General availability of the new Amazon OpenSearch Service engine gives enterprises a log-analytics option AWS says can cut storage costs by 70% while improving price-performance.
AI workloads are driving the need: Dynatrace found a 93% jump in log volume over the past year, pushing organizations to exclude an average 86% of log data to control cost and capacity.
Apache Parquet storage, Lucene indexes, and query routing through Apache Calcite and DataFusion let the engine run search and analytical aggregation in the same query while keeping the same console, APIs, security model and networking setup.
Migration could slow uptake because customers must create a new domain rather than enable the engine inside an existing one, and teams using OpenSearch DSL may need to rewrite dashboards, alerts and automation.
Analysts said cheaper retention could extend compliance and incident-response visibility and reduce observability tool sprawl, but adoption will hinge more on migration friction than on the technology itself.
With enterprises discarding 86% of log data, can AWS's new engine truly end the dilemma between crippling costs and critical visibility?
As AI data costs explode, does this specialized AWS engine signal the end of one-size-fits-all data systems?
Beyond the 70% cost saving, what is the hidden price of the mandatory migration and workflow overhaul for AI logs?
Amazon OpenSearch Serverless NextGen: 20x Faster Auto-Scaling and Scale-to-Zero for AI Workloads
Overview
In June 2026, Amazon launched OpenSearch Serverless NextGen across all commercial AWS regions, marking a major architectural evolution in response to the rise of agentic AI applications. These new AI workloads are bursty and unpredictable, making traditional data infrastructure inefficient due to over- or under-provisioning. AWS rebuilt the service to deliver true serverless capabilities, allowing dynamic scaling without the overhead of always-on resources. This overhaul ensures that compute can scale rapidly and even down to zero when not in use, directly addressing the needs of modern AI applications and enabling significant cost savings and operational efficiency.