Artificial intelligence is changing how businesses create and run smart applications. As more companies want faster and more secure AI tools, they are turning to new technology that makes handling data easier and brings intelligence to every part of their operations.
At Oracle AI World, Oracle introduced the new OCI Zettascale10 computing cluster. This powerful computer uses NVIDIA GPUs to handle even the most demanding AI workloads and can reach up to 16 zettaflops of AI computing power. The cluster also connects many GPUs efficiently using NVIDIA Spectrum X Ethernet, which was specially designed for AI projects.
Oracle also announced more integration with NVIDIA technology. Now, Oracle databases include support for NVIDIA NIM microservices, which help developers build smarter AI features. The NVIDIA AI Enterprise software is now directly available in the Oracle Cloud Infrastructure (OCI) console, making it easier for businesses to access top AI tools alongside over 200 cloud services Oracle offers.
As Oracle and NVIDIA work closer together, they are helping companies use AI for many different tasks. These include speeding up database processes, making data processing faster, and supporting different types of AI projects. The new Oracle Database 26ai now supports integration with NVIDIA NeMo Retriever, so developers can use advanced AI models easily for tasks like searching through large data sets or supporting chatbots that give accurate information.
NVIDIA provides tools like NeMo Retriever for collecting and transforming data and large language models that help generate accurate responses. Oracle’s new Private AI Services Container makes it easier to run these kinds of AI services whether your data is in the cloud or in your own data center. Soon, this service will also let users speed up certain data-heavy tasks using NVIDIA GPUs, making work like creating search indexes much quicker.
NVIDIA’s accelerated computing is now part of the Oracle AI Data Platform, which brings together all a company’s data, AI models, and management tools. With built-in NVIDIA GPU options, businesses can run intensive AI workloads and use NVIDIA RAPIDS Accelerator for Apache Spark to make analytics and machine learning tasks faster. Developers can use GPU acceleration for their Apache Spark jobs without needing to change their code.
Oracle is also using NVIDIA’s NeMo Curator library and Nemotron models to make video and document understanding much easier. This process helps organize, label, and prepare video and text data so companies can train better AI models. NVIDIA’s NeMo Retriever Parse model even helps extract valuable information from complex documents, which is useful for building smarter applications.
The Oracle AI Hub gives businesses one place to build, deploy, and manage their AI tools. Through this hub, companies can use NVIDIA NIM microservices with a simple setup and deploy various AI models without needing to write any code. This makes it easy for organizations to try out the newest AI features in their own products.
Finally, NVIDIA AI Enterprise is now fully available within Oracle Cloud Infrastructure. This means businesses can use the complete suite of AI tools from NVIDIA directly through Oracle’s cloud, with strong support and security features. This integration makes building and running AI solutions at large scale much simpler and more cost-effective.
With all these advances and a close partnership between the two companies, Oracle and NVIDIA are making enterprise AI easier to use, more powerful, and accessible for more organizations.
Original article and image: https://blogs.nvidia.com/blog/nvidia-oracle-accelerate-enterprise-ai-data-processing/
