Oracle repositions lakehouse for AI agents with shared semantic layer
Oracle claims its new lakehouse approach supports real-time AI agent queries across multiple data sources without copying data.
Key points
- Oracle pitches its Autonomous AI Lakehouse as an open layer that sits across existing data estates rather than replacing them.
- The system uses a catalog that can mount other catalogs, including Databricks Unity Catalog, Postgres and MySQL databases.
- Oracle’s Lake Cache speeds up queries by moving the most-used Iceberg tables into Exadata’s flash cache.
Oracle is shifting its data strategy to support artificial intelligence agents as primary data consumers, according to a report by SiliconANGLE AI.reported The company describes its Autonomous AI Lakehouse as an open layer that integrates with existing enterprise data rather than replacing current systems.Oracle is updating its data tools so AI programs can access information instantly from many different sources.Oracle’s Autonomous AI Lakehouse introduces a shared semantic layer that mounts external catalogs to support low-latency AI agent queries.
Evidence: reported
Oracle rethinks the lakehouse as AI agents become data consumersaifleet-siliconangle
What was released or claimed
Nipun Agarwal, senior vice president of software engineering at Oracle, stated that lakehouses are evolving from analytics platforms into the decision backbone for AI. He made these remarks during an exclusive broadcast on theCUBE at Oracle AI World. Agarwal emphasized that the new approach combines a shared semantic layer with fast, real-time access to data across various platforms.
Who it affects
The changes primarily impact enterprises that use multiple data sources and lakehouses. Organizations relying on human analysts for reporting are now facing environments where AI agents query corporate data alongside humans. These agents require platforms that can interpret information scattered across different clouds, formats and rival systems.
What is still unknown
It remains unclear how many customers have already migrated to this new model or what specific performance metrics Oracle has validated. The material does not specify if there are licensing costs associated with mounting external catalogs like Databricks Unity Catalog within the Oracle layer.
The takeaway
Oracle views its offering as complementary to other lakehouses rather than competitive. The company believes enterprises will maintain multiple data sources and lakehouses simultaneously. The shared semantic layer allows data to be queried in place, reducing the need to copy information between systems.
What changes for users
Enterprises can query data in place across Oracle, Databricks, Postgres and MySQL without moving it, according to the source.reported Real-time access becomes essential because AI agents demand on-demand data rather than relying on scheduled refreshes of copied datasets. Oracle announced that table and column descriptions the system generates apply to non-Oracle sources as well.
Evidence: reported
Oracle rethinks the lakehouse as AI agents become data consumersaifleet-siliconangle
Why it matters
- Teams can reduce data duplication by querying existing catalogs directly through Oracle’s mounting capability.
- Real-time access requirements shift infrastructure needs from batch processing to low-latency caching solutions.
- Support for Apache Iceberg and Unity Catalog integration allows broader compatibility with existing data estates.
- The semantic layer standardizes descriptions across disparate data sources, simplifying agent interactions.
Topics: Oracle, data lakehouse, AI agents, semantic layer, Apache Iceberg
Updated
Then / now / next
Lakehouses were primarily used by customers for running traditional analytics and reporting tasks.
Oracle claims lakehouses have become the decision backbone for AI agents requiring real-time data.
Watch for further details on how the shared semantic layer handles complex queries across rival systems.
Story timeline
- Then Lakehouses were primarily used by customers for running traditional analytics and reporting tasks.
- 04 Nov 2026 Oracle rethinks the lakehouse as AI agents become data consumers
- Now Oracle claims lakehouses have become the decision backbone for AI agents requiring real-time data.
- Next Watch for further details on how the shared semantic layer handles complex queries across rival systems.
Ask the article
Local keyword search of this article; no generated answers.Did this article give you the evidence you needed?
Your choice stays on this device. No public tally.Questions readers ask
Does Oracle’s new lakehouse replace existing data platforms?
No, Oracle positions it as an open layer that sits across existing data estates rather than replacing them.
Which external catalogs can be mounted?
The system can mount Databricks Unity Catalog, Postgres and MySQL databases.
How does Oracle speed up queries for AI agents?
Oracle’s Lake Cache moves the most-used Iceberg tables into Exadata’s flash cache for faster access.
Connected stories
Reading receipt
- Oracle pitches its Autonomous AI Lakehouse as an open layer that sits across existing data estates rather than replacing them.
- The system uses a catalog that can mount other catalogs, including Databricks Unity Catalog, Postgres and MySQL databases.
- Oracle’s Lake Cache speeds up queries by moving the most-used Iceberg tables into Exadata’s flash cache.
Still open: How does the system handle conflicts when the same data exists in multiple mounted catalogs?
Still open: What are the specific latency improvements provided by the Lake Cache feature?
