Oracle plans a business semantic layer and AI catalog for Autonomous AI Lakehouse

Oracle’s October 26 Autonomous AI Lakehouse announcement combines expanded Apache Iceberg support with plans to give AI agents a shared view of enterprise data and its business meaning. The distinction is between being able to query a table and knowing which definitions, relationships and metrics should govern the answer.

Two central pieces—AI Data Catalog and Business Semantic Layer—are described as becoming available within the next 12 months. They should not be presented as generally available features today. Oracle also attaches a future‑product disclaimer: the release outlines product direction, is not a delivery commitment or contractual term, and should not be relied on for purchasing decisions. Development, release, timing and pricing may change at Oracle’s discretion.

The [planned AI Data Catalog](https://www.oracle.com/news/announcement/ai‑world‑oracle‑advances‑autonomous‑ai‑lakehouse‑with‑business‑semantic‑layer‑and‑expanded‑support‑for‑iceberg‑data‑2026‑10‑26/) would let customers mount external catalogs, Oracle and non‑Oracle databases, Iceberg tables, applications and object‑store data without duplicating it. Oracle describes live querying through its engine and metadata enrichment covering business semantics, lineage, relationships and retrieval signals. The Business Semantic Layer is described as using catalog annotations, imported Oracle AI Database SQL annotations and AI‑suggested analytic views. Those reusable views capture dimensions, measures, metrics, hierarchies, calculations and KPIs. This supplies a proposed place to express business definitions; the announcement does not independently establish the accuracy of an agent’s resulting interpretation. Oracle also describes role‑based, object‑level controls through a single catalog control plane. That is the vendor’s proposed governance model, not verification that a specific deployment enforces every upstream permission.

[Expanded Iceberg support](https://www.oracle.com/news/announcement/ai‑world‑oracle‑advances‑autonomous‑ai‑lakehouse‑with‑business‑semantic‑layer‑and‑expanded‑support‑for‑iceberg‑data‑2026‑10‑26/) is described as letting teams create, populate, update and delete Iceberg tables using SQL, along with reusable data products and time‑travel functionality. The release does not supply a complete feature‑by‑feature availability matrix, so these announced capabilities should not inherit a blanket launch‑status claim. Oracle also describes Data Studio as a workspace for discovery, preparation, transformation pipelines and analysis, with natural‑language, SQL and Python interfaces. A Business Simulation Engine integrates Essbase for multidimensional what‑if work, including write‑back, hierarchy rollups, calculations and allocations. Those are interface and workflow descriptions, rather than evidence of improved forecast quality.

The [release](https://www.oracle.com/news/announcement/ai‑world‑oracle‑advances‑autonomous‑ai‑lakehouse‑with‑business‑semantic‑layer‑and‑expanded‑support‑for‑iceberg‑data‑2026‑10‑26/) says Autonomous AI Lakehouse is available across OCI, AWS, Microsoft Azure, Google Cloud, Exadata Cloud@Customer and OCI Dedicated Region. That existing‑platform availability does not establish that every newly described catalog, semantic or governance feature is enabled in every environment. Individual rollout status, prerequisites and pricing remain separate checks. The emphasis also differs from Oracle’s [Deep Trust AI architecture](https://opentools.ai/news/oracle‑deep‑trust‑ai‑architecture‑database‑object‑api‑apex‑controls), which describes where database, access and application‑runtime controls should operate. This lakehouse announcement focuses on the data agents query, the definitions they use and the related preparation and simulation workflows. A team evaluating the roadmap needs both the relevant data definitions and the controls that govern access to them. Oracle headquarters signage photographed on January 18, 2005. Archive company context, not an image of the Autonomous AI Lakehouse interface or deployment. Photo: [Peter Kaminski](https://www.flickr.com/people/peterkaminski/), [Flickr](https://www.flickr.com/photos/35034359460@N01/3772015/) / [Wikimedia Commons](https://commons.wikimedia.org/wiki/File:Oracle_headquarters.jpg), [CC BY 2.0](https://creativecommons.org/licenses/by/2.0/deed.en). Resized and cropped for display.