AWS Is Building Its Databases for AI Agents
Techstrong TV Interviews
•
17m
AI agents now use more enterprise data than ever. Alan Shimel talks with G2 Krishnamoorthy, VP of AWS Database Services, about how AWS databases are changing for this agentic world. They cover the new Duck Labs integration in Aurora PostgreSQL, MCP servers and serverless scale.
The breadth of AWS databases
Krishnamoorthy began his career on SQL Server and Azure data services at Microsoft. Later, he worked on backend infrastructure at Meta before he joined Amazon. Today his team runs a wide range of AWS databases. For example, Aurora and RDS handle relational work, and DynamoDB delivers key-value speed at any scale. Meanwhile, DocumentDB, Neptune and ElastiCache cover documents, graphs and in-memory data. The goal is simple: give builders the right tool for each task.
Duck Labs powers analytics in Aurora PostgreSQL
Open engines like Postgres give customers portability. At the same time, AWS competes on scale, security and price performance. To link operational and analytical data, Aurora PostgreSQL now uses Duck Labs as its analytical query engine. As a result, an app can query Iceberg or Parquet files directly. Customers no longer need to copy that data back into the database.
Making AWS databases agent-ready
Agents now build more apps and more analytics. So AWS launched MCP servers that let agents query data across AWS databases through the Model Context Protocol. In addition, serverless options fit agent workloads, which tend to be spiky and short-lived. Price performance also matters, because every token can drive more database activity.
An AI-ready data foundation for AWS databases
Looking toward re:Invent and 2027, Krishnamoorthy says data is the real differentiator. After all, everyone can use the same frontier models. Therefore, AWS is focused on an open data foundation. It brings together Iceberg, a shared catalog and AWS Context, so agents can understand how enterprise data fits together.
AWS databases keep evolving as agentic AI takes hold. Explore more DevOps and cloud videos and the latest Techstrong TV interviews.
For more information please visit aws.amazon.com
Up Next in Techstrong TV Interviews
-
The DevOps Standard Finally Defines D...
DevOps has gone more than a decade without a formal definition. Alan Shimel talks with Marc Hornbeek, CEO and founder of Engineering DevOps Consulting, and Markus Bause, VP of product at PeopleCert, about The DevOps Standard, the new DevOps Institute book that finally defines DevOps and lays out ...
-
Open Mainframe Project Opens Mainfram...
A Neutral Home for Mainframe Open Source
John Mertic is director of program management at the Linux Foundation. He has spent 11 years with the Open Mainframe Project. He and Alan Shimel once co-hosted the Open Mainframe podcast together.The project began around Linux on IBM Z after IBM's LinuxO...
-
AI Agents Need Deterministic Building...
In this Techstrong TV interview, Mike Vizard talks with Adam Jacob, co-founder and CEO of Swamp Club. They discuss why deterministic AI agents are the key to moving AI from prototype to production. Swamp Club has raised $7 million in seed funding. Its bet is that the next phase of AI adoption wil...