OpenSearch Five Years On: AWS Share Falls From 91% to 72%
Cloud Platform EngineeringIn OpenSearch’s first year, AWS accounted for 91% of contribution activity across the project’s repositories. In 2025–26 it’s 72%. For platform teams that chose OpenSearch to avoid depending on one vendor, that’s progress, but it also means one company still does nearly three-quarters of the work.
The Number
The New Stack reports the figures from the OpenSearch Software Foundation. AWS launched OpenSearch in January 2021 after Elastic moved Elasticsearch and Kibana off Apache 2.0, Elastic said at the time it wanted to stop cloud providers such as AWS offering Elasticsearch as a service without giving anything back, and accused AWS of violating its trademark. AWS moved OpenSearch to the Linux Foundation in September 2024 to put it under vendor-neutral governance. Amazon-affiliated engineers made up 95% of maintainers in the core repository in April 2022 and 71% today. AWS holds four of 15 voting seats on the Technical Steering Committee, but an AWS employee, Andrew Ross, was just reelected chair.
Deployment is shifting faster. In the 2026 Open Data Infrastructure Report from the Linux Foundation and the OpenSearch Software Foundation, 60% of surveyed users run self-managed OpenSearch in the cloud, up from 37% in 2024. Use of Amazon’s managed OpenSearch Service rose from 44% to 48%.
How the Math Works
The foundation’s executive director, Bianca Lewis, calls the donation “a watershed moment” because AWS “couldn’t close OpenSearch and they couldn’t change the license.” Her reading of AWS’s motive: “AWS don’t make money from applications. AWS make money from compute.”
The perception problem is measurable too. Among 71 surveyed non-adopters, 24% cite concerns about vendor neutrality and ties to AWS. And the report shows many deployments are still experimental or replaceable. Lewis points to multiyear Splunk and Elastic contracts that companies won’t abandon mid-term, so they may be testing OpenSearch alongside tools they’ve already paid for.
Elastic, meanwhile, has added AGPL as a licensing option for Elasticsearch and Kibana and made up with AWS, so the original reason for the fork has partly faded.
What It Signals
The product has moved on from its origins. Lewis admits early OpenSearch “was Elasticsearch with a different name”, with the same plugins and much of the same documentation. It now pitches vector and hybrid search for retrieval-augmented generation, and OpenSearch 3.9 in September added retrieval improvements for AI applications. “If you decide that you wanted your AI infrastructure on an open platform, I think [OpenSearch] today is just the default choice,” Lewis says, a claim from the foundation’s own director.
AWS keeps building on it. Stack Archive covered how AWS rebuilt OpenSearch Serverless for agentic workloads, the managed path most AWS customers will meet first.
The So What
Platform teams choosing a search or vector store for RAG should treat OpenSearch’s independence as real in licensing and still concentrated in contributors. If avoiding lock-in is the reason to pick it, run it self-managed, track the non-AWS share of maintainers over time, and keep data and query layers portable enough to move between OpenSearch and Amazon’s managed service.
Source: The New Stack
Content created with AI assistance and reviewed for accuracy.
Join the conversation
Stack Insiders is our free community for readers who want to go deeper — share resources, ask questions, and connect with others across every vertical we cover.
Join Stack Insiders →