Also known as: Snowflake Horizon.
At a glance
| Metadata ingestion and connectors | Snowflake objects natively; catalog-linked databases for AWS Glue and OneLake; external Iceberg catalogs |
| Lineage | Column level in Snowsight; Enterprise Edition or higher |
| Business glossary | Not documented as a glossary; semantic views and tags provide business definitions |
| Stewardship and workflows | Not documented as workflows; Internal Marketplace for data products |
| Access policies | RBAC, masking, row access, aggregation and projection policies (Enterprise or higher for policies) |
| Classification | Automatic sensitive data classification (Enterprise or higher) |
| Search and discovery | Cortex-generated descriptions, semantic views and tags; search UI not detailed on the pages checked |
| Deployment | Built into Snowflake on AWS, Azure and Google Cloud |
| Licensing | Included in Snowflake credits; Iceberg REST API 0.5 credit per million calls |
For Snowflake customers, Horizon Catalog is the governance layer already in the account: policies, lineage, classification and quality checks apply wherever the data is queried, including from Iceberg REST engines. Most of those controls need Enterprise Edition or higher, so check your edition first, and use a neutral catalog if much of your estate is outside Snowflake.
Overview
Snowflake Horizon Catalog is not a separate product: it is the set of catalog, discovery and governance capabilities inside Snowflake, which Snowflake describes as covering data inside and outside Snowflake across engines and clouds. It includes object tagging, sensitive data classification, masking and row access policies, access history, column-level lineage, Cortex-generated descriptions, semantic views for AI agents, data quality monitoring, the Trust Center, an Internal Marketplace for data products and replication for business continuity.
For interoperability Horizon Catalog integrates Apache Polaris and exposes the Horizon Iceberg REST Catalog API, so engines such as Spark, Trino, Flink, DuckDB, PyIceberg, Dremio and StarRocks can read and write Snowflake-managed Iceberg v2 and v3 tables using existing Snowflake roles and policies, with vended credentials. Catalog-linked databases and external tables connect to AWS Glue and Microsoft OneLake. Databricks Unity Catalog has no announced support for the API, according to the documentation.
Key features
- Masking and row access policies that also apply to Iceberg REST engines (Enterprise Edition or higher)
- Automatic sensitive data classification and object tagging
- Column-level data lineage in Snowsight (Enterprise Edition or higher)
- Data quality monitoring with data metric functions and root cause analysis
- Horizon Iceberg REST Catalog API: external engines read and write Snowflake-managed Iceberg tables
- Access history auditing and Trust Center security monitoring
- Semantic views, Cortex-generated descriptions and an Internal Marketplace for data products
Integrations and ecosystem
Pros and cons
Pros
- Already in every Snowflake account, with no separate product to deploy
- Policies follow the data to external Iceberg REST engines
- Read and write access to Snowflake-managed Iceberg tables from Spark, Trino, DuckDB and others
- Lineage, classification and quality checks in the same platform as the data
Cons
- Most governance controls need Enterprise Edition or higher
- Iceberg REST access covers only Snowflake-managed Iceberg tables, not native Snowflake tables
- Centred on Snowflake; no documented business glossary or stewardship workflows
Pricing
Horizon Catalog is part of Snowflake and has no separate list price; you pay for Snowflake credits and storage as in our Snowflake profile. Several governance features need Enterprise Edition or higher (On Demand credits in AWS US East at USD 3.00 for Enterprise versus USD 2.00 for Standard, per the Service Consumption Table effective 4 November 2026): masking policies, row access policies, data classification, access history, aggregation and projection policies, data metric functions and data lineage. Object tagging is in Standard, although some tagging features need Enterprise. The Horizon Iceberg REST Catalog API is available in all editions and is billed at 0.5 credit per million API calls as Cloud Services, with billing scheduled to start in the second half of 2026, plus cross-region egress where relevant.
Pricing checked on the vendor's official page on 7 November 2026. Prices change; confirm before buying.
Who it is best for
Snowflake customers on Enterprise Edition or higher who want governance and discovery in the platform and open Iceberg access for other engines.
Typical use cases
- Masking PII columns once and enforcing it in Snowflake and in Spark or Trino via Iceberg REST
- Classifying sensitive data automatically and tagging it for policy enforcement
- Tracing column lineage before changing a table
- Monitoring table quality with data metric functions and root cause analysis
Who should look elsewhere
- Estates where Snowflake is one of many platforms and a cross-platform business catalog is needed; see Atlan or Collibra.
Verdict
Horizon Catalog makes Snowflake a strong governance platform for its own data, and the Iceberg REST API now extends those policies to other engines. Its value depends on edition, since policies, lineage and classification need Enterprise or higher, and it does not replace a neutral catalog for estates that span many platforms.
Frequently asked questions
Does Snowflake Horizon Catalog cost extra?
There is no separate licence: it is part of Snowflake. Many governance features need Enterprise Edition or higher, which costs more per credit, and Horizon Iceberg REST API calls are billed at 0.5 credit per million as Cloud Services.
Which Horizon features need Enterprise Edition?
Masking policies, row access policies, data classification, access history, aggregation and projection policies, data metric functions and data lineage need Enterprise Edition or higher. Object tagging and the Iceberg REST Catalog API are available in Standard.
Can Spark or Trino write to Snowflake Iceberg tables?
Yes, through the Horizon Iceberg REST Catalog API, which supports read and write for Snowflake-managed Iceberg v2 and v3 tables. Native Snowflake tables and externally managed tables are not readable this way, and writes do not support CTAS, equality deletes, branches or tags.
How does Horizon Catalog compare with Databricks Unity Catalog?
Both are platform-native governance layers with fine-grained policies, lineage and Iceberg REST access. Unity Catalog is included in Databricks Premium and Enterprise tiers and has an open source edition; Horizon features depend on Snowflake edition, and Snowflake documents no support for Unity Catalog as a client of its Iceberg REST API.
Sources
- Snowflake Horizon Catalog (documentation)
- Horizon Catalog product page
- Snowflake: Overview of editions
- Snowflake: Data lineage
- Query Iceberg tables with external engines through Horizon Catalog
Checked November 2026.
Listed in: Data Catalog & Governance.
How we research reviews and what star ratings mean: our editorial method.