Elastic Launches AI Vector Database Weeks After Posting 15% Revenue Growth, Still Running a GAAP Loss

Elastic (NYSE: ESTC) announced Elasticsearch Vector Database on September 11, a serverless product the company says lets developers build AI search and retrieval systems without assembling the infrastructure by hand, according to a company statement published on ir.elastic.co.

The product targets retrieval-augmented generation, or RAG, and AI agent workloads, the two dominant patterns enterprises are using to bolt large language models onto their own data. Elastic says the database automates document chunking, embedding, indexing, and query-time retrieval, tasks that normally require stitching together separate tools.

What The Product Actually Does

Elastic built the system around a single field type that handles indexing, embeddings, and chunking together, plus hybrid search that runs full-text and vector queries against the same index, according to the company's announcement. The database supports text, image, and multi-modal vectors in one index, and it plugs into Elastic's Inference Service for managed GPU-based embedding and reranking, including third-party models from Jina AI.

The headline technical claim is Elastic's Better Binary Quantization, or BBQ, which the company says cuts vector memory use by up to 32 times while keeping search recall high, allowing the system to scale to hundreds of billions of vectors. Crypto Briefing reported that the underlying VectorDB index mode and automatic calibration features had already been generally available since August 2026, meaning the core technology had roughly a month of production use before Elastic packaged it into a named product.

On pricing, Elastic is charging based on data volume and search capacity rather than the compute-unit pricing common among dedicated vector database vendors, a structure company general manager Ajay Nair said avoids surprise bills for customers running AI workloads at scale.

The Numbers Behind The Launch

The launch landed less than three weeks after Elastic reported first-quarter fiscal 2027 results on August 27, for the period ended July 31. Total revenue hit $478 million, up 15% year over year, according to Yahoo Finance. Current remaining performance obligations grew 21% and total remaining performance obligations grew 27%, both measures of contracted future revenue.

Sales-led subscription revenue, which strips out Elastic's older Monthly Elastic Cloud consumption model, grew 18% to $399 million. Elastic added more customers spending over $100,000 annually than in any prior quarter, pushing that group past 1,800, up from more than 1,720 in the fourth quarter of fiscal 2026 and over 1,550 a year earlier. Net expansion rate held near 111%, and the company generated $143 million in adjusted free cash flow for the quarter.

Elastic is guiding second-quarter fiscal 2027 revenue to $486 million to $487 million and projecting a swing to positive GAAP operating margin, with full-year non-GAAP operating margin guided to roughly 19.4%, up from 16.2% in the quarter just reported.

The growth numbers come with a caveat. Elastic posted a GAAP operating loss of $24 million for the quarter, a negative 5% operating margin, and a GAAP net loss per share of $0.16. Non-GAAP figures were healthier, with $77 million in non-GAAP operating income and $0.70 in non-GAAP earnings per share, but the gap between the two accounting methods means Elastic is still not profitable by standard measures investors use to judge a company's bottom line.

A Split Market Verdict

Elastic's stock has not moved in lockstep with the growth story. Shares closed at $83.39, down 9.17% over the prior seven days and down 3.80% over 30 days, according to Simply Wall St, even after a 90-day return of 38.73% that reflected earlier momentum. The stock's total shareholder return is down 5.57% over the past year and down 47.99% over five years.

One widely followed valuation model cited by Simply Wall St pegs fair value at $74.52 using an 8.55% discount rate, implying the stock trades above what that model considers justified. Analyst price targets on the stock span from the mid-$50s to just above $100, a wide range that reflects genuine disagreement over how fast AI-driven demand will show up in Elastic's actual earnings.

That valuation gap cuts both ways. Elastic trades at roughly 23.3 times earnings, below the broader U.S. software sector average of 30.1 times and well below a peer group average near 46.6 times, according to Simply Wall St's own numbers, an argument for the stock being cheap relative to comparable companies even if one specific fair-value model says otherwise. Investors weighing the story have two data points to reconcile: a business posting double-digit growth and a fresh AI product, against a stock that has fallen nearly 10% in the past week and remains unprofitable under GAAP accounting. Elastic's next test comes when it reports second-quarter fiscal 2027 results against its own guidance of $486 million to $487 million in revenue and a promised return to positive GAAP operating margin.