Oracle to offer Google’s Gemini models through OCI, widening enterprise AI choice

Austin, Texas and Sunnyvale, California, Aug 14, 2025 — Oracle and Google Cloud said they are expanding their partnership to make Google’s Gemini models available to Oracle customers through the OCI Generative AI service, starting with Gemini 2.5.

The move gives Oracle clients access to multimodal and coding-focused capabilities for tasks such as content understanding, software development, workflow automation, and retrieval-based research, according to the companies.

Oracle plans to expose the broader Gemini portfolio through new integrations with Vertex AI. The roadmap includes models for video, image, speech, and music generation, as well as industry offerings like MedLM for healthcare.

The companies also said Gemini models delivered via Vertex AI are expected to be offered as an option inside Oracle Fusion Cloud Applications at a later date, which would extend AI features across finance, HR, supply chain, sales, service, and marketing. Oracle customers can fund usage with existing Oracle Universal Credits.

“Leading enterprises are using Gemini to power AI agents across a range of use cases,” said Thomas Kurian, CEO of Google Cloud. He added that access from within Oracle environments is intended to lower friction for deploying agents that support developers and data integration work.

Clay Magouyrk, president of Oracle Cloud Infrastructure, said the addition of Gemini reflects Oracle’s model-choice strategy for enterprises and its focus on secure, cost-effective AI.

Google positions Gemini for enterprise use with support for large context windows, response grounding with current Google Search data, and established encryption and data-handling policies. Oracle highlights proximity to enterprise data on OCI, along with security, adaptability, and scale.

The company pointed to its bare-metal GPU instances and broader AI infrastructure for running generative models, natural language applications, computer vision, and recommendation systems.

The announcement underscores growing demand for agentic AI, where systems can plan and act across tools and data sources.

For Oracle customers, the near-term impact will depend on model availability inside OCI regions, performance and latency under real workloads, and pricing under the Universal Credits model. The integration path into Fusion Cloud Applications will be a key marker for how quickly line-of-business teams can adopt these capabilities without building custom stacks.

Items to watch include timelines for additional Gemini variants, any region-specific support, and the scope of administrative controls for governance and privacy. Enterprises evaluating the service will likely assess citation behavior in grounded outputs, rate limits, and monitoring features that support compliance.