How Much Does a Cloud Data Warehouse Cost on AWS, Azure, or GCP?
A cloud data warehouse costs $50-$200/mo at 100GB with light query loads and $2,000-$10,000/mo at 10TB with heavy analytical queries. BigQuery's per-query pricing ($6.25/TB scanned) often wins for ad-hoc workloads with variable query volume.
Cloud data warehouses (Redshift, Synapse, BigQuery) price storage and compute differently. Some charge per-query, others per-cluster-hour. Your cost depends on data volume, query frequency, and concurrency. Use our calculator to compare data warehouse costs across all three providers.
Data Warehouse Pricing Models
Redshift charges per-node-hour, ranging from $0.25 to $13.04/hr depending on node type and size. Azure Synapse charges per-DWU-hour (Data Warehouse Unit). BigQuery charges per-TB scanned at $6.25/TB with 10GB of free storage. Serverless options (BigQuery on-demand, Redshift Serverless) eliminate idle costs but charge more per query. Choosing between provisioned and serverless depends on whether your query patterns are predictable or ad-hoc.
Storage vs Compute Costs
Storage is relatively cheap across all providers at $20–$25/TB/month. Compute is where costs diverge dramatically. A Redshift dc2.large cluster runs approximately $180/month. An equivalent Synapse DW100c costs roughly $1,200/month. BigQuery with 1TB of queries per month costs just $6.25 in compute. For ad-hoc analytics workloads with variable query volume, BigQuery's per-query pricing model often wins on cost efficiency.
Data Warehouse Costs by Data Volume
At 100GB of data with light query loads, expect $50–$200/month. At 1TB with moderate query frequency, costs range from $300–$1,500/month depending on the provider and provisioning model. At 10TB with heavy analytical queries, costs reach $2,000–$10,000/month. Partitioning tables and using columnar storage formats reduce query costs by 50–80% by limiting the amount of data scanned per query.
Key Cost Factors
- Compute: Query processing nodes or serverless credits
- Storage: Raw and transformed data retention
- Data ingestion: ETL and loading pipeline costs
- BI tool connections: Concurrent query slots
- Backup: Disaster recovery and snapshots
- Cross-region replication: Multi-region availability
Frequently Asked Questions
How much does a data warehouse cost on Redshift vs Synapse vs BigQuery?
Redshift charges per-node-hour ranging from $0.25 to $13.04/hr depending on node type and size — a dc2.large cluster runs roughly $180/month. Azure Synapse charges per-DWU-hour, with an equivalent DW100c costing roughly $1,200/month. BigQuery charges per-TB scanned at $6.25/TB with 10GB of free storage, so 1TB of queries per month costs just $6.25 in compute. At 100GB with light query loads, expect $50-$200/month; at 10TB with heavy queries, $2,000-$10,000/month.
What is the cheapest data warehouse for ad-hoc analytics?
BigQuery's per-query pricing model often wins for ad-hoc analytics workloads with variable and unpredictable query volume, since you only pay for data actually scanned rather than provisioned cluster time. For predictable, steady query workloads, provisioned options like Redshift or Synapse can be cheaper since you're not paying a per-query premium. Serverless options (BigQuery on-demand, Redshift Serverless) trade higher per-query cost for zero idle cost.
What drives the cost of a cloud data warehouse?
Compute for query processing is where costs diverge most dramatically between providers and pricing models. Storage is relatively cheap and consistent across providers at $20-$25/TB/month. Data ingestion for ETL and loading pipelines, BI tool connections requiring concurrent query slots, and backup/disaster recovery snapshots add further cost. Partitioning tables and using columnar storage formats reduce query costs by 50-80% by limiting data scanned per query.
How accurate are these data warehouse cost estimates?
Estimates use published on-demand list prices for Redshift, Synapse, and BigQuery in US regions, assuming 730 hours per month for provisioned clusters, last verified 2026-07-15. Query-based pricing (BigQuery) is estimated from typical scan volumes rather than your actual query patterns, so real costs can vary more than for provisioned options. See /methodology.