Elastic (ESTC) Bets Big On AI Search With New Database

On September 11, Elastic (NYSE:ESTC) introduced its Elasticsearch Vector Database, a serverless offering designed to enable developers to create large-scale vector search and AI applications without the need for manual infrastructure assembly. The new database streamlines processes such as document chunking, embedding and reranking model hosting, index configuration, and query-time retrieval.

Elastic incorporated expert-tuned defaults, a unified field type for indexing, embeddings, and chunking, and hybrid search capabilities that blend full-text and vector retrieval within a single query. The system can scale to handle hundreds of billions of vectors using Elastic's Better Binary Quantization, which reduces vector memory by up to 32 times while maintaining high recall.

Pricing is based on data and search capacity, as opposed to the more opaque compute units typically used in pure-play vector database pricing.

Elastic's first-quarter fiscal 2027 results, released on August 27, demonstrated accelerating growth as total revenue reached $478 million, a 15% increase year-over-year. Current remaining performance obligations rose by 21%, while adjusted free cash flow generated during the quarter amounted to $143 million. Guidance for the second quarter of fiscal 2027 forecasts revenue between $486 million and $487 million, with a shift to a positive GAAP operating margin.

Non-GAAP figures indicate $77 million in operating income and $0.70 in non-GAAP diluted earnings per share, highlighting the gap between GAAP and non-GAAP profitability driven by factors like stock-based compensation. The company continues to invest in growth through share repurchases, totaling about 0.8 million shares for approximately $40 million, and the acquisition of Deductive AI, an observability platform, though both moves are not considered significant red flags.