This page outlines the current rate limits and quota limits for BigQuery BI Engine.
BI Engine is subject to the following capacity limits:
The maximum size of a BI Engine reservation is 100 GB per project per location. This limit does not affect the size of the tables that you query. BI Engine loads in-memory only the columns used in your queries, not the entire table. The correct compression algorithm is automatically chosen based on the data types that are stored in memory.
The data model size is limited to 10 GB per table. If you have a 100 GB reservation per project per location, BI Engine limits the reservation per table to 10 GB. The rest of the available reservation is used for other tables in the project.
BI Engine uses spare capacity to automatically replicate data across multiple nodes when running multiple concurrent queries, reducing overall available capacity. Replicated data is deleted automatically when concurrency demands are lowered, freeing up the used capacity.
Changes in capacity limits are announced via the BI Engine release notes.
If you run a query that produces query results larger than the size of your BI Engine capacity, the BI Engine self-tuning feature uses BigQuery slots to run the query. When BigQuery slots are used to run a query, you are charged based on BigQuery on-demand query pricing for the query job. When slots are used to run a query, all BigQuery quotas and limits on query jobs apply.