Iceberg, Delta Lake, Hudi, Paimon, DuckLake: State of Lakehouse Formats in 2026
As of mid-2026, Apache Iceberg has emerged as the industry's de facto interoperability standard, with major cloud providers, Snowflake, Databricks, and DuckDB all supporting it natively. However, the broader lakehouse table format landscape has not consolidated, as each competing format addresses distinct technical needs. Delta Lake remains dominant within the Databricks ecosystem, Hudi has reached its 1.0 release with advanced indexing capabilities, and Apache Paimon — originating from the Flink community — focuses on streaming-native table design. DuckLake, introduced in 2025 from the DuckDB project, challenges conventional metadata architecture with a simpler approach. Each format retains a dedicated community and specific workloads where it outperforms the others, making format choice a matter of use case rather than a settled standards debate.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Discussion (0)
Log in to join the discussion and vote.
Log in