Amazon Launches OpenSearch Serverless With 60% Lower Costs for AI Agents

10 articles · Updated · AWS Blog · May 28

Amazon said the next-generation OpenSearch Serverless is now generally available, giving AI-agent developers a managed search and vector engine that scales from zero to thousands of requests per second.
Up to 60% cost savings versus clusters provisioned for peak capacity come from scale-to-zero pricing, while capacity scales up to 20 times faster than the previous generation and resources are created in seconds.
Vercel integration lets users create or connect OpenSearch collections from its console, and AWS said developers can move from prototype to production in minutes with tools including Kiro, Cursor and Claude Code.
At launch, the service supports full-text and vector search collection types, is available in all AWS commercial regions where OpenSearch Serverless already runs, and charges by OpenSearch Compute Units plus storage.

Can Amazon's serverless AI engine cut costs by 60% without the performance pitfalls of cold starts?

Is the push for serverless AI databases about developer freedom or a new way for cloud giants to dominate the market?