re:Invent 2026 News & Analysis

Announcements with the way forward for CXOs and leadership.

AWS re:Invent 2026 runs November 30 to December 4 in Las Vegas. Daily analysis from the NuVista AI team starts December 1.

What to watch for in data and agents

The re:Invent 2026 session catalog points to four areas that matter for data and AI teams planning 2027.

Wednesday, December 2

The Data & AI keynote: getting agents into production

The keynote focuses on three things teams change to move agents out of the demo: how agents reach their data, how teams measure what agents are worth, and the limits teams set.

What it means: Teams planning agent pilots for 2027 should settle data access, measurement and limits before choosing a model.

ANT308

Graviton‑powered RG instances for Amazon Redshift

The session covers new RG instances, a built‑in data lake engine that reads and writes Apache Iceberg tables in place, and multi‑warehouse architecture.

What it means: Teams planning a warehouse migration should check whether RG instances and in‑place Iceberg change their target architecture.

ANT316

Agents that build data pipelines from a single prompt

Agents use the Model Context Protocol (MCP) to build pipelines across AWS Glue, Amazon Redshift and Amazon MWAA Serverless.

What it means: Data engineering leads should decide now how agent‑built pipelines get reviewed before they run in production.

ANT441

Text‑to‑SQL with guardrails

The Amazon Redshift MCP Server lets agents write SQL from plain English. The session focuses on stopping query hallucination, prompt injection and PII exposure.

What it means: Before text‑to‑SQL reaches business users, build guardrails for hallucination, injection and PII into the design.

Sessions worth booking for data teams

Five sessions our team recommends for data and analytics leaders.

ANT308

What’s new in Amazon Redshift

Graviton‑powered RG instances, native Apache Iceberg read and write, and multi‑warehouse architecture.

ANT301

Graviton. Serverless. Apache Iceberg. Your guide to a modern data warehouse strategy

A clear architecture plan for a faster, open warehouse.

ANT316

From prompt to pipeline: Agentic data engineering

Agents building pipelines across AWS Glue, Amazon Redshift and Amazon MWAA Serverless from one prompt.

ANT314

Accelerating data analytics on Apache Iceberg with Amazon S3 Tables

Faster analytics on open table formats.

ANT441

Build a well‑governed natural language‑to‑SQL agentic solution for Amazon Redshift

Security gates for text‑to‑SQL agents in production.

Reserve seats in the official re:Invent 2026 session catalog.

Keynote schedule

All keynotes take place at The Venetian and stream live. Our analysis of each one is published here the same day.

Monday, November 307:30 to 8:45 PM PT

Developer Keynote

Speaker to be announced by AWS

Analysis on Dec 1
Tuesday, December 18:30 to 10:30 AM PT

CEO Keynote

Matt Garman, CEO, AWS

Analysis on Dec 1
Wednesday, December 28:30 to 10:00 AM PT

Data & AI Keynote

Speaker to be announced by AWS

Analysis on Dec 2
Wednesday, December 23:00 to 4:00 PM PT

Partner Keynote

Ruba Borno, Vice President, AWS Specialists & Partners

Analysis on Dec 2
Thursday, December 38:30 to 10:00 AM PT

Technology Innovations Keynote

Speaker to be announced by AWS

Analysis on Dec 3

Times from the AWS re:Invent 2026 agenda. Watch live on the AWS keynotes page.

Talk to our team at re:Invent 2026

Bimal Tandel, John Aguilar and Vinay Tandel are on the expo floor all week.