DataHub vs Collibra Pricing and Data Lineage for Teams Choosing a Catalog

DataHub vs Collibra is a choice between an open-source catalog with lineage built in and a governance platform that sells lineage per source type. DataHub Cloud lists at $75,000 a year on AWS Marketplace for up to 20 monthly active users, and DataHub Core is free under Apache 2.0. Collibra lists at $170,000 a year for one platform unit, and its lineage add-ons run $10,861.42 per technical lineage source and $27,153.61 per BI tool on the US GSA schedule. For a typical cloud stack, Collibra costs about three times as much at list.

Search for DataHub vs Collibra and page one is review aggregators: a CB Insights profile, two TrustRadius grids, an rfp.wiki score with a 44 percent confidence label, and a few pages about SAP Data Hub, which is a different product. None of them puts a dollar figure next to either name. Both vendors send you to sales. Both also sell through AWS Marketplace, where a transactable listing has to carry a real price, and Collibra's resellers file an itemized US price list on the federal GSA schedule. Those are the numbers this comparison uses.

DataHub vs Collibra pricing at a glance

DataHub figures from AWS Marketplace listing prodview-sum2rwvczvwvi, re-read 8 October 2026. Collibra AWS figures from prodview-6gqg3yc2e2rsu, re-read 7 October 2026. Collibra lineage lines from the US GSA schedule, read 7 October 2026. All are US list, before private offers and negotiated discounts.

The headline numbers already favor DataHub by more than half. The gap widens once lineage enters the quote, because on DataHub lineage is part of the product and on Collibra it is a set of line items.

What DataHub costs

DataHub has two editions. DataHub Core is the open-source project under Apache 2.0, and it includes the catalog, search, ingestion connectors, a business glossary and column-level lineage. It costs nothing to license and real money to run: a production deployment needs Kafka, a relational database, Elasticsearch or OpenSearch and the DataHub services on top, and your team owns upgrades and uptime.

DataHub Cloud is the managed product sold by the company formerly called Acryl Data, which renamed itself DataHub in 2025. Its only published price is the AWS Marketplace listing: one dimension, "Discover & Govern", described as "Up to 20 Monthly Active Users", at $75,000.00 for 12 months. That is $3,750 per active user per year. No 24 or 36-month term is offered there, and the listing states that fees are non-cancellable and non-refundable except as required by law. Cloud adds what Core lacks: freshness, volume, schema and column monitoring, anomaly detection, fine-grained access control and a 99.5 percent uptime SLA.

The open question in any DataHub Cloud quote is the user count. The listing never defines a monthly active user. A catalog is meant to be opened by analysts across the company, so ask in writing whether read-only viewers count before you sign. The full breakdown, including what Core costs to operate, is on our DataHub pricing page.

What Collibra costs

Collibra's AWS listing sells one dimension, the Collibra Cloud Platform, at $170,000 for 12 months, $340,000 for 24 and $510,000 for 36. The longer terms are exactly two and three times the one-year price, so committing longer buys nothing at list. The US federal schedule itemizes what sits underneath: a Standard cloud platform at $166,346.53 a year, Creator seats at $6,142.16 each, and the lineage and integration lines priced one by one.

What each package includes comes from the pricing document Collibra filed with the UK Digital Marketplace. Standard bundles 20 Creators, three metadata connectors, one technical lineage and one BI integration. Premier adds an ETL integration and a second BI tool. Only Ultimate removes the counting. Every source type past the bundled count is an annual add-on. Seats, modules and services are on our Collibra pricing page.

Data lineage is where DataHub and Collibra differ most

On DataHub, lineage is a feature of ingestion. Each connector reports table and column edges as it pulls metadata, and SQL-based lineage comes from a parser built on sqlglot that DataHub says reaches 97 to 99 percent accuracy. DataHub also documents what that parser cannot read. Scalar UDFs point lineage at their inputs without understanding the function, and dynamic identifiers are not supported. A missing edge looks exactly like a table nobody reads, so that list is worth knowing before you trust the graph for impact analysis.

On Collibra, lineage is a cloud-only product called Collibra Data Lineage, priced per source type. The GSA schedule lists Custom Technical Lineage at $10,861.42 a year and a BI integration at $27,153.61. Collibra's own documentation, dated July 29, 2026, lists the edges plainly: Power BI DirectQuery and live connections are not supported, BigQuery complex structures are not supported, and column-level lineage is not generated for tables created by SQL statements unless you add a shared storage connection. The CLI lineage harvester reached end of life on July 31, 2026, so lineage now runs through Edge. We price every line and quote every limit on Collibra data lineage cost.

Collibra's breadth is real, though. It scans Informatica PowerCenter, SSIS, IBM DataStage, SAP HANA calculation views, Teradata BTEQ and MicroStrategy, and ties each column to a governed business term with an owner. DataHub's connector catalog is wide in the modern stack and thinner in legacy ETL.

Is DataHub cheaper than Collibra?

Yes, at every published price point. Take a team tracing Snowflake and Postgres through dbt into Tableau and Power BI, with 20 people using the catalog. DataHub Cloud covers that at $75,000 a year at list, lineage included, as long as the active-user count stays at 20. On Collibra Standard, one database and one BI tool are bundled. The second database is $10,861.42, dbt counts as an ETL integration at about $27,172 (the UK add-on price converted), and the second BI tool is $27,153.61. That comes to about $231,534 a year at list, or roughly $156,500 more than DataHub. Self-hosting DataHub Core takes the license line to zero and replaces it with infrastructure and the engineering time to run it.

If your purchasing team runs vendor selection through formal procurement software, put both list prices in the request before the first call. Vendors anchor discounts on the number they think you know, and in this category most buyers know none.

Is DataHub a replacement for Collibra?

For a data or analytics engineering team, often yes. DataHub covers discovery, a glossary, ownership, column-level lineage and, in Cloud, data monitoring, which is what most engineering-led teams actually use a catalog for. For a governance office, usually no. Collibra's stewardship workflows, policy management, approval chains and business lineage tied to governed assets are the product, and DataHub's equivalents are lighter. Regulated reporting stacks with decades of legacy ETL also fit Collibra's scanner list better.

Which should you buy?

Buy DataHub when the team is engineering-led, the stack is a cloud warehouse with dbt, and you want lineage and discovery without line-item pricing. Start on Core if you have the platform engineers, or Cloud if you would rather pay $75,000 than run Kafka. Our OpenMetadata vs DataHub comparison covers the other open-source option.

Buy Collibra when a governance office owns the program, auditors need every report column tied to a governed term, and legacy ETL or SAP sits in the data path. Budget the lineage lines separately from the platform fee, because that is where the quote grows.

Buy neither when the real problem is narrower. Many catalog evaluations start because a column changed, three dashboards broke, and nobody could say what depended on it. That is a lineage and impact question. DataTrail connects to Snowflake, BigQuery, Redshift, Databricks or Postgres with a read-only role, parses query history and dbt artifacts into column-level lineage, and shows every downstream model and dashboard before a change ships through impact analysis. Freshness and schema-change alerts are included, and Starter, Team and Scale run $708 to $5,748 a year billed annually, with Enterprise above that, all on the pricing page and none with a user meter. It does not replace a glossary or stewardship workflows. The wider field is in our data catalog tools comparison.

DataHub vs Collibra FAQ

Is DataHub better than Collibra?

It depends on who owns the catalog. DataHub is the better fit for engineering-led teams that want open source, column-level lineage included and a lower price, with DataHub Cloud at $75,000 a year for 20 active users. Collibra is the better fit for governance offices that need stewardship workflows, policy management and lineage across legacy ETL, at $170,000 a year before lineage add-ons.

Is DataHub free?

DataHub Core is free to download and self-host under Apache 2.0, including the catalog, glossary and column-level lineage. Running it is not free, because production needs Kafka, a relational database, a search index and the DataHub services. Data monitoring and anomaly detection are only in the paid DataHub Cloud, which lists at $75,000 a year on AWS Marketplace.

Does Collibra include data lineage?

Partly. Collibra Standard bundles one technical lineage and one BI integration, according to its UK G-Cloud filing. Each extra source type is an annual add-on: $10,861.42 for a technical lineage integration and $27,153.61 for a BI tool on the US GSA schedule. Only the Ultimate package includes unlimited lineage, ETL and BI integrations.

Can DataHub import lineage from Collibra?

DataHub does not list a Collibra lineage connector among its ingestion sources, and most teams that run both use them for different jobs. A more common pattern is DataHub or a lineage-first tool for engineering, reading the warehouse directly, while Collibra stays the system of record for business terms and policies. Check DataHub's current connector list before planning a migration.

Prices change and both vendors negotiate. Treat every figure here as a list price to check your quote against, and ask each account team to confirm the user and source counts in writing.