanced AI reasoning. This aims to reduce the time it takes to deploy reasoning models. NVIDIA AI Enterprise will be available as a deployment image for OCI bare-metal instances and Kubernetes clusters using OCI Kubernetes Engine. OCI Console customers benefit from direct billing and customer support through Oracle.

Organisations can deploy OCI’s 150+ AI and cloud services with NVIDIA in the data centre, the public cloud, or at the edge. This feature offers an integrated AI stack to help address data privacy, sovereign AI, and low-latency requirements.

Enterprises can also leverage OCI AI Blueprints to get no-code deployment recipes that enable them to run AI workloads without having to make decisions about the software stack or manually provision the infrastructure. The blueprints offer hardware recommendations for NVIDIA GPUs, NIM microservices, and prepackaged observability tools. NVIDIA Blueprints provide developers with a unified experience across the NVIDIA stack, providing reference workflows for enterprise AI use cases, and supporting them to build custom AI applications.