Jan,

Apache Iceberg unlocks things that weren't practical before: time travel queries, schema evolution without breaking downstream consumers, partition pruning that actually works at scale. But those capabilities only hold if your platform is Iceberg-native, not just Iceberg-compatible.

Compatibility layers add translation overhead and create drift risk. Dremio is built on Iceberg natively. That means you get full table format control, branch-and-merge workflows for safe data changes, and agents that can query your lakehouse directly without custom connectors.

Here's what Iceberg-native architecture makes possible in practice:

A 2026 Introduction to Apache Iceberg

Alec Merced
Head of Developer Relations, Dremio