Oracle AI World Oct. 28: Loaiza says guardrails alone won't cut it

Oracle's messaging ahead of its Las Vegas event centers on a shift in how database modernization is framed. Traditionally justified by efficiency and cost savings, modernization is now being pitched by Oracle executives as an AI readiness decision that impacts all parts of the business. Dave Vellante of theCUBE Research describes the strategic question for Oracle customers as whether they are simply reducing infrastructure cost or creating a better platform for repeatable AI execution.

The product areas Oracle plans to showcase reflect that framing. Juan Loaiza argues that dropping AI on top of an architecture never designed for it will not work, and that Oracle is architecting AI, data and AppDev together to ensure trust across the stack. Arpan Shah of Oracle Database said a key focus in Las Vegas is showing how features are designed from the data layer to the app layer to help customers move from AI prototypes to trusted AI apps at scale. Oracle earlier this year outlined a three-pronged strategy addressing database security, patching, testing and lifecycle management.

On the data architecture side, Oracle's Autonomous AI Lakehouse combines Autonomous Database capabilities with the Apache Iceberg open table format. Nipun Agarwal says the lakehouse is transforming from an analytic platform into a decision backbone for enterprise AI, serving both analytic and AI workloads. Oracle's platform capabilities were also extended this year with APEXlang, a human-readable application definition language that Mike Hichwa says lets coding agents such as Codex or Claude generate applications. Separately, Oracle announced Exadata VM Live Migration in June, which Jia Shi says allows a live Oracle Database to move between hosts with RDMA connections intact, a capability Shi describes as previously impossible.