Oracle's AI Vector Search, OCI Compute and Generative AI Service

Oracle has made a series of announcements that fortify its commitment to AI and high-performance computing. First, an update to Database 23c incorporates AI Vector Search capabilities, significantly enhancing data queries. Second, Oracle announced the upcoming availability of Oracle Cloud Infrastructure (OCI) Compute instances featuring next-gen processors from NVIDIA and Ampere Computing. Lastly, Oracle revealed the limited availability of its OCI Generative AI service, which aims to automate business processes while maintaining data security and privacy.

The Database 23c's new AI Vector Search feature allows the database to perform semantic searches across business data, promising improved search accuracy and speed. A set of new features, including vector data types, vector indexes, and vector search SQL operators, lets the database handle unstructured data like documents and images as vectors, allowing for semantic similarity queries.

The update also adds Retrieval Augmented Generation (RAG), a Generative AI technique that can deliver natural language responses to queries without exposing sensitive information. Oracle's aim is to make it easier for developers to include advanced AI functionalities in their applications without requiring specialized machine learning expertise. These enhancements join updates to Oracle Autonomous Database, Oracle APEX, and various other services, adding to Oracle's drive to offer scalable and efficient database solutions for all business sizes.

Oracle is also set to launch new OCI Compute instances, powered by next-generation NVIDIA and Ampere Computing processors. The NVIDIA-powered instances could provide up to a 30x improvement in AI inference performance and a 4x improvement in AI model training, compared to previous versions. Meanwhile, Ampere-powered instances are expected to provide leading price-performance and the industry's highest processor core count. These new instances aim to cater to a range of workloads, from AI to cloud-native applications and video transcoding.

Finally, Oracle announced the limited availability of its OCI Generative AI service, developed in partnership with Cohere. The managed service supports Large Language Models that can be integrated into applications via an API. It's expected to work seamlessly with Oracle's AI Vector Search and will eventually be a part of Oracle's suite of SaaS applications. The service comes with various models tailored for different enterprise needs and runs on Oracle's AI Supercluster architecture, giving businesses control over their costs and performance.