What running DynamoDB and Keyspaces at scale takes in 2026
Watch on-demand: Alex DeBrie and AWS on what's changed for these databases, from DynamoDB's brand-new vector search to the problems every team runs into once the tables start piling up.
Our speakers
What's new and what's next for DynamoDB and Keyspaces
If you run DynamoDB or Keyspaces at scale, the ground just shifted under you. AI agents are changing how you build on these databases, DynamoDB just shipped Vector Search, and even how you model your data is worth a fresh look. Meanwhile the old questions haven't gone anywhere: do you actually know what's in these thousand tables, and can you get one record back without restoring the whole thing?
That's a lot to track, so we brought in the people who'd know. Alex DeBrie, author of The DynamoDB Book, is building one of the first live demos of DynamoDB vector search anywhere, live attendees only. He's joined by AWS's Vikas Panghal, who leads Keyspaces and part of DynamoDB, and Eon's Ohad Maislish.
What you'll walk away with:
- DynamoDB or Keyspaces: which one, and why the choice sticks
- Where vector search fits, and what AI agents actually make easier
- How AI is reshaping the way you model and build on these databases
- What an AI agent can surface about what's really in your tables
- How to recover a single record or table at scale, past the 35-day PITR window and across accounts and regions