BigQuery

BigQuery is Google Cloud's serverless data warehouse for SQL analytics, machine learning and streaming data, billed by data scanned or by slot capacity.

BigQuery lets a company query all of its data in SQL without provisioning or scaling a single server. Google builds it and sells it through Google Cloud, beside Looker, Dataflow and Dataproc. Storage and compute run as separate layers, so each one grows on its own as data volume and query load change. The product now reaches past warehousing into data preparation, machine learning and generative AI on the same tables. Queries run on slots, units of compute that autoscale inside reservations and wait in a queue when capacity is busy. Data arrives through batch loads, the BigQuery Data Transfer Service, the Storage Write API and Pub/Sub BigQuery subscriptions, while SQL-based continuous queries act on streaming rows as they land. Tables keep a time travel window, clone into writable copies, and feed materialized views that refresh incrementally. External tables read files in Cloud Storage, Bigtable and Google Drive in place, and Apache Iceberg tables let Spark and other open source engines work on one copy. BigQuery ML trains models in SQL, AI functions summarize text and score sentiment, and Python runs through user-defined functions and BigQuery DataFrames. The natural home for BigQuery is an organization that already runs applications on Google Cloud and wants its analytics in the same account. Analysts who think in SQL get a warehouse with no servers to provision, and data scientists get notebooks in Colab Enterprise and the Data Science Agent beside the tables they model. Data engineers can hand pipeline building and error detection to the Data Engineering Agent. Business users who never write SQL can put questions in plain language to the Conversational Analytics Agent. Governance teams get column-level data masking, row-level security and BigQuery sharing, which exposes datasets to other organizations without replicating them. Day to day, BigQuery holds the central copy of company data that dashboards and models read from. Events streamed in through Pub/Sub can trigger a continuous query the moment they land, and the results can feed dashboards in Looker. Teams moving off another warehouse use it as the landing point for migration, then layer geospatial analysis and AI functions onto the same tables. Managed disaster recovery replicates datasets to a second region for teams that need failover. BigQuery Omni carries the same SQL to data held in AWS and Azure regions, for companies whose records are spread across more than one cloud.