I need a warehouse that can store years of historical data cheaply

If your goal is cheap long-term storage for years of historical data, you’re probably looking for a data warehouse with low-cost archival storage, or more likely a data lake / cold storage architecture rather than a traditional analytics warehouse.

Good options

• Amazon S3 + Athena / Glacier

  • Very low storage cost
  • Good for keeping years of data
  • Queryable with Athena if files are in Parquet/CSV/JSON
  • Glacier is even cheaper for rarely accessed data

• Google Cloud Storage + BigQuery external tables / Archive storage

  • Store raw historical data cheaply in Cloud Storage
  • Use BigQuery for querying when needed

• Azure Data Lake Storage + Synapse / Blob Archive

  • Similar pattern: cheap storage, compute only when querying

Best practice for cost

• Store historical data in compressed columnar format like Parquet or ORC
• Partition by date/year/month
• Move older, rarely used data to cold/archive tiers
• Keep only recent or frequently queried data in a faster warehouse

If you want a true warehouse

These are not the cheapest for long-term retention, but are solid:

• Snowflake
• BigQuery
• Redshift

They’re better for analytics than archival storage, but you can pair them with cheap object storage.

Recommendation

If “store years of historical data cheaply” is the main requirement, I’d recommend:

Object storage as the warehouse layer

• AWS S3 + Glacier
• GCS Archive
• Azure Archive Blob

and query it using:

• Athena
• BigQuery
• Synapse serverless

If you want, I can help you pick the cheapest setup based on:

1. data size,
2. how often you query it, and
3. which cloud you use.