cloudparts

Onehouse

A managed lakehouse from the creator of Apache Hudi that runs in your own cloud account: it ingests data into Hudi, Iceberg or Delta Lake tables in your buckets, keeps them optimized, syncs them to other catalogs and runs Spark, SQL, Trino, Flink and Ray on them.

What it is
Lakehouses
Type
Commercial product
Made by
Onehouse
License
Proprietary
Facts checked
23 October 2026
Website

What it is

Onehouse sells the parts of a lakehouse that teams otherwise assemble from open-source pieces. The tables are open formats in your own object storage: Apache Hudi, which Onehouse’s founder Vinoth Chandar created at Uber, plus Iceberg and Delta Lake through Apache XTable, which Onehouse also started. Around them run managed services: OneFlow ingests from Kafka, database change streams and cloud storage; the Table Optimizer runs compaction, clustering, cleaning and file sizing; and OneSync registers the same tables and their permissions in catalogs such as AWS Glue, Unity Catalog and Snowflake, so other engines can read them without a copy.

Since 2025 Onehouse also runs the compute. The Onehouse Compute Runtime manages clusters inside your VPC, on any cloud or on your own Kubernetes, while Onehouse handles upgrades, patches and on-call. On it run Quanton, Onehouse’s own engine for Spark and SQL ETL jobs; Open Engines, managed Trino, Flink and Ray clusters; and Lakegres, a Postgres-compatible endpoint for low-latency queries on Hudi and Iceberg tables. The platform is proprietary; the company also maintains open-source projects and free tools, listed below. Onehouse was founded in 2021 and has raised $68 million from Greylock, Addition and Craft Ventures.

Why teams pick it

  • Data stays yours: everything runs in your cloud account, and tables are open Hudi, Iceberg or Delta Lake files that other engines can read without Onehouse
  • Table maintenance handled for you: compaction, clustering, cleaning and file sizing run asynchronously, so ingestion doesn’t stall on them
  • One copy of data for several platforms: OneSync, built on XTable, exposes the same tables to Snowflake, Databricks, Athena and others
  • Built by the people behind Hudi, the format it runs on, with the incremental, record-level processing Hudi was designed for

Watch out for

  • A commercial platform: the services and the Quanton engine are proprietary, and the Quanton Kubernetes operator needs an Onehouse account to run jobs
  • Performance claims are Onehouse’s own benchmarks and vary between pages, from “2-3x better price-performance” to “5x faster performance at up to 70% lower cost” for Quanton
  • Hudi is the home format: Iceberg and Delta Lake copies made by XTable carry only copy-on-write and read-optimized views, and the free LakeView tool supports only Hudi tables on S3 and GCS
  • Much of the compute side is new, from Quanton in 2025 to Lakegres in 2026, and Open Engines launched as a private preview

Alternatives

What all lakehouses have in common →Compare side by side →
Amazon S3 TablesAWSA bucket type for Apache Iceberg tables: S3 stores them, acts as their catalog, and compacts and cleans them up for you.Apache AmoroIncubatingOpen sourceAn incubating Apache “lakehouse management system”, started at NetEase as Arctic, that watches Iceberg tables and keeps them fast on its own: it compacts small files, rewrites deletes and expires old snapshots and files in the background.DatabricksCommercialA managed lakehouse platform built on Apache Spark and Delta Lake, with Unity Catalog for governance, on AWS, Azure or Google Cloud.DremioCommercialA SQL lakehouse engine built on Apache Arrow, with its own Iceberg catalog and “reflections” that speed up queries without anyone changing them.LakeOpsCommercialA commercial “control plane” for Apache Iceberg lakehouses that connects to your catalog, compacts and maintains tables on its own based on their health, and routes SQL from one endpoint to Trino, Snowflake, Athena, DuckDB, StarRocks or ClickHouse.Microsoft FabricAzureMicrosoft’s software-as-a-service analytics platform: Spark, a SQL warehouse, pipelines, a KQL eventhouse and Power BI all reading and writing one tenant-wide lake, OneLake, paid for from shared capacity.SpiralCommercialA closed-source storage and query engine for multimodal AI data (video, images, sensor traces and tables) that keeps everything as Vortex files in your own object storage, versions every change and streams results to GPUs; access is by request.Tiger LakeBetaCommercialTiger Cloud’s built-in Iceberg connector, in public beta: it copies a Postgres table or TimescaleDB hypertable into Apache Iceberg, then keeps it current from the write-ahead log.

How it fits with other technologies

Onehouse’s projects and products#

What the company has built, open and closed, oldest first. Dates are from Onehouse’s announcements.

Apache Hudi

open source, Apache

The table format Vinoth Chandar co-created at Uber in 2016 for record-level upserts and incremental reads. Onehouse’s ingestion and table services are built on it, and the company says it invests in the open project rather than an enterprise fork.

Apache XTable

open source, incubating

Announced as OneTable in February 2023 and open-sourced with Microsoft and Google that November. It translates metadata so one set of Parquet files reads as Hudi, Iceberg and Delta Lake, and entered the Apache Incubator in February 2024.

LakeView

free, June 2024

An observability service for Hudi tables: dashboards, a searchable timeline, skew and file-size insights, compaction backlog for merge-on-read tables, and Slack or email alerts. An Apache 2.0 extractor in your account pushes only metadata, never data files; it supports Hudi on S3 and GCS.

Table Optimizer

managed, June 2024

Runs compaction, incremental clustering, cleaning and file sizing on existing tables, written by Onehouse or by your own pipelines, asynchronously from ingestion. The launch post names Hudi, Iceberg and Delta Lake.

Onehouse Compute Runtime

January 2025

The managed runtime under all Onehouse services: clusters run in your VPC with budget caps and autoscaling, while Onehouse handles upgrades, patches and on-call. Onehouse calls it 100% compatible with open-source Spark.

Open Engines

April 2025

Managed clusters of open-source engines on the Compute Runtime over the same tables: Flink for streams, Trino for BI and Ray for AI and data science. It launched in private preview.

Quanton

May 2025

Onehouse’s engine for Spark and SQL ETL. Because pipelines rerun the same plan over the same tables, it tries to do only the work that changed since the last run. Jobs run on Spark or SQL clusters, or on your own Kubernetes through the Quanton operator for kubeflow’s Spark operator.

OneFlow

August 2025

Managed ingestion from streams such as Kafka, database change data capture and cloud storage into open tables in your buckets, using index-based incremental merge-on-read writes.

OneSync

managed

Registers tables in other catalogs, such as AWS Glue, Unity Catalog, Snowflake and Google Data Catalog, translating formats with XTable, and keeps table privileges in step across them.

Lakegres

2026

A Postgres-wire-compatible serving layer for Hudi and Iceberg tables, with a columnar cache tied to table commits and record-level and secondary indexes for point lookups. Launched in February 2026 as LakeBase and renamed in July after a naming clash with Databricks.

See it in a real stack

No architecture uses Onehouse yet. Drop it into the playground and wire it up yourself.Open playground →
Where it came from2016–2026

Vinoth Chandar co-created Hudi at Uber in 2016 to keep a Hadoop data lake up to date within minutes. He then watched companies move from warehouses to data lakes and struggle to run both: building a lakehouse took engineers who understood “at least 3 to 4 distributed systems or databases” and months to more than a year of work.

He founded Onehouse in 2021 and launched it in February 2022 with an $8 million seed round from Greylock and Addition. The first product was a managed data layer on Hudi that handled ingestion, compaction, clustering and catalog sync: “not another query engine, but a self-managing data layer.” A $25 million Series A followed in February 2023, along with OneTable, a converter between Hudi, Delta Lake and Iceberg that became Apache XTable.

In June 2024 a $35 million Series B led by Craft Ventures brought total funding to $68 million, and Onehouse launched LakeView and the Table Optimizer. From 2025 it moved into compute, with the Compute Runtime, Open Engines, the Quanton engine and OneFlow ingestion, and in 2026 the Lakegres serving layer.

Why it’s built this way

Build on open-source Hudi, not a fork
Chandar wrote at launch that Onehouse was “not here to fork an enterprise version of Hudi”, so tables stay readable by any engine that supports the open format.
Bridge the table formats instead of picking one
Users must be free to choose the right tool, Chandar wrote, and shouldn’t be locked into one table format; OneTable, later Apache XTable, translates metadata so one copy of the data serves Hudi, Iceberg and Delta Lake readers.
Run in the customer’s cloud account
The Compute Runtime offers a bring-your-own-cloud model in the customer’s VPC with a managed experience, so data stays inside the customer’s compliance boundary.
An engine tuned for repeated ETL runs
ETL pipelines rerun the same SQL plan on the same tables, so Quanton aims to do only the minimal work each run instead of recomputing from scratch.

Timeline

  1. 2016

    Vinoth Chandar and colleagues build Hudi at Uber

  2. 2021

    Onehouse is founded

  3. 2022

    Onehouse launches with an $8 million seed round (February)

  4. 2023

    $25 million Series A and OneTable (February); OneTable open-sourced (November)

  5. 2024

    $35 million Series B; LakeView and Table Optimizer launch (June)

  6. 2025

    Compute Runtime (January), Open Engines (April), Quanton (May), OneFlow (August)

  7. 2026

    Lakegres, first called LakeBase, launches in February and is renamed in July

Sources16 · facts checked 23 October 2026

Improve this page on GitHub