Elastic Went Back to Open Source. Now It Has to Prove the Fork Didn't Already Win.

$1.37 billion: Elastic N.V.'s revenue in fiscal 2025, up 17% year-over-year, according to its 10-K filing Source. Elastic Cloud, the company's SaaS product, grew 26% and now represents 46% of total revenue. The numbers are solid. The story is complicated.

In January 2021, Elastic changed the license of Elasticsearch and Kibana from Apache 2.0 to the Server Side Public License, a non-open-source, source-available license designed to prevent cloud providers — specifically AWS — from offering hosted Elasticsearch as a service without contributing back Source. The move was correct in theory and destructive in practice. AWS responded by creating OpenSearch, a fork of the last Apache 2.0-licensed version of Elasticsearch, and launched it as an AWS managed service in September 2021 Source. OpenSearch inherited the developer community Elastic had spent a decade building. It also inherited the enterprise customers who had built on Elasticsearch assuming the project was sustainably open.

Elastic announced in August 2024 that it was adding the AGPL license, a Free Software Foundation-approved open-source license, as an option alongside the SSPL Source. The announcement's tone was an apology. Elastic's CTO, Shay Banon, acknowledged that the SSPL "worked" in preventing Amazon-style behavior but "created a rift in the community." The AGPL addition was positioned as a fix, a way to call Elasticsearch open source again and rejoin the ecosystem the license change had fractured.

The relevant question for investors is not whether the AGPL move was the right thing to do. It was. What matters is whether the community and customer goodwill translates into adoption that moves the revenue needle, or whether OpenSearch has already captured enough of the addressable market that Elastic's return to open source carries no commercial consequence.

OpenSearch's growth suggests the latter is a real risk. AWS does not disclose OpenSearch revenue separately, but the service is available in 32 AWS regions and is integrated with every AWS data service from S3 to SageMaker. AWS's OpenSearch Serverless offering, which launched in January 2023, eliminates cluster management entirely and competes directly with Elastic Cloud on managed search Source. The developer community around OpenSearch is active. The project has over 10,000 GitHub stars, a regular release cadence, and a governance model that includes non-Amazon maintainers. It is not a ghost fork that Elastic can wish away.

Elastic faces a second threat that has nothing to do with AWS. Procurement teams across the Fortune 500 are asking whether a dedicated search platform is even necessary once an LLM can answer questions about a company's own data. Elastic's answer is the Elasticsearch Relevance Engine, which combines vector search with traditional keyword search in a hybrid retrieval model built to feed RAG pipelines. The offering is competitive. But the addressable market for search infrastructure is being redefined by the same AI tools Elastic hopes will drive its next growth phase. Once every database vendor adds vector search and every cloud provider ships a managed RAG pipeline, the question of whether enterprise search is a standalone product or a feature bundled into everything else stops being rhetorical.

Elastic's return to open source is the right strategic move, made three years late. The reputational damage has been repaired. The commercial damage from the fork remains to be measured. If Elastic Cloud can grow through the AI transition and the AGPL license can repatriate a meaningful share of the OpenSearch install base, the stock is undervalued. If OpenSearch has already become the default managed Elasticsearch service in the AWS ecosystem — which is where most Elastic workloads run — the return to open source is a footnote to a market share shift that already happened.

Sometimes the apology lands. Sometimes the fork already ate your lunch.