Rockset

Real-time cloud-native search and analytics database service acquired by OpenAI in 2024.

Updated April 2026

Overview

Website: rockset.com
Acquired by: OpenAI
Segment: Data Platforms & Warehouses
Posture: Analytical Warehouse & Lakehouse

Product overview

Rockset provided a serverless, real-time analytics database supporting SQL queries on semi-structured data from sources like DynamoDB, Kafka, and S3, with automatic indexing for sub-second ingest and low-latency queries including search, aggregations, and vector similarity. It used a converged index (inverted, columnar, row) built on RocksDB for high-concurrency applications like personalization, dashboards, and AI RAG. Used by developers and enterprises for operational analytics; distinguished by schema-on-need, disaggregated compute-storage, and no-ETL streaming ingest. Acquired by OpenAI in June 2024; public service shut down September 30, 2024, with technology integrated into OpenAI products. ,

Revenue model

Usage-based: compute (virtual instances from $0.32/hour shared, higher for dedicated/premium), tiered storage ($0.25/GB-month first 100GB decreasing to $0.13/GB+), free developer tier (limited ingest/storage). No fixed subscriptions.

Moat

Proprietary Technology
Scale Advantages
Data Flywheel

Rockset's key competitive moat is its proprietary technology for real-time indexing and vector search on unstructured data, enabling sub-second SQL queries at scale that are difficult for competitors to replicate without equivalent engineering expertise.