Topic
The warehouse compute tax
ELT does not just cost the vendor subscription. It also adds 20–50% to the warehouse compute bill: Snowflake ~42%, BigQuery ~28%, Databricks ~45%, Redshift ~20%. No vendor calculator surfaces this. Sync frequency is the dominant multiplier: a 15-minute sync is roughly 1.6x a daily sync.source - pulled 2026-06-21
Why the tax exists
The pipeline vendor moves the data; the warehouse does the work of writing, indexing, clustering, and materialising. The cost of that work is real but invisible to the pipeline vendor's calculator. Every sync triggers warehouse compute, and the higher the frequency, the higher the multiplier.
Multipliers by warehouse
- › Snowflake: 42% - credit consumption on warehouse ingestion + clustering keys.
- › BigQuery: 28% - slot-on-demand or reservation; lower because of column-store efficiency.
- › Databricks: 45% - DBU consumption on Auto Loader, Delta Live Tables.
- › Redshift: 20% - provisioned clusters absorb spikes without billing impact.
- › Self-hosted Postgres: 5% - pure infra, no managed-service markup.
What minimises the tax
Batched syncs (daily or hourly), wide-table replications, and column-store-friendly schemas. What maximises it: 15-minute or streaming syncs, narrow-table CDC, and complex clustering keys on Snowflake. When pricing ELT, always price the warehouse tax alongside.
Related
Written by Oliver Wakefield-Smith, Founder of Digital Signet. Independent reference, no vendor sponsorship.
Sources logged at /sources. Pricing-change history at /changelog. Last reviewed 2026-06-21.