Oracle expands its AI lineup with Gemini

Oracle and Google Cloud are working together to prepare Oracle’s main database for AI on Google Cloud, as part of a larger plan for 2025. On October 14, Oracle introduced new AI database services for Oracle Database on Google Cloud, launched the long-term support AI Database “26ai", and made its Autonomous AI Lakehouse accessible to everyone.

This announcement follows two important developments this year: Oracle has made Google’s Gemini models available to its customers, and OpenAI has enhanced its services on Oracle’s cloud. These actions clearly show Oracle’s strategy to make its data platform a leading choice for enterprise AI while helping teams with their current workflows.

How it works

If you’re already using Oracle’s previous AI version, you can easily update to 26ai with the October 2025 update, there’s no need for a complete upgrade or to recertify your app. A key feature of this update is the built-in vector search, which lets you run “find similar” queries and retrieve information from live data. This feature is included at no extra cost.

On the analytics side, the new Autonomous AI Lakehouse uses Apache Iceberg, an open table format. This means your data can stay where it is while you integrate AI and analytics across various cloud platforms. It’s available on Oracle Cloud, AWS, Azure, Google Cloud, and Exadata Cloud@Customer.

Deployment & Details

On the geography front, Oracle and Google Cloud expanded their services to Australia (Melbourne) on October 23, with plans for more regions to follow. This is significant for latency and data residency concerns, especially if your users or regulators have specific requirements about where data is stored.

In terms of commerce, there’s now a formal partner program and marketplace options, allowing enterprises to purchase and support Database@Google Cloud through familiar Oracle and Google channels instead of relying on one-off deals.

From a technology perspective, Database@Google Cloud in Australia operates on the Oracle Exadata Database Service within Oracle Cloud Infrastructure, located inside a Google Cloud data center, essentially creating a “cloud-next-to-cloud” setup. This arrangement keeps Oracle closely integrated with the rest of the Google Cloud stack (like BigQuery and Vertex/Gemini), reducing the need for cross-cloud transfers while maintaining daily operational workflows for teams.

What now / who benefits

Data teams: You can now pilot semantic search and retrieval on live tables using 26ai’s vector features, there’s no need to set up or sync a separate vector store.

Analytics teams: Leverage the Iceberg-based lakehouse to query across clouds without needing to copy data first. This approach helps you avoid the common pitfalls of moving or duplicating data, which often hinder AI pilot projects.

App teams on Google Cloud: As new Database@Google Cloud regions become available, you can keep Oracle systems of record close to Big Query and Vertex-style services, helping to reduce latency and meet residency requirements.

Big picture: With Gemini accessible to Oracle customers and OpenAI expanding its capabilities on Oracle, the combination of multi-cloud and multi-AI solutions is shifting from being an exception to a supported strategy. This context underscores the significance of the recent database developments.