SageMaker Data Agent adds Snowflake SQL analytics and interactive charting

AWS expanded its SageMaker Data Agent with three notebook capabilities in Unified Studio: running SQL analytics directly against Snowflake sources, managing materialized views, and building interactive charts from natural-language prompts. The headline integration is querying Snowflake alongside native AWS data in one agentic workflow, which AWS demonstrated through a cross-platform fraud-detection scenario spanning both data estates.

The Snowflake bridge is strategically interesting because Snowflake and AWS both compete and partner heavily; letting a SageMaker agent query Snowflake natively reduces friction for the many enterprises whose data is split across the two. Practitioners can schedule aggregations via materialized views and turn prompts into visualizations without hand-writing SQL or plotting code — a step toward conversational analytics on top of warehouse data.

The release is one thread in AWS's dense Summit-week output (Lambda MicroVMs, AgentCore Payments, multi-tenant RAG, the AgentCore harness). Collectively they push the message that AWS wants to own the enterprise agent and data-agent stack end to end. The caveat with natural-language-to-SQL agents remains correctness and governance: generated queries against production warehouses need guardrails, and AWS's pairing of agents with Verified Permissions for multi-tenant RAG hints it knows authorization is the hard part. Watch adoption among joint Snowflake/AWS customers.