2025 was a defining year for lakeFS.
Across open source and Enterprise editions, we shipped major capabilities that expanded lakeFS from a powerful data versioning layer into a control plane for AI-Ready Data – spanning structured and unstructured data, multiple public and private clouds, and a growing ecosystem of analytics and ML engines.
Here’s our Top 10 lakeFS product releases of 2025, in no particular order.
1. lakeFS Becomes an Iceberg REST Catalog
This was the year lakeFS officially became an Iceberg REST Catalog.
With this release, users can:
- Version Apache Iceberg tables using lakeFS commits and branches
- Access those tables from any Iceberg-compatible engine (Spark, Trino, Flink, Snowflake, and more!)
- Manage structured (Iceberg) and unstructured data (files) together in a single repository
This unlocks a powerful new workflow: atomic versioning across tables, images, videos, models, and metadata – all governed by the same commit semantics and quality and governance controls.
2. Multiple Storage Backends in a Single lakeFS Installation
We added a ”behind the scenes”, yet highly valuable, enterprise feature: the support for multiple storage backends side-by-side, providing unified versioning and seamless access across siloed data stored in different systems.
For example, you can now manage
- One repository backed by Amazon S3
- Another backed by Google Cloud Storage
- Another backed by an on-premise NetApp StorageGRID
all within the same lakeFS deployment.
This enables hybrid and multi-cloud architectures without duplicating infrastructure or operational overhead.
3. Metadata Search (via the Iceberg REST Catalog)
Metadata became a first-class citizen in lakeFS this year.
We added support for:
- Exposing object metadata
- Querying user-defined tags and system metadata attached to objects
- Surfacing this metadata through the lakeFS Iceberg REST Catalog
The result: faster data discovery, richer governance, and more powerful integrations with downstream query engines. These capabilities are backed by lakeFS itself, so all queries are reproducible by default
4. IAM Authentication for lakeFS API Access
Security and ease-of-use got a big boost with native IAM authentication.
On AWS, users can now:
- Authenticate to the lakeFS API using an existing IAM role
- Eliminate long-lived access keys
- Align lakeFS access with existing cloud security practices
This makes lakeFS a more secure system to deploy in enterprise environments and provides a more cloud-native experience.
5. lakeFS Mount: Major Performance Overhauls + Windows Support
We significantly revamped lakeFS Mount, focusing on performance and stability.
Key improvements:
- Up to 50% faster reads and writes
- Better behavior under concurrent workloads
- More predictable performance for large data operations
- Windows support, enabling lakeFS Mount for teams working in Windows-based environments
For teams using lakeFS Mount as a simple and highly performant bridge between file-based tools and lakeFS versioning, this was a game-changer and Windows support made it accessible to even more workflows across the organization.
6. Squash Merge for Cleaner Histories
Squash merge arrived in 2025, bringing a familiar Git-style workflow to data.
It allows:
- Merging multiple commits from a source branch
- Producing a single logical commit on the destination branch
This results in a cleaner commit history which is more easily auditable, and clearer lineage and provenance for changes.
7. GeoJSON Preview Support in the lakeFS UI
Geospatial data became easier to work with thanks to native GeoJSON viewing in the lakeFS UI.
Users can now:
- Open and inspect GeoJSON files directly
- Validate outputs without leaving lakeFS
- Collaborate more easily on spatial datasets
A small feature with outsized impact for GIS and location-aware workflows.
8. A Complete lakeFS UI Overhaul
While we spent a lot of time on complex enterprise-scale features, we also thought of our many users and shipped a fully refreshed UI:
- Faster page loads
- Improved readability
- Clearer navigation
- Better usability across the board
Whether browsing objects, reviewing commits, or comparing branches, the new UI makes lakeFS more approachable for both engineers, analysts and data scientists.
9. Redis-Compatible, MemoryDB-Backed KV Store
We introduced a Redis-compatible key-value store option (including AWS MemoryDB support) for managing uncommitted data.
Benefits include:
- Lower latency
- Higher throughput
- In write-heavy workloads, up to 5× performance improvements compared to PostgreSQL or DynamoDB
This dramatically improves performance for large scale batch and streaming workloads.
10. A Stronger Foundation for the Future
Beyond headline features, 2025 also delivered:
- Performance optimizations across the system
- Improved interoperability with modern data engines
- A clearer separation between control plane, metadata, and storage
For lakeFS open source, this translated into 778 PRs that got merged this year.
Together, these changes set the stage for what’s next: deeper governance, richer metadata, and even tighter integration with the open data ecosystem.
2025 transformed lakeFS from “Git for data lakes” into something bigger:
A control plane for AI-Ready Data
Thank you to the community and customers who shaped these releases – and stay tuned for what’s coming next! Did this make you curious? Try lakeFS now or book a demo with our friendly sales team.
