Data Lakehouse

A data lakehouse is a unified architecture that merges the cheap, flexible storage of a data lake with the high-performance management and ACID transactions of a data warehouse.

Modern data teams use the lakehouse to eliminate the friction of maintaining two separate systems for BI and machine learning. By implementing a metadata layer (like Delta Lake or Apache Iceberg) directly on top of low-cost cloud object storage, this architecture provides schema enforcement, data governance, and sub-second SQL performance. It serves as a single source of truth where analysts run complex reports and data scientists train models on the same raw data: no more expensive ETL pipelines or redundant data silos. Whether processing petabytes of streaming IoT logs or structured ERP records, the lakehouse scales storage and compute independently to keep costs low and reliability high.

https://www.databricks.com/glossary/data-lakehouse