Accelerate AI Anywhere with Oracle Cloud and NVIDIA
Build, train, and deploy AI with Oracle Cloud Infrastructure (OCI) and NVIDIA. From large-scale GPU clusters to NVIDIA Vera CPUs for agentic workloads and accelerated enterprise software, Oracle and NVIDIA help you bring AI into production with infrastructure and services built for performance, scale, and flexibility.
Oracle and NVIDIA expand AI infrastructure and enterprise AI capabilities.
- Explore NVIDIA
Vera Rubin NVL72 on OCI
Discover infrastructure for demanding AI training and inference workloads.
- Power agents with NVIDIA Vera CPUs
Bring dedicated CPU capacity to the services that keep AI agents moving.
- Oracle deploys thousands of NVIDIA Blackwell
GPUs
Oracle has stood up and optimized the first wave of liquid-cooled NVIDIA GB200 NVL72 racks. These racks are live and available now from NVIDIA DGX Cloud and OCI.
Power the work behind every AI agent
AI agents depend on more than model inference. Tool calls, code execution, retrieval, and orchestration all rely on CPUs. NVIDIA Vera CPUs on OCI provide dedicated compute for these workloads, with bare metal and virtual instance options that don’t require Rubin GPUs.
-
Agent orchestration
Coordinate sessions, dispatch tools, and process results.
-
Sandbox execution
Run generated code, tests, simulations, and reward logic.
-
Retrieval and analytics
Support workloads that depend on memory capacity and bandwidth.
AI infrastructure solutions
-
Solution Architecture
Leverage the power of NVIDIA GPUs and advanced deep learning frameworks with ready-to-deploy automation.
Hands-on lab -
Solution Architecture
Learn how Agent Workflow running on OCI helps you configure different AI models securely with ready-to-deploy automation.
Hands-on lab -
Solution Architecture
Run a simple RAG-enabled chatbot in OKE using NVIDIA NIM, Qdrant, Gradio, and ready-to-deploy automation.
Hands-on lab -
Solution Architecture
Harness the capabilities of NVIDIA GPUs and NIMs to gain practical experience in deploying scalable fraud detection solutions with ready-to-deploy automation.
Hands-on lab -
Solution Architecture
Automate the deployment of Morpheus on OCI leveraging NVIDIA RAPIDS, Triton, and other tools.
Hands-on lab -
AI Solution
Deploy NVIDIA NIM on OCI Kubernetes Engine for scalable, efficient inference using OCI Object Storage and NVIDIA GPUs for optimal performance.
Learning Path -
AI Solution
Learn to deploy large language models (LLMs) on OCI, accelerated by NVIDIA GPUs with bare metal machines.
Learning Path -
AI Solution
Learn how to run distributed multinode training with NVIDIA GPUs on OCI for efficient deep neural network training.
Learning Path
NVIDIA AI Enterprise on OCI Marketplace
-
Data preparation
Accelerate data science workflows using NVIDIA RAPIDS Accelerator on Oracle Cloud Infrastructure (OCI) Data Flow.
-
Model training and customization
Build and customize generative AI models using NVIDIA DGX™ Cloud on OCI.
-
Inference (RAG)
Improve the accuracy and relevance of LLM outputs by building retrieval-augmented generation (RAG) pipelines using NVIDIA NeMo Retriever, Oracle Database 23ai, and Oracle Cloud Infrastructure Kubernetes Engine (OKE).
-
Inference
Accelerate and scale model deployment with NVIDIA NIM microservices and OKE.
Bring NVIDIA acceleration to enterprise AI
Extend accelerated computing into inference, database indexing, and business optimization.
NVIDIA Dynamo on OKE
Support disaggregated inference for containerized AI workloads.
NVIDIA cuVS
Accelerate indexing with Oracle Vector Index Service.
NVIDIA cuopt
Apply accelerated optimization with OCI Enterprise AI.
Now generally available: The largest, fastest AI supercomputer in the cloud
We’re excited to announce the general availability of Oracle Cloud Infrastructure Supercluster with NVIDIA H200 Tensor Core GPUs.
Enabling customers to streamline AI innovation
-
Simplified operations for agentic AI
Deliver a full-stack AI platform that powers agentic AI—intelligent systems that autonomously perceive, reason, and act. With NVIDIA AI Enterprise available natively through the OCI Console, enterprises can quickly and easily access more than 160 AI tools for training and inference while getting direct billing and customer support.
For rapid AI inference, NVIDIA NIM inference endpoints in OCI Marketplace offer a scalable, low-complexity solution for deploying AI-powered assistants, copilots, and real-time applications.
-
Industry-leading AI infrastructure
Stay at the forefront of generative AI innovation by achieving up to 260 exaFLOPS of performance with Hopper GPUs and 2.4 zettaFLOPS with Blackwell GPUs. With these OCI Superclusters, you can train trillion-parameter models faster and deploy them at scale.
Take advantage of OCI Compute with bare metal instances and no virtualization overhead, ultrafast RDMA cluster networking, petabyte-scale file storage, and orchestration tools such as OCI Kubernetes Engine to accelerate AI workloads at any scale.
-
Powering AI anywhere
Optimize inferencing and run AI anywhere with NVIDIA technologies on OCI’s distributed cloud. Deploy NVIDIA L4 GPUs on an edge appliance and scale up to the largest supercomputing infrastructure in the public cloud, or expand AI infrastructure in your data center.
Governments and regulated industries can choose flexible deployment models to address strict data sovereignty and compliance needs. NVIDIA AI Enterprise on OCI accelerates the development and deployment of production-ready AI and is available anywhere in OCI’s distributed cloud.
Featured AI partners and customers
Infuse AI into applications using Cohere large language models
Oracle has collaborated with Cohere to power Oracle Cloud Infrastructure’s generative AI services. Leveraging the performance of OCI to train their models, Cohere is working with Oracle to bring enterprise AI technology to businesses.
Modal Labs
Modal Labs lets you run data/AI jobs in the cloud by just writing a few lines of Python. Customers use Modal to deploy GenAI models at large scale, fine-tune LLM models, run protein folding simulations, and much more. Modal Labs uses Oracle's bare metal A10 instances because of the unbeatable combination of price and performance.
Scaling medical knowledge and creating reliable AI for the health benefit of every person
Headquartered in San Francisco, California, Evidium is a health technology startup that has created a referenced AI platform to give healthcare organizations grounded and trustworthy AI. To power its model training, the company leverages GPUs on OCI for its diverse product line.
Fireworks AI
Founded by leaders from PyTorch and Meta, Fireworks AI offers the fastest and highest quality platform to serve generative AI models aimed at accelerating product innovation and disruption. The company selected Oracle Cloud Infrastructure to run inferencing and training workloads.
GenAI-powered search and chat with your own data
Legionintel is a generative AI integration platform on OCI that’s trusted by the world’s most secure organizations. The company offers high-quality attributed outputs, faster time to value, and seamless integration with source-of-truth applications.
Altair tests and runs complex simulations overnight on OCI
Altair is all about transforming enterprise decision-making by leveraging the convergence of simulation, high performance computing, and artificial intelligence. The company flips its cutting-edge design applications into overdrive by developing and running them on high performance Oracle Cloud Infrastructure.
Get started with OCI AI infrastructure accelerated by NVIDIA
Learn more about Oracle and NVIDIA AI services
Oracle and NVIDIA are expanding access to accelerated AI computing in the cloud so organizations can solve their most complex business challenges.
Related information
- Oracle Delivers Sovereign AI Anywhere Using NVIDIA Accelerated Computing
- OCI AI Blueprints: Go from Zero to Hero While Deploying AI Workloads on OCI
- First Principles: Inside Zettascale OCI Superclusters for Next-Gen AI
- Now Generally Available: The Largest, Fastest AI Supercomputer in the Cloud
- Agentic Workflow on OCI Using NVIDIA NIMs
- MLPerf Training Benchmark 4.0 Results on OCI GPU Superclusters
- GPU-Based Assets on OCI AI Solutions Hub
- Accelerating AI Workloads with OCI Supercluster (PDF)
- Latest Oracle and NVIDIA announcements
- General availability of OCI Compute with Vera CPUs
- NVIDIA Vera CPUs on OCI
- OCI Achieves NVIDIA Exemplar Cloud Validation