Oracle announces Base Database Cloud@Customer, a hybrid cloud service for medium-sized, high-performance workloads. The solution runs on the new Data Infrastructure Cloud@Customer X11 and allows organizations to run Oracle AI Database, applications, and AI agents on-premises, with infrastructure fully managed by Oracle.
More and more organizations are moving Oracle databases to the cloud, but not everyone can or wants to migrate their data there. Consider companies that want to stay close to existing systems due to data residency requirements, regulations, or low-latency needs. For this group, Oracle is now introducing an intermediate solution.
Base Database Cloud@Customer is based on the Base Database Service, which is already available on OCI and in multicloud environments. It runs on the Cloud@Customer model, where Oracle installs the hardware in the customer’s data center and manages it remotely. Unlike the more powerful Exadata Cloud@Customer, this new service is aimed at smaller organizations, departments, and remote sites that do not require that level of scale.
The infrastructure under the hood
The foundation is an 8U rack-mountable Engineered System with two Oracle X11 Compute servers and a shared storage shelf. Each server has 60 usable x86 cores and 660 GB of memory. The all-flash storage starts at 11.6 TB and can be expanded to 47.2 TB. Network traffic runs over 10/25 GbE.
Databases can automatically scale within a single server or be distributed across both servers for high availability. Customers run both Oracle AI Database 26ai and the older 19c. The infrastructure is also prepared for Real Application Clusters (RAC), with automated configuration.
Private AI behind your own firewall
In addition to the database, customers can also host application VMs on the same infrastructure. Of particular interest is the combination with Oracle Private AI Services Container and the AI Database Private Agent Factory. This allows organizations to run agentic AI workloads entirely on-premises, behind their own firewall, and isolated from external AI providers.
This AI focus is part of a broader strategy. Oracle positions 26ai as an AI-native database that combines AI Vector Search, transaction processing, and analytics into a single engine. Oracle is building a converged database for OLTP, analytics, and vector searches.
The pricing model is pay-per-use. Customers scale compute resources as needed and do not pay for licenses on unused cores during periods of low activity. Oracle Cloud Operations manages and monitors the underlying infrastructure.