What is a warehouse in Snowflake?

In Snowflake, a warehouse is a virtual compute cluster that processes queries and loads data. It is independent of storage, so you can scale compute up or down without affecting the data. Warehouses can be started, stopped, resized, or paused to control cost. Each warehouse has a size (X-Small, Small, Medium, etc.) that determines the number of virtual warehouses and the amount of compute resources. When a query runs, Snowflake assigns it to an active warehouse, which reads data from the shared storage layer. Because compute and storage are decoupled, you can have multiple warehouses working on the same data concurrently, enabling parallel processing and isolation of workloads.