Iceberg Needs an Open Catalog, Not a Walled Garden

Apache Iceberg promises vendor-neutral data storage via open table formats, but proprietary catalogs can recreate the same lock-in at a higher layer. The catalog defines the real working contract around Iceberg tables — governing permissions, schema evolution, multi-engine coordination, and governance. When a catalog favors one compute engine over others, customers lose the architectural freedom Iceberg was meant to provide. As AI and real-time operational workloads demand fresh data and low-latency access, the catalog layer becomes even more critical. The next lakehouse battleground is not Iceberg support itself, but whether vendors preserve genuine openness after adoption.