The modern data stack was built for humans asking questions. Google just rebuilt its for agents taking action.

At Cloud Next, Google introduced Agentic Data Cloud, positioning its data platform for AI agents rather than only human-driven analytics.

The stack centers on an automated Knowledge Catalog, a cross-cloud BigQuery lakehouse built on Apache Iceberg, and a Data Agent Kit that replaces custom pipelines with agent tools.

Google is also adding private AWS connectivity and zero-copy federation to support governed access across clouds and better semantic understanding of enterprise data.

The move signals a push to make data platforms agent-ready, so AI systems can discover, interpret, and act on enterprise data across clouds and SaaS apps.