Microsoft makes OneLake catalog the core of its data control
Microsoft tells that agents must have data in one system and the context of that data: what it means, who can see it and if it is correct.
At Ignite, Microsoft made the OneLake catalog the main part of its data governance strategy. OneLake is one system where a company unifies its data in one copy, with no move or copy of the data. The catalog is where it finds, controls and protects that data. Microsoft tells that new functions put data, context and governance together for AI agents.
How it works
OneLake links data from many clouds and on site systems into one view. Its catalog controls data in the open Delta Lake and Apache Iceberg formats. Microsoft tells that it is one of the only lakes that does this for the 2.
The numbers
These are use numbers from Microsoft:
- More than 240,000 organizations use the catalog.
- It controls more than 350 million data items.
- It is in more than 100 apps, such as Teams and Excel.
What is not known
Microsoft gave no test results of how this helps agents. It is not clear which new functions are available at this time, or at what price.
Insight: Each large vendor wants to be the catalog that agents read first
Microsoft tells that agents must have one copy of data. Confluent told on 4 November 2026 that agents can use data that is not current. In October, AWS showed AWS Context, where an agent sees only the data that its user sees. This reads as a fight to own the catalog, because it selects which data an agent can use. Look for catalogs from different clouds that work together.
Sources
- OneLake: The unified foundation for data, context, governance, and AIcommunity.fabric.microsoft.com
Posted