AWS has launched a new version of OpenSearch Serverless, designed to efficiently manage the demands of AI agents transitioning from pilot to production stages. This update separates compute resources from storage, allowing for rapid scaling and cost efficiency by reducing idle compute power. Previously, AWS's serverless architecture required at least one running instance due to tightly coupled compute and storage resources. The shift comes as bot traffic, including AI crawlers and intelligent assistants, now constitutes 31% of all HTTP traffic, according to Cloudflare. AWS's move aligns with industry trends, as companies like Google, Databricks, and Microsoft also adapt their systems to accommodate increasing machine-generated traffic. The new OpenSearch Serverless version integrates with platforms like Vercel and Kiro, facilitating the deployment of search and vector database backends for AI applications.