OpenSearch has emerged as a formidable open source alternative to proprietary search platforms, doubling its downloads to 1.4 billion since transitioning to Linux Foundation governance almost two years ago.
In a conversation at KubeCon and CloudNativeCon, Bianca Lewis, executive director at the OpenSearch Software Foundation, discussed the platform’s evolution since its 2021 fork from Elasticsearch. What began as an AWS-led initiative to maintain an open source search solution has transformed into a vendor-neutral project with contributions from over 400 companies and more than 1.5 million monthly page visits.
Watch more videos from KubeCon & CloudNativeCon Europe
From AWS project to vendor-neutral platform
OpenSearch originated when Elasticsearch changed its license to a more restrictive model in 2021. AWS led the fork to continue providing an open source alternative, maintaining the former Elasticsearch service as the new OpenSearch service. However, the critical turning point came when AWS donated the project to the Linux Foundation.
“AWS decided OpenSearch is an Apache V2 license, completely open source and should not be under the management of any vendor,” Lewis explained. “Since the Linux foundation took over the ownership of OpenSearch, we’ve seen a phenomenal growth rate in the amount of contributors, downloads, participation and commits in the open source project.”
When the project was donated, OpenSearch had 700 million downloads. Today, that number has reached 1.4 billion, demonstrating sustained momentum in enterprise adoption.
Open source philosophy versus vendor control
Lewis emphasized that the decision between OpenSearch and proprietary alternatives shouldn’t focus solely on feature comparison. Instead, organizations should consider business models and data sovereignty.
“The true question that you should ask is not a focus of features, but what is the business model in data sovereignty that I need,” Lewis stated. “If you go in with the vendor-controlled projects, you’re giving your data and control of pricing over to the vendor.”
For developers concerned about feature parity, Lewis argued that OpenSearch’s pace of innovation matches or exceeds proprietary vendors. “The pace of innovation is immense. So obviously you can fulfill your developer requirements and needs on OpenSearch as easily as you can from any other vendor,” she noted.
Vector search and AI capabilities
Like other modern data platforms, OpenSearch has embraced vector search as a foundational capability. “All major databases today, if they’re not vector databases, they’re not relevant,” Lewis observed, noting that this applies to both Elasticsearch and OpenSearch.
The platform takes a model-agnostic approach to AI, allowing companies to choose which models they want to use rather than being locked into a vendor’s preferred solution. “Being an open source project, companies can choose which models they want to use,” Lewis explained.
Safety mechanisms for agentic AI
As organizations deploy agentic AI systems, OpenSearch has built native safety mechanisms that many proprietary platforms still lack. The platform provides full trace telemetry to monitor AI agent behavior throughout their entire lifecycle.
“You can trace back, and you can form anomaly detection and things that are querying things that shouldn’t be querying and trace why this happened in real time and alert on that,” Lewis described. This capability allows organizations to identify and respond to problematic AI behavior before it causes issues.
Enterprise deployments at scale
Recent case studies demonstrate OpenSearch’s readiness for large-scale enterprise deployments. Nvidia announced at GTC that OpenSearch has become the AI backbone of Nemo platform and agentic AI for Nvidia hardware. Atlassian, a premier member of the OpenSearch Foundation, runs over 300 search clusters with billions of documents at more than 99.99% uptime on OpenSearch. Changi Airport, one of the world’s largest airports, uses OpenSearch for its entire retail search infrastructure across more than 1,000 retail shops. The Changi deployment extends beyond simple location searches to include a complete recommendation system. “If you purchase something at a store, it’s got a whole recommendation system built behind that,” Lewis explained. “It’s changing the reality of those recommendations in real time and a massive scale that wasn’t possible before.”
Migration paths and compatibility
OpenSearch continues to see migrations from both Elasticsearch installations and other search platforms like Solr. However, Lewis noted that migration strategies are typically gradual rather than immediate wholesale transitions. A million-dollar ElasticSearch user is going to become a million-dollar OpenSearch user next month. Migration happens much more organically when there is a new use case. For organizations running older versions of Elasticsearch, migration to the corresponding OpenSearch versions remains relatively straightforward. However, newer versions of both platforms have diverged significantly, each with distinct strengths optimized for different use cases.
Future direction and community governance
As executive director, Lewis emphasized that the project’s direction is determined by community needs rather than any single entity’s vision. “I don’t say I want to change anything because one, it’s not up to me, it’s up to the community. That’s the whole idea of open source,” she stated. The recent announcements and case studies provide validation that the community-driven approach is working. With 2025 focused on vector search databases and 2026 expected to emphasize agentic AI, OpenSearch appears positioned to support enterprise needs across observability, search, and security monitoring use cases.
For organizations evaluating search and analytics platforms, OpenSearch offers a compelling combination of open-source flexibility, vendor neutrality, and enterprise-grade capabilities, backed by a growing community of contributors and users.