Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
Databricks announced on June 4, 2024, that it had agreed to acquire Tabular, the data-management company founded by Apache Iceberg creators Ryan Blue, Daniel Weeks, and Jason Reid. The price was not disclosed. The deal was strategically important because it brought Iceberg expertise into the company behind Delta Lake, helping Databricks strengthen support for multiple open table formats rather than requiring every customer to migrate immediately to Delta Lake.
As of 2026, the acquisition has translated into a broader Databricks Iceberg strategy that includes Unity Catalog-managed Iceberg tables, foreign Iceberg tables, an Iceberg REST Catalog, and support for Iceberg specification versions 1, 2, and 3. That does not make Delta Lake and Iceberg identical, eliminate catalog choices, or guarantee feature parity across every engine.
The short version
- What happened: Databricks announced a definitive agreement to acquire Tabular on June 4, 2024.
- Who Tabular was: A data-management and open-table-format company founded by Ryan Blue, Daniel Weeks, and Jason Reid, the original creators of Apache Iceberg.
- Price: Databricks did not disclose the purchase price. Secondary reports cited an estimated range of $1 billion to $2 billion, but that figure was not confirmed by Databricks.
- Why it mattered: Databricks gained direct Iceberg expertise while trying to make Delta Lake, Iceberg, and Apache Hudi more interoperable.
- What customers should understand: Databricks now presents a more credible multi-format lakehouse, but support still depends on the table type, catalog, runtime, cloud, engine, and operation.
The acquisition was not a purchase of a physical storage company, nor did Databricks buy Apache Iceberg itself. Apache Iceberg remains an open-source project. Databricks acquired Tabular’s people, products, and expertise in table management, metadata, catalogs, and interoperability on top of cloud object storage.
Free tools Windows power users keep installed
One-click scans. No signup required.
What Tabular built
Tabular operated at the data-management layer above object storage. Its work centered on Apache Iceberg, an open table format for analytical data stored in systems such as Amazon S3 and other cloud object stores.
#1 Best Overall
- Wiley
- Language: english
- Book - storytelling with data: a data visualization guide for business professionals
That distinction matters. A table format manages how files and metadata are organized so engines can safely perform operations such as reads, writes, schema changes, partition evolution, snapshots, and time travel. A catalog helps engines discover tables and coordinate access. Neither is the same as manufacturing or selling the underlying storage hardware.
Tabular was founded by Ryan Blue, Daniel Weeks, and Jason Reid, who were closely associated with the creation and development of Apache Iceberg. The value to Databricks was therefore broader than a conventional software product acquisition. It included engineering talent, deep knowledge of Iceberg’s design, experience with open table metadata, and credibility with organizations building multi-engine data platforms.
Tabular described the transaction as a definitive agreement to join Databricks in its announcement. Databricks subsequently described Tabular as acquired and framed the combination as bringing together the original creators of Iceberg and contributors to Delta Lake.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Why Databricks wanted Tabular
Databricks created Delta Lake and uses it as the default format for Databricks tables. Delta Lake adds a transaction-log and table-management layer to Parquet-based data in cloud object storage. It is deeply integrated with Databricks compute, governance, optimization, and data-engineering workflows.
Apache Iceberg developed a different kind of market position. It became a widely adopted open table format supported by ecosystems including Apache Spark, Apache Flink, Trino, Snowflake, AWS, and other platforms. Apache Hudi is another open table format used for incremental processing and data-lake workloads.
For enterprise buyers, the format question is not merely technical. It affects:
- Whether several query and processing engines can work with the same data.
- How difficult it is to move between cloud providers or platforms.
- Whether data must be duplicated for different tools.
- Which catalog controls metadata and permissions.
- How governance, maintenance, streaming, updates, and deletes are implemented.
- How much leverage a customer has when negotiating with a platform provider.
Databricks publicly emphasized interoperability. Its acquisition announcement said the company wanted to bring together the teams behind Delta Lake and Iceberg and improve compatibility among major open table formats, including Hudi. The move also placed Databricks in a stronger position as the lakehouse control point expanded beyond storage and compute into catalogs, governance, and AI workloads.
Delta Lake and Apache Iceberg are related, not interchangeable
| Area | Delta Lake | Apache Iceberg |
|---|---|---|
| Basic role | A transaction-log and table-management layer for data in object storage. | An open table format for analytical workloads in object storage. |
| Databricks relationship | Created by Databricks and the default format for Databricks tables. | Supported by Databricks, but developed independently as an open-source project. |
| Typical strengths | Deep integration with Databricks operations, governance, optimization, and engineering workflows. | Cross-engine adoption, schema evolution, hidden partitioning, time travel, and multiple specification versions. |
| Catalog model | Can be used with catalog and platform arrangements that differ by deployment. | Can be used with multiple catalog implementations, including REST-based and cloud-oriented catalogs. |
| Interoperability | Databricks supports interoperability mechanisms such as UniForm. | Supported across a broad ecosystem, but actual capabilities vary by engine, catalog, and client version. |
Databricks’ Delta Lake documentation describes Delta as the foundation of its lakehouse and default table format. Its Iceberg documentation describes Iceberg capabilities including schema evolution, time travel, hidden partitioning, and support for multiple specification versions.
“Open format” also has limits. A table can use an open format while the surrounding catalog, authentication system, governance controls, optimization service, or compute engine remains vendor-specific. Support for reading a table does not automatically mean support for safe concurrent writes, row-level deletes, streaming updates, schema changes, or every advanced feature.
What Delta Lake UniForm does
Databricks positioned Delta Lake UniForm as a compatibility mechanism. It allows Delta tables to be exposed through Iceberg and Hudi-compatible interfaces while retaining the underlying data rather than creating an entirely separate copy for each format.
The important technical point is that interoperability depends on metadata, not only on the fact that both systems may use Parquet files. Table engines need compatible metadata describing snapshots, schemas, partitions, transactions, and other table state. UniForm generates the relevant compatibility metadata asynchronously so supported engines can access the same underlying data through another interface.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
UniForm can reduce the need for immediate migrations and duplicated datasets, but it does not turn every Delta table into a universally writable native Iceberg table. These are different scenarios:
- Reading through another interface: An engine may read a Delta table using an Iceberg-compatible view of its metadata.
- Native format conversion: Data and metadata are converted or rewritten into a table that follows another format’s native rules.
- Two independently writable representations: Separate Delta and Iceberg tables are maintained and synchronized, creating additional operational complexity.
Customers must verify whether the chosen engine can write, update, delete, commit concurrently, evolve schemas, and preserve snapshots through the compatibility path. They must also account for metadata-generation delays, unsupported features, client-version requirements, and differences in how engines interpret table specifications.
What changed for Databricks customers by 2026
The acquisition’s practical importance is clearer in Databricks’ current product documentation than it was at announcement time. Databricks now documents support for both Delta Lake and Apache Iceberg, including several ways to use Iceberg data:
- Unity Catalog-managed Iceberg tables: Iceberg tables managed through Databricks’ governance and catalog layer.
- Foreign Iceberg tables: Tables managed by an external catalog and made visible to Databricks. In the documented workflow, these are read-only and have more limited platform support.
- Iceberg REST Catalog access: External engines can connect to Databricks-managed Iceberg tables through a REST Catalog interface.
- External-engine access: Databricks documents access paths for engines including Apache Spark, Apache Flink, and Trino, subject to requirements and supported operations.
- Specification support: Databricks documents support for Iceberg specification versions 1, 2, and 3.
Databricks announced general availability in May 2026 for Unity Catalog-managed Iceberg tables, foreign Iceberg tables, and Iceberg v3 features in its release notes.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsThe current documentation also states that managed and foreign Iceberg table use requires Unity Catalog. For the documented AWS workflows, Databricks Runtime 16.4 LTS or later is required, and managed-table workflows may require serverless compute. Databricks recommends Iceberg client version 1.9.2 or later in its current documentation. Requirements can differ by cloud, table type, operation, and deployment.
For external access, Databricks documents its Iceberg REST Catalog and client configuration. Customers still need to validate authentication, credential vending, cloud-storage permissions, networking, firewall rules, private connectivity, and the exact read/write behavior of each external engine.
What this means for different customers
Organizations already using Iceberg
They may be able to use Databricks compute, governance, or engineering features without converting every existing table to Delta. That can reduce migration pressure, but it does not guarantee that all existing Iceberg workflows will behave identically inside Databricks.
Databricks-first organizations
They can continue using Delta Lake as the native operating environment while making selected tables accessible to Iceberg-compatible engines. This may be useful when a data science, streaming, or analytics team has standardized on a different engine.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchMulti-cloud and multi-engine organizations
The acquisition is most relevant where data must be shared among engines such as Spark, Flink, Trino, Snowflake, or cloud-native services. The benefit is greatest when avoiding a full copy-and-convert cycle matters more than having every workload use the same vendor platform.
Buyers seeking neutrality
Databricks’ Iceberg support can reduce format lock-in while increasing the importance of Unity Catalog, Databricks compute, credential vending, and platform governance. Those are separate architectural dependencies. A format may be portable even when the complete operating environment is not.
Rank #4
Open source: benefit and concern
Databricks framed the acquisition as a commitment to open formats and open-source data infrastructure. That could produce meaningful benefits:
- More engineering resources devoted to compatibility.
- Better practical interoperability between Delta Lake and Iceberg ecosystems.
- Less pressure on customers to perform expensive format migrations.
- More mature support for external engines and catalogs.
There are also legitimate concerns, although they are not proof that Databricks has abandoned Iceberg or violated open-source governance:
- Databricks may continue to prioritize Delta Lake and the economics of its own platform.
- Open table formats can be surrounded by proprietary catalog, governance, optimization, and security services.
- Concentrating important Iceberg expertise inside a major commercial platform could concern users who prefer neutral ecosystem stewardship.
- Marketing claims about compatibility may be broader than actual support for writes, deletes, streaming, or advanced table features.
Open-source project governance, commercial ownership of a company employing project contributors, openness of the table format, and neutrality of the surrounding service are different questions. They should not be treated as interchangeable.
Competitive significance
The transaction fit into a broader contest over who controls the modern data platform: the table format, catalog, execution engine, governance system, and AI workload environment.
- Snowflake: A managed cloud data platform with strong SQL analytics, governance, sharing, and Iceberg-oriented capabilities.
- AWS: A composable approach using services such as S3, Glue, Athena, and EMR, giving customers substantial control but requiring more architectural assembly.
- Microsoft Fabric and OneLake: Closely integrated with Microsoft 365, Power BI, Azure, and Microsoft identity services.
- Google Cloud: A major cloud platform with its own analytics, storage, governance, and AI services.
- Dremio: Focused on lakehouse query, semantic access, and open-data architectures.
- Starburst and Trino-based deployments: Focused on federated query access across data sources and catalogs.
- Independent open-source deployments: Offer greater control and neutrality but require organizations to operate more of the catalog, security, maintenance, and reliability stack themselves.
It is reasonable to view the acquisition as part of Databricks’ response to competitive pressure from Snowflake and cloud providers. It is too narrow to say the deal was solely defensive against Snowflake. Databricks’ stated rationale centered on open formats and interoperability, while the commercial effect was to make its lakehouse platform more relevant to customers who would not accept a Delta-only strategy.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Risks and limitations buyers should not overlook
Read compatibility is not write compatibility
An external client may be able to read a table while lacking safe support for updates, deletes, concurrent commits, schema evolution, snapshot maintenance, or metadata written by another engine. Test each operation required by the workload.
Foreign tables are not equivalent to managed tables
Databricks documents foreign Iceberg tables as read-only in the relevant workflow, with limited platform support. That is materially different from a fully managed, writable Iceberg table.
Runtime and cloud matter
Documentation for one Databricks cloud or runtime should not be generalized to every deployment. Version requirements, serverless dependencies, supported clients, identity flows, and feature availability can change by cloud and table type.
Catalog choice remains important
Decide whether the system of record for table metadata is Unity Catalog, AWS Glue, a Hive Metastore, Snowflake Horizon Catalog, another REST Catalog, or a different service. The catalog affects discovery, permissions, credentials, concurrency, and exit options.
Metadata evolution can create friction
REST Catalog access and UniForm reduce integration work, but they still depend on metadata generation, client versions, specification support, and feature mapping. No supported interface means that every engine can safely perform every operation on every table.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →Open formats do not remove infrastructure costs
Customers still pay for object storage, requests, compute, network transfer, catalog operations, governance, optimization, support, and potentially data duplication. The acquisition does not eliminate those cost categories.
Questions to answer before adopting Databricks for Iceberg
- Which catalog owns the table metadata?
- Will Databricks only read the data, or must it write, update, and delete?
- Which Iceberg specification version is required?
- Are row-level changes, schema evolution, partition evolution, and concurrent writes supported by every participating engine?
- Are external clients read-only or read/write?
- Does the workflow require Databricks Runtime 16.4 LTS or later?
- Is serverless compute required?
- How will external clients authenticate and obtain cloud-storage credentials?
- Are networking, IAM, firewall, private-link, and cross-cloud requirements satisfied?
- Which optimizations are portable, and which depend on Databricks?
- How will compaction, retention, metadata cleanup, and table maintenance be handled?
- What happens if the organization later moves away from Databricks?
How to interpret the reported purchase price
Databricks’ primary acquisition announcement did not disclose a transaction value. Some secondary reporting cited an estimated range of $1 billion to $2 billion, reportedly based on Wall Street Journal reporting, but Databricks did not confirm that figure in its announcement.
The accurate description is therefore: the price was undisclosed; reports put the value somewhere between $1 billion and $2 billion, but Databricks did not confirm that figure. It should not be presented as a verified $1 billion, $1.5 billion, or $2 billion purchase price.
The bottom line
Databricks’ Tabular acquisition was a control-point strategy, not simply a storage purchase. Databricks acquired the people and product expertise needed to make Iceberg compatibility credible while retaining Delta Lake and Databricks as its preferred native operating environment.
Recommended Free Tools
The result is a stronger multi-format lakehouse story. Existing Iceberg users may have more ways to use Databricks without converting all their data, and Databricks customers have more options for sharing tables with external engines. But the acquisition did not merge Delta Lake and Iceberg, end the format competition, or remove the need to choose catalogs, engines, governance models, and cloud architecture.
For buyers, the most defensible conclusion is that Iceberg became too important for Databricks to ignore. Format choice may now be less disruptive, but platform choice still carries significant technical, operational, and financial consequences.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
