What is ELT (Extract, Load, Transform)?
ELT loads raw data into a warehouse first and transforms it there using the warehouse's compute, the modern inversion of ETL that suits cloud data warehouses and lets analysts model data in SQL after the fact.
ELT, Extract, Load, Transform, is a data integration pattern that flips the order of the classic ETL pipeline. Instead of transforming data in a separate processing layer before it lands, ELT extracts data from sources, loads it raw into a cloud data warehouse, and performs transformations inside the warehouse using its own scalable compute. This ordering became dominant because columnar cloud warehouses (BigQuery, Snowflake, Redshift, ClickHouse) made storage cheap and in-warehouse SQL transformation fast, so there is little reason to pre-process before loading. ELT's advantages are flexibility and reproducibility: because the raw data is preserved, analysts can re-model it later without re-extracting, and transformation logic lives as version-controlled SQL (often managed by tools like dbt) rather than buried in an opaque pipeline. The tradeoff is that you store more raw data and push transformation cost onto the warehouse. ELT is the standard architecture of the modern data stack, contrasted with ETL (transform-first, suited to constrained targets) and reverse ETL (pushing modeled warehouse data back out to operational tools). For an AI agent querying organizational data, understanding which tables are raw versus modeled, and where the transformation logic lives, is valuable shared context.