Data Plane · Lakehouse Platforms

Starburst

Commercial data platform based on Trino, for querying data across many sources.

Overview

Starburst sells Trino: the open-source distributed SQL engine created by the people who built Presto, several of whom co-founded Starburst. Like Dremio it owns no storage and runs no Spark; its strength is querying across many systems at once.

Two products. Starburst Galaxy is the fully managed service, handling provisioning, maintenance and security. Starburst Enterprise is a supported self-hosted distribution of Trino, adding connectors, performance features and security on top.

Its architecture pattern has a name: the Icehouse, meaning Trino as the engine and Apache Iceberg as the table format, with Starburst automating compaction, purging and table optimisation.

Format support is the widest here: Iceberg, Delta Lake, Hudi and Hive, with Iceberg REST catalogues including S3 Tables, Polaris, Lakekeeper and Unity Catalog. Federation reaches Redshift, BigQuery, Snowflake, Synapse, MongoDB, Cassandra, Salesforce, SAP HANA and many more.

Governance is a layer called Gravity: roles and privileges, tags and policies, row-level security and column masking, with attribute-based access and SCIM on the higher tiers.

The AI angle is AIDA, an assistant that answers questions in natural language within the governed layer, plus an MCP endpoint so external agents in Claude, LangChain or Cursor can query through the same permissions. Token usage is billed separately.

Note that Trino itself belongs to the non-profit Trino Software Foundation, not to Starburst, which is a genuine difference from vendors that own their engine.

Key features and capabilities

The same headings are used for every lakehouse platforms entry, so two tools can be read side by side.

Architecture and storage
  • No storage tier: it queries your object store and source systems in place
  • Galaxy organises work as clusters for compute and catalogs for connections
  • Autoscaling on higher tiers, with fault-tolerant and accelerated execution modes
  • Enterprise deploys to Kubernetes or bare metal
Engines and workloads
  • One engine: Trino distributed SQL. No Spark or second proprietary engine
  • Streaming ingest from the Pro tier, with a documented Kafka pipeline
  • BI through client drivers rather than a bundled tool
  • Enterprise adds a caching service and query optimiser
Table formats and catalogues
  • All four major table formats: Iceberg, Delta Lake, Hudi and Hive
  • Iceberg REST catalogues including S3 Tables, Polaris, Lakekeeper and Unity Catalog
  • Object storage on AWS, Azure and Google Cloud
  • Federation to Redshift, BigQuery, Snowflake, Synapse, MongoDB, Cassandra, Salesforce and SAP HANA
Security and governance
  • Gravity provides roles, privileges, tags and policies
  • Row-level security and column masking, enforced even for natural-language questions
  • Attribute-based access control and SCIM from the Enterprise tier
  • Advanced governance integrations and compliance tooling on the top tier
AI and machine learning
  • AIDA answers questions in natural language across connected systems
  • An MCP endpoint lets external agents query through the same governed layer
  • Answers are grounded in shared definitions and enforced policy
  • No model training, serving or vector search
Ecosystem and integrations
  • A vendor-run developer centre with hands-on tutorials
  • REST API and data products in the Enterprise distribution
  • Client drivers for the usual BI and SQL tools
  • No marketplace; dbt support is not published
Where it runs and what it costs
  • Galaxy on AWS, Azure and Google Cloud, across 16 published regions
  • Clusters and data sources must be in the same cloud and region
  • Enterprise self-hosted on Kubernetes, bare metal or on-premises
  • Private networking to data sources from the Enterprise tier

Pricing

Usage-basedFree tier; Pro from $0.50 a credit

Galaxy is priced in credits, a universal unit of compute, with the rate rising by tier: Pro from $0.50, Enterprise from $0.75 and Mission-Critical from $1.00 a credit. A free tier runs up to three clusters for ever. The 30-day trial includes $500 of compute and Enterprise features, then drops to the free tier. Starburst Enterprise, the self-managed distribution, is quote-only, and AIDA token usage is billed separately at a rate that is not published.

Vendor pricing page →

Demos and videos

About Starburst

Starburst was founded in 2017 by Justin Borgman with co-creators of Presto and Trino, including Martin Traverso, and is based in Boston. It has raised $414m in total, most recently a $250m Series D in February 2022 at a $3.35bn valuation, led by Alkeon with Andreessen Horowitz, Coatue, Index and Salesforce Ventures among the investors. Galaxy and Enterprise are commercial distributions; the upstream Trino project is Apache 2.0 and governed by the Trino Software Foundation rather than by Starburst.

Founded 2017 · Boston, Massachusetts · starburst.io

Other lakehouse platforms tools

Databricks

Data Plane · Lakehouse Platforms

Lakehouse platform built on Apache Spark and open table formats, governed through Unity Catalog.

  • Commercial

Dremio

Data Plane · Lakehouse Platforms

Lakehouse query engine and semantic layer built on Apache Iceberg and Apache Arrow.

  • Open core

Microsoft Fabric

Data Plane · Lakehouse Platforms

SaaS analytics platform that unifies data engineering, warehousing, real-time analytics and Power BI on OneLake.

  • Cloud service

Drafted with AI assistance and checked against the vendor’s own documentation.