Oracle opens AI Database@AWS in the Paris region with an upcoming interconnection between OCI and AWS

Oracle has announced the launch of Oracle AI Database@AWS in the AWS Europe Paris region, along with plans for a native interconnection between Oracle Cloud Infrastructure (OCI) and AWS. These announcements underscore Oracle’s multicloud strategy in the realm of AI databases, complementing a previously established partnership with Google Cloud around Gemini Enterprise that was unveiled on May 5.

The architecture of enterprise databases has traditionally followed a straightforward model: Oracle, with its Exadata and Autonomous Database solutions, stood on one side, while American hyperscalers offered their managed services and analytical tool ecosystems on the other. However, the shift initiated in 2023 and accelerated in 2025 through Database@Azure, Database@Google Cloud, and Database@AWS has blurred these lines, physically integrating Oracle’s infrastructure within the regions of its primary competitors. Recent announcements extend this trend on two fronts—geographic availability for France and native network connectivity between OCI and AWS.

AWS Europe Paris Region Opens for Oracle AI Database

With the Oracle AI Database@AWS offering, customers in the AWS Europe Paris region can now access Oracle Exadata Database Service, Oracle Autonomous AI Database, and Oracle Autonomous AI Lakehouse, all running on a dedicated OCI infrastructure hosted within AWS. French clients can migrate their existing Oracle AI Database and Oracle Exadata installations with minimal changes, preserving the functionality, architecture, and performance levels they currently enjoy in their local environments.

Zero-ETL integrations streamline data unification between Oracle and AWS for advanced analytics, machine learning models, and generative AI services, including Amazon Bedrock. The distribution of this service will occur via AWS Marketplace, certified partners like Accenture and Atos, as well as through Oracle Multicloud Universal Credits. The Paris region adds to a total of fifteen existing regions across Asia, North America, Australia, Frankfurt, Ireland, London, and Zurich, with six additional regions planned, including Melbourne, Milan, Spain, Stockholm, São Paulo, and Northern California.

Upcoming Native OCI–AWS Interconnection in Virginia

The announcement made on May 12 achieves a separate milestone. Oracle plans to establish a private, managed connectivity between Oracle Interconnect and AWS Interconnect—multicloud, joining the existing native interconnections already available with Microsoft Azure and Google Cloud. The initial deployment is expected in the AWS US East (Northern Virginia) region later this year. This connectivity will support comprehensive multicloud deployments and hybrid architecture setups without relying on third-party network operators or additional physical equipment.

The operational stakes are significant given the workloads involved. Oracle claims that twenty-six partner cloud regions are already interconnected via OCI Interconnect, mainly with Microsoft Azure and Google Cloud. The upcoming AWS connection completes the coverage with the three major U.S. hyperscalers, providing mutual Oracle-AWS clients with a native network foundation to synchronize their Oracle AI Database systems with AWS’s analytics and AI services, all without needing manually configured VPN tunnels or third-party private gateways.

A Triangle of Oracle, AWS, and Google Cloud Around Gemini Enterprise and Worldline

The Paris announcement and the AWS interconnection fit within a busy quarter for Oracle. On May 5, the company expanded its partnership with Google Cloud concerning Oracle AI Database@Google Cloud, now available in fifteen regions. The new Oracle AI Database Agent for Gemini Enterprise enables users to query governed Oracle data in natural language, without needing to write SQL queries or move data. OCI GoldenGate facilitates real-time migration of databases to Oracle AI Database@Google Cloud with integration into Google BigQuery, while Oracle Autonomous AI Lakehouse now directly reads BigQuery Iceberg tables without duplication.

Worldline is highlighted as a reference customer in this context. This European payment service provider supports its transaction processing platform through Oracle Exadata Database Service hosted in Oracle AI Database@Google Cloud, ensuring low latency and high throughput necessary for real-time processing on a global scale. The Worldline case exemplifies what this type of multicloud architecture can achieve in critical regulated workloads.

The Role of Databases in AI Architecture

Together, these three announcements point toward a singular intention. Oracle aims to establish a physical presence within each of its competitors, delivering a marketing promise of multicloud freedom, while creating a structural effect of ubiquity. As Oracle expands its deployment regions—fifteen with AWS, fifteen with Google Cloud, and the same with Microsoft Azure, along with six AWS regions and two Google Cloud regions forthcoming—customers retain the technical flexibility to migrate their workloads from one hyperscaler to another, all while remaining within the Oracle ecosystem. Once the databases are standardized across this configuration, moving out becomes a multi-year systems migration project.

The overarching thesis emphasizes the place of databases within AI architecture. The Oracle AI Database Agent for Gemini Enterprise queries governed Oracle data without relocating it, zero-ETL integrations unify data between Oracle and AWS without pipelines, and Oracle Autonomous AI Lakehouse reads BigQuery Iceberg tables without duplication. Oracle positions its AI Database as the non-movable anchor of serious enterprise AI architecture, relegating hyperscaler tools to layers of inference or analysis connected to it. The Paris launch and the expansion of European regions preemptively neutralize concerns about geographical data localization while waiting for European decisions on sovereignty to solidify. Worldline’s role as a reference client with Google Cloud serves the same purpose, demonstrating that a European payment operator subject to DORA can support its transaction processing on a combination of Oracle and an American hyperscaler.

European CIO Perspectives: Sovereignty and Cloud Act Exposure

For French and European IT leaders, Oracle’s multicloud strategy opens technical options without altering the legal equation. The AWS Europe Paris region hosting Oracle AI Database@AWS remains exposed to the extraterritoriality of the Cloud Act, regardless of the physical storage location. The guidelines from ANSSI on SecNumCloud and the Schrems II ruling from the CJEU continue to serve as reference frameworks for evaluating these offers; to date, none of the three Oracle constructs satisfy the immunity criteria against foreign laws required by SecNumCloud.

The focus thus shifts to assessing the specific workloads involved. For Oracle enterprise databases that do not fall under critical operator regulations, Oracle’s multicloud framework simplifies modernization and connectivity to the AI services of hyperscalers while maintaining application compatibility. However, for regulated workloads that need to align with NIS2, DORA, or specific sector regulations, the combination of Oracle and an American hyperscaler remains to be evaluated against sovereignty requirements and hybrid models under European licensing, whether purely digital or operated in France, such as Bleu or S3ns. The transition between these two approaches is evaluated workload by workload, with no one-size-fits-all answer.

Scroll to Top