Onehouse Open Data Lakehouse Platform

Onehouse is a Sunnyvale, California company founded and led by Vinoth Chandar, the original creator and PMC Chair of Apache Hudi. It provides open data infrastructure that runs inside the customer's own cloud account, so companies can build a lakehouse on open table formats without handing data to a proprietary warehouse.

The platform covers ingestion, transformation, serving and table management. OneFlow ingests from cloud storage, database CDC and Kafka-compatible streams incrementally; Table Optimizer runs compaction and clustering; OneSync keeps tables and permissions in sync across Glue, Unity, Snowflake and Hive catalogs; and Apache XTable converts between Hudi, Iceberg and Delta. The Quanton engine runs Spark and SQL jobs, Open Engines manages Trino, Flink and Ray clusters, and Lakegres offers a Postgres-compatible endpoint for BI and AI tools. It deploys in a VPC on any cloud or on Kubernetes, including on-premise.

• Incremental ingestion: OneFlow loads from object storage, database CDC and Kafka-compatible streams.
• Quanton engine: Runs existing Spark and SQL ETL jobs without code changes.
• Open Engines: Managed Trino, Flink and Ray clusters on the same tables.
• Table Optimizer: Continuous compaction, clustering and cleaning of lakehouse tables.
• OneSync and XTable: Sync across Glue, Unity, Snowflake and Hive and convert between Hudi, Iceberg and Delta.
• Lakegres serving: Postgres-compatible endpoint for SQL, BI and AI tools.

A Vietnamese e-commerce or fintech company running Spark on EMR can move ETL to Quanton to lower cloud cost. A bank can ingest core banking changes via CDC into an open lakehouse that both BI tools and data science teams use. A group with several cloud providers can keep one copy of data readable by multiple engines.

Data stays in open formats in the customer's own account, so any engine can read it and vendor lock-in is reduced. Onehouse states that Quanton delivers 2-3x better price-performance for ETL and guarantees at least 50% lower Spark infrastructure cost for existing Hudi users. Automated table maintenance and cross-catalog sync cut manual data engineering work, and a free trial and free Spark cost analyzer are available.

Enterprises are shifting ETL spend away from proprietary warehouses toward open lakehouses where storage and compute can be chosen separately. Interoperability among Hudi, Iceberg and Delta and the need to keep data inside one's own cloud account for compliance are driving demand for managed open lakehouse services.

Headquartered at 150 Mathilda Place, Sunnyvale, California; founder and CEO Vinoth Chandar created Apache Hudi. Steward of open-source Apache Hudi and Apache XTable (Incubating). Launched the Quanton engine in May 2025. Pricing is pay-as-you-go or a flat yearly fee.